The critical energy infrastructure is undergoing two significant transformations: the rapid increase in renewable distributed energy resources (DER) and the digitalization of the energy sector, collectively shaping what is known as the Internet of Energy (IoE). Artificial intelligence (AI) has become a widely adopted tool for effectively allocating energy and managing sector-related resources, where ensuring fairness is essential. While inherent unfairness in AI systems is well acknowledged, little attention has been given to evaluating this unfairness and its real-world implications within the context of the IoE. In this study, we take a first step to elucidate the unfairness in AI-powered IoE systems induced by malicious users. We introduce Unfairness Score (UScore), a novel metric designed to evaluate the unfairness of machine learning models in real-world IoE scenarios. We then extensively evaluate unfairness attacks using three IoE tabular datasets, demonstrating that AI model fairness can be compromised through data poisoning, whether in centralized learning (CL) or federated learning (FL) settings. Notably, such compromises can occur when malicious users tamper with only a small subset of the data they control. Finally, we propose a novel approach that unifies fairness and differential privacy (DP) by leveraging DP as a provable defense mechanism. This approach provides a universally applicable solution to unfairness attacks, regardless of whether the learning tasks are classification or regression, and is effective in both FL and CL settings. Our contributions represent a significant step in addressing unfairness and privacy concerns in AI-powered IoE systems.

Web 3.0, built on blockchain technology, prioritizes user privacy and autonomy, presenting new opportunities for financial systems while also complicating the regulation of illicit activities. In this study, we present a novel infrastructure named Pseudo Fine-tuning (PFT) that provides account matching services to combat financial crimes on account-based blockchains such as money laundering through coin-mixing services. The significance of PFT lies in overcoming the need for real labels to fine-tune language models for account matching, given the limited availability of labeled account pairs for the task. Specifically, our design involves (1) crafting pseudo-labeled pairs from transactions of an account across different periods, and (2) fine-tuning language models to distill knowledge from pseudo pairs, which is transferable to the target task. We provide an in-depth analysis to investigate the inherent knowledge acquired during the PFT process and the conditions conducive to its effectiveness. Comprehensive experiments on real-world datasets collected from coin-mixing services and ENS name services, corroborate that the framework delivers pronounced enhancements over state-of-the-art approaches. Our implementation is released at https://github.com/git-disl/PFT .

The dynamic nature of streaming data often introduces distribution shifts that challenge typical text classification models. This paper proposes an online learning framework tailored for streaming text classification under distribution shifts. First, we decompose a neural network-based text classification model into distinct modules and analyze the varying impact of updating these modules under different types of shifts. Based on this insight, we define three novel indicators to efficiently measure the extent of distribution shifts without evaluating the entire model. These indicators enable the development of predictive models that dynamically optimize module update strategies, balancing learning efficiency and accuracy in real-time. To the best of our knowledge, this is the first approach to systematically adapt model updates according to a trade-off between efficiency and accuracy in online text classification. Extensive experiments on real-world streaming datasets demonstrate the effectiveness of our method, which consistently outperforms both static update strategies and state-of-the-art online text classification models. Our code and data are publicly available athttps://github.com/bigbases/modular-online-adaptation.

Graph Contrastive Learning (GCL) has demonstrated substantial promise in enhancing the robustness and generalization of recommender systems, particularly by enabling models to leverage large-scale unlabeled data for improved representation learning. However, in this article, we reveal an unexpected vulnerability: the integration of GCL inadvertently increases the susceptibility of a recommender to targeted promotion attacks. Through both theoretical investigation and empirical validation, we identify the root cause as the spectral smoothing effect induced by contrastive optimization, which disperses item embeddings across the representation space and unintentionally enhances the exposure of target items. Building on this insight, we introduce a bi-level optimization attack method, named graph Contrastive Learning Recommendation Attack (CLeaR), which deliberately amplifies spectral smoothness and enables a systematic investigation of the susceptibility of GCL-based recommendation models to targeted promotion attacks. Our findings highlight the urgent need for robust countermeasures; in response, we further propose a Spectral-Irregularity Mitigation framework, named SIM, which accurately detects and suppresses targeted items without compromising model performance. Extensive experiments on multiple benchmark datasets demonstrate that, compared to existing targeted promotion attacks, GCL-based recommendation models exhibit greater susceptibility when evaluated with CLeaR, while SIM effectively mitigates these vulnerabilities.

Developing grounding techniques for LLMs poses two requirements for interactive environments, i.e., (i) the presence of rich knowledge beyond the scope of existing LLMs and (ii) the complexity of tasks that require strategic reasoning. Existing environments fail to meet both requirements due to their simplicity or reliance on commonsense knowledge already encoded in LLMs for interaction. In this article, we present PokéLLMon, a new benchmark enriched with fictional game knowledge and characterized by the intense, dynamic, and adversarial gameplay of Pokémon battles, setting new challenges for the development of grounding and reasoning techniques in interactive environments. Empirical evaluations demonstrate that existing LLMs lack game knowledge and struggle in Pokémon battles. We investigate grounding techniques that leverage feedback and game knowledge, and provide a thorough analysis of reasoning methods from a new perspective of action consistency. Additionally, we introduce higher-level reasoning challenges when playing against human players. The implementation of our benchmark is released at: https://github.com/git-disl/PokeLLMon .

Deep graph clustering aims to uncover latent structural patterns within user behavior graphs to effectively group users based on behavioral similarity, thereby enhancing the efficiency of personalized service recommendation. However, most existing generative clustering methods heavily rely on encoder–decoder architectures and their pretraining procedures, which not only increase training complexity but also result in insufficient robustness and unstable optimization. To address these limitations, this paper proposes an Anchor-Guided Graph Clustering Network (AGCN) that eliminates the need for pretraining. The proposed method introduces an anchor-guided learning strategy that automatically selects high-confidence anchor samples farthest from the distribution boundary and utilizes the k-means++ algorithm to generate pseudo-labels in the diffusion feature space. Furthermore, AGCN integrates an anchor expansion strategy and a commonality enhancement mechanism to significantly improve clustering discriminability and recommendation accuracy. Theoretical analysis and extensive experiments on multiple real-world datasets demonstrate that the proposed method outperforms state-of-the-art approaches in terms of both clustering accuracy and stability.

Gradient leakage attacks pose a significant privacy threat in distributed learning-as-a-service APIs. Existing literature on gradient leakage defense relies on gradient perturbation for preventing privacy leakage. However, determining where and how much to perturb the gradient offers different capabilities for preventing gradient leakage. This paper presents GradCloak, a principled approach to guiding gradient perturbation with theoretical robustness bounds in federated learning as a service, aiming to find the minimum required noise for simultaneously achieving privacy protection, competitive accuracy, and preventing gradient leakage attacks. The paper is organized into three major components.First, we formulate the gradient leakage threats and their adverse effect. We categorize the attack into two broad types: leakage during local training and leakage before global aggregation.Second, we investigate different gradient perturbation approaches. We analyze and compare these gradient perturbation methods, which are performed at the federated server, with those performed at the participating client(s).Third, we introduce three robustness properties of robust perturbation against gradient leakage threats, formulated bythe anonymization boundfor training data robustness,the perturbation boundfor gradient robustness, andthe distribution robustness boundfor perturbed gradients. We conduct extensive evaluations on eight benchmark datasets to demonstrate that specific settings of gradient perturbation exist that best balance privacy, accuracy, and leakage prevention. Code is available athttps://github.com/git-disl/GradCloak.

Search engine personalization, while enhancing user satisfaction, can lead to information disparities. Previous studies on this topic face limitations, such as the absence of context-aware data collection, superficial URL-level analysis, and human-dependent annotations. We propose FAIR-SE, a Framework for Analyzing Information dispaRities in Search Engines that addresses these challenges through AWS Lambda-based concurrent data collection and LLM-generated persona-based content analysis. We collected search results across four user contexts (Search History, Geo-location, Language Preference, and Access Environment) and analyzed them through four analytical perspectives (Political Leaning, Topic-specific Stance, Subjectivity, and Bias). Experiments conducted on two globally prominent search engines across nine controversial topics demonstrate the efficacy of FAIR-SE regarding benchmark accuracy, persona consistency, and ability to reflect real-world discourse patterns across diverse topics. Our statistical analysis identifies distinct search engine characteristics and demonstrates significant information disparities in our case studies examining regional disparities in search results. Our code and datasets are publicly available at: https://github.com/bigbases/FAIR-SE.

Adversarial perturbations are useful tools for exposing vulnerabilities in neural networks. Existing adversarial perturbation methods for object detection are either limited to attacking CNN-based detectors or weak against transformer-based detectors. This paper presents an Attention-Focused Offensive Gradient (AFOG) attack against object detection transformers. By design, AFOG is neural-architecture agnostic and effective for attacking both large transformer-based object detectors and conventional CNN-based detectors with a unified adversarial attention framework. This paper makes three original contributions. First, AFOG utilizes a learnable attention mechanism that focuses perturbations on vulnerable image regions in multi-box detection tasks, increasing performance over non-attention baselines by up to 30.6%. Second, AFOG's attack loss is formulated by integrating two types of feature loss through learnable attention updates with iterative injection of adversarial perturbations. Finally, AFOG is an efficient and stealthy adversarial perturbation method. It probes the weak spots of detection transformers by adding strategically generated and visually imperceptible perturbations which can cause well-trained object detection models to fail. Extensive experiments conducted with twelve large detection transformers on COCO demonstrate the efficacy of AFOG. Our empirical results also show that AFOG outperforms existing attacks on transformer-based and CNN-based object detectors by up to 83% with superior speed and imperceptibility. Code is available at https://github.com/zacharyyahn/AFOG.

Federated learning (FL) is a privacy-preserving machine learning algorithm that enables multiple clients to collaborate. To respond to non-independent and identically distributed (non-IID) environments between clients, personalized FL (PFL) has been actively investigated. The typical PFL model consists of two parts: 1) the head (i.e., classifier) for the final classification and 2) the body (i.e., feature extractor) for extracting representations from local datasets. The head is maintained separately in each client for personalization; the body is aggregated for generalization. FedSDP introduces a bridge layer, called a personalized layer, between the head and the body to preserve individual, non-shared local prototypes for each client. A personalized layer decouples the body and head, strengthening the generalization and personalization, respectively. Based on this architecture, this study proposes a new PFL framework, Federated Self-Derived Prototypes (FedSDP), to dynamically balance personalization and generalization. To this end, we introduce two dynamic adjustments for generating self-derived prototypes: 1) global-local similarity weight (GL-Sim Weight) and 2) personalization early stopping indicator (P-Stop Indicator). GL-Sim Weight based on the similarity between the global and local prototypes is utilized to adjust the degree of personalization of each local model. PStop Indicator is calculated based on the changed degree of local parameters in each client, determining the early stopping for personalization in the client and further concentrating on generalization. Our comprehensive experiments demonstrate that FedSDP outperforms existing state-of-the-art FL frameworks, showing superior effectiveness in non-IID settings. Our code and data are available at https://github.com/bigbases/FedSDP.

Harmful fine-tuning attack poses serious safety concerns for large language models' fine-tuning-as-a-service. While existing defenses have been proposed to mitigate the issue, their performances are still far away from satisfactory, and the root cause of the problem has not been fully recovered. To this end, we in this paper show that \textit{harmful perturbation} over the model weights could be a probable cause of alignment-broken. In order to attenuate the negative impact of harmful perturbation, we propose an alignment-stage solution, dubbed Booster. Technically, along with the original alignment loss, we append a loss regularizer in the alignment stage's optimization. The regularizer ensures that the model's harmful loss reduction after the simulated harmful perturbation is attenuated, thereby mitigating the subsequent fine-tuning risk. Empirical results show that Booster can effectively reduce the harmful score of the fine-tuned models while maintaining the performance of downstream tasks. Our code is available at https://github.com/git-disl/Booster

Safety aligned Large Language Models (LLMs) are vulnerable to harmful fine-tuning attacks – a few harmful data mixed in the fine-tuning dataset can break the LLMs’s safety alignment. While several defenses have been proposed, our evaluation shows that existing defenses fail when some specific training hyper-parameters are chosen – a large learning rate or a large number of training epochs in the fine-tuning stage can easily invalidate the defense. To this end, we propose Antidote, a post-fine-tuning stage solution, which remains agnostic to the training hyper-parameters in the fine-tuning stage. Antidote relies on the philosophy that by removing the harmful parameters, the harmful model can be recovered from the harmful behaviors, regardless of how those harmful parameters are formed in the fine-tuning stage. With this philosophy, we introduce a one-shot pruning stage after harmful fine-tuning to remove the harmful weights that are responsible for the generation of harmful content. Despite its embarrassing simplicity, empirical results show that Antidote can reduce harmful score while maintaining accuracy on downstream tasks.

The intrinsic qualities of the underwater acoustic channel (UAC), such as multipath propagation and motion-induced Doppler shift, may result in large direction of arrival (DOA) estimation bias. In this paper, we propose a two-dimensional DOA estimation method for the underwater acoustic signals with gain-phase errors, coined as ML-ANM (Maximum Likelihood-Atomic Norm Minimization). We assume that the number of received signals satisfies the parameter identifiability condition. First, we propose a bimodal atomic norm involving noncoherent and coherent signals under the maximum likelihood estimation (MLE) framework, which allows simultaneous processing of noncoherent and coherent signals without decorrelation. To avoid the direct calculation of gain-phase errors, we transform the primal MLE optimization model into an ML-ANM optimization problem with two positive semidefinite constraints. Second, we utilize oblique projection to divide the optimal problem into two optimal subproblems involving noncoherent and coherent signals, and prove the equivalence of the problems and their solutions. Then, we obtain the global optimal solutions for the two optimal subproblems. Third but not the least, we provide the Cramér-Rao lower bound (CRB) and prove that the asymptotic error covariance matrix of ML-ANM achieves the CRB. Extensive experiments through realistic ocean simulations demonstrate the effectiveness of the ML-ANM method.

This article presents FusionShot, a focal diversity-optimized few-shot ensemble learning approach for boosting the robustness and generalization performance of pre-trained few-shot models. The article makes three original contributions. First, we explore the unique characteristics of few-shot learning to ensemble multiple few-shot (FS) models by creating three alternative fusion channels. Second, we introduce the concept of focal error diversity to learn the most efficient ensemble teaming strategy, rather than assuming that an ensemble of a larger number of base models will outperform those sub-ensembles of smaller size. We develop a focal diversity ensemble pruning method to effectively prune out the candidate ensembles with low ensemble error diversity and recommend top- \( K \) FS ensembles with the highest focal error diversity. Finally, we capture the complex non-linear patterns of ensemble few-shot predictions by designing the learn-to-combine algorithm, which can learn the diverse weight assignments for robust ensemble fusion over different member models. Extensive experiments on representative few-shot benchmarks show that the top-K ensembles recommended by FusionShot can outperform the representative state-of-the-art (SOTA) few-shot models on novel tasks (different distributions and unknown at training) and can prevail over existing few-shot learners in both cross-domain settings and adversarial settings. For reproducibility purposes, FusionShot trained models, results, and code are made available at https://github.com/sftekin/fusionshot .

Continual learning aims to progressively acquire new knowledge while retaining previously learned information, addressing the challenge of catastrophic forgetting. This paper introduces a novel continual learning method, called SPrint, which is the first research effort to devise the principles of self-paced learning for solving continual learning problems. SPrint dynamically adapts the complexity of samples for both new and previous tasks in response to the model’s current learning capacity. It employs a self-paced loss function for sampling new tasks and a forgetting occurrence for sampling previous tasks with replay memory. Through extensive empirical evaluation, we demonstrate that SPrint consistently outperforms state-of-the-art methods in various continual learning benchmarks. Our source code is publicly available at https://github.com/bigbases/SPrint.

This study presents an adaptive online learning method for text classification under distribution shifts. We formulate a typical neural network-based text classification model as multiple logical modules. By leveraging the characteristics of the modules, we introduce three novel indicators to effectively measure the degree of dynamic distribution shifts without evaluating the model. To enhance online learning, we tactically trade off between learning efficiency and accuracy based on distribution shifts measured in real time. To the best of our knowledge, this is the first effort to adapt the model to the preference of learning efficiency or accuracy for online text classification. Extensive experiments on real-world streaming text datasets show that our method outperforms the best static update strategy and state-of-the-art online text classification models. Our code and data are available at https://github.com/bigbases/online-learning-text.

In this paper, we present a neural network-enabled data distribution aware sorting method, coined as NN-sort. Our approach explores the potential of developing deep learning techniques to speed up large-scale sort operations, enabling data distribution aware sorting as a deep learning service. Compared to traditional pairwise comparison-based sorting algorithms, which sort data elements by performing pairwise operations, NN-sort leverages the neural network model to learn the data distribution and uses it to map large-scale data elements into ordered ones. Our experiments demonstrate the significant advantage of using NN-sort. Measurements on both synthetic and real-world datasets show that NN-sort yields 2.18× to 10× performance improvement over traditional sorting algorithms.

With the recent advances in vision transformers and large language models (LLMs),finetuning costly large mod-els on downstream learning tasks poses significant chal-lenges under limited computational resources. This pa-per presents a REsource and ComputAtion-efficient Pruning framework (RECAP) for the finetuning of transformer-based large models. RECAP by design bridges the gap between efficiency and performance through an iterative process cycling between pruning, finetuning, and updating stages to explore different chunks of the given large-scale model. At each iteration, we first prune the model with Taylor-approximation-based importance estimation and then only update a subset of the pruned model weights based on the Fisher-information criterion. In this way, RE-CAP achieves two synergistic and yet conflicting goals: re-ducing the GPU memory footprint while maintaining model performance, unlike most existing pruning methods that re-quire the model to be finetuned beforehand for better preser-vation of model performance. We perform extensive exper-iments with a wide range of large transformer-based archi-tectures on various computer vision and natural language understanding tasks. Compared to recent pruning techniques, we demonstrate that RECAP offers significant im-provements in GPU memory efficiency, capable of reducing the footprint by up to 65%.

Graph Neural Networks (GNNs) have become an important graph feature learning paradigm that is extensively applied to graph inference tasks. However, GNNs still have limitations in some aspects such as representation capability, interpretability, etc. To this end, this paper introduces a novel hybrid Fuzzy-Logic Graph Neural Network (FL-GNN) that synergistically combines Fuzzy Neural Network (FNN) and GNN to effectively capture and aggregate local information flows within the graph. FL-GNN exhibits three distinctive features. First, we incorporate a specialized structure fuzzy rule to enhance FL-GNN’s graph inference capability, surpassing representative GNN models. Second, we augment the interpretability of FL-GNN by integrating analytical exploration methods from two perspectives: Fuzzy Inference System and Message Passing Algorithm (MPA). Lastly, we refine the structure of FL-GNN based on MPA to optimize its calculation complexity and consequently improve learning efficiency. Experimental results show that FL-GNN outperforms existing representative graph neural networks for graph inference tasks.

Hybrid cloud provides an attractive solution to microservices for better resource elasticity. A subset of application components can be offloaded from the on-premises cluster to the cloud, where they can readily access additional resources. However, the selection of this subset is challenging because of the large number of possible combinations. A poor choice degrades the application performance, disrupts the critical services, and increases the cost to the extent of making the use of hybrid cloud unviable. This paper presents Atlas, a hybrid cloud migration advisor. Atlas uses a data-driven approach to learn how each user-facing API utilizes different components and their network footprints to drive the migration decision. It learns to accelerate the discovery of high-quality migration plans from millions and offers recommendations with customizable trade-offs among three quality indicators: end-to-end latency of user-facing APIs representing application performance, service availability, and cloud hosting costs. Atlas continuously monitors the application even after the migration for proactive recommendations. Our evaluation shows that Atlas can achieve 21% better API performance (latency) and 11% cheaper cost with less service disruption than widely used solutions.

Recent advancements in conversational large language models (LLMs), such as ChatGPT, have demonstrated remarkable promise in various domains, including drug discovery. However, existing works mainly focus on investigating the capabilities of conversational LLMs on chemical reactions and retrosynthesis. While drug editing, a critical task in the drug discovery pipeline, remains largely unexplored. To bridge this gap, we propose ChatDrug, a framework to facilitate the systematic investigation of drug editing using LLMs. ChatDrug jointly leverages a prompt module, a retrieval and domain feedback module, and a conversation module to streamline effective drug editing. We empirically show that ChatDrug reaches the best performance on all 39 drug editing tasks, encompassing small molecules, peptides, and proteins. We further demonstrate, through 10 case studies, that ChatDrug can successfully identify the key substructures for manipulation, generating diverse and valid suggestions for drug editing. Promisingly, we also show that ChatDrug can offer insightful explanations from a domain-specific perspective, enhancing interpretability and enabling informed decision-making.

Predicting information diffusion is a fundamental task in online social networks (OSNs).Recent studies mainly focus on the popularity prediction of specific content but ignore the correlation between multiple pieces of information.The topic is often used to correlate such information and can correspond to multi-source information.The popularity of a topic relies not only on information diffusion time but also on users' followership.Current solutions concentrate on hard time partition, lacking versatility.Meanwhile, the hop-based sampling adopted in state-of-the-art (SOTA) methods encounters redundant user followership.Moreover, many SOTA methods are not designed with good modularity and lack evaluation for each functional module and enlightening discussion.This paper presents a novel extensible framework, coined as HIF, for effective popularity prediction in OSNs with four original contributions.First, HIF adopts a soft partition of users and time intervals to better learn users' behavioral preferences over time.Second, HIF utilizes weighted sampling to optimize the construction of heterogeneous graphs and reduce redundancy.Furthermore, HIF supports multi-task collaborative optimization to improve its learning capability.Finally, as an extensible framework, HIF provides generic module slots to combine different submodules (e.g., RNNs,

The new paradigm of fine-tuning-as-a-service introduces a new attack surface for Large Language Models (LLMs): a few harmful data uploaded by users can easily trick the fine-tuning to produce an alignment-broken model. We conduct an empirical analysis and uncover a \textit{harmful embedding drift} phenomenon, showing a probable cause of the alignment-broken effect. Inspired by our findings, we propose Vaccine, a perturbation-aware alignment technique to mitigate the security risk of users fine-tuning. The core idea of Vaccine is to produce invariant hidden embeddings by progressively adding crafted perturbation to them in the alignment phase. This enables the embeddings to withstand harmful perturbation from un-sanitized user data in the fine-tuning phase. Our results on open source mainstream LLMs (e.g., Llama2, Opt, Vicuna) demonstrate that Vaccine can boost the robustness of alignment against harmful prompts induced embedding drift while reserving reasoning ability towards benign prompts. Our code is available at https://github.com/git-disl/Vaccine.

Recent studies show that Large Language Models (LLMs) with safety alignment can be jail-broken by fine-tuning on a dataset mixed with harmful data. For the first time in the literature, we show that the jail-break effect can be mitigated by separating two states in the fine-tuning stage to respectively optimize over the alignment and user datasets. Unfortunately, our subsequent study shows that this simple Bi-State Optimization (BSO) solution experiences convergence instability when steps invested in its alignment state is too small, leading to downgraded alignment performance. By statistical analysis, we show that the \textit{excess drift} towards the switching iterates of the two states could be a probable reason for the instability. To remedy this issue, we propose \textbf{L}azy(\textbf{i}) \textbf{s}afety \textbf{a}lignment (\textbf{Lisa}), which introduces a proximal term to constraint the drift of each state. Theoretically, the benefit of the proximal term is supported by the convergence analysis, wherein we show that a sufficient large proximal factor is necessary to guarantee Lisa's convergence. Empirically, our results on four downstream fine-tuning tasks show that Lisa with a proximal term can significantly increase alignment performance while maintaining the LLM's accuracy on the user tasks. Code is available at https://github.com/git-disl/Lisa.

Well-trained deep neural networks (DNNs) treat all test samples equally during prediction. Adaptive DNN inference with early exiting leverages the observation that some test examples can be easier to predict than others. This paper presents EENet, a novel early-exiting scheduling framework for multi-exit DNN models. Instead of having every sample go through all DNN layers during prediction, EENet learns an early exit scheduler, which can intelligently terminate the inference earlier for certain predictions, which the model has high confidence of early exit. As opposed to previous early-exiting solutions with heuristics-based methods, our EENet framework optimizes an early-exiting policy to maximize model accuracy while satisfying the given per-sample average inference budget. Extensive experiments are conducted on four computer vision datasets (CIFAR-10, CIFAR-100, ImageNet, Cityscapes) and two NLP datasets (SST-2, AgNews). The results demonstrate that the adaptive inference by EENet can outperform the representative existing early exit techniques. We also perform a detailed visualization analysis of the comparison results to interpret the benefits of EENet.

Language models (LMs) have demonstrated superior performance in detecting fraudulent activities on Blockchains. Nonetheless, the sheer volume of Blockchain data results in excessive memory and computational costs when training LMs from scratch, limiting their capabilities to large-scale applications. In this paper, we present ZipZap, a framework tailored to achieve both parameter and computational efficiency when training LMs on large-scale transaction data. First, with the frequency-aware compression, an LM can be compressed down to a mere 7.5% of its initial size with an imperceptible performance dip. This technique correlates the embedding dimension of an address with its occurrence frequency in the dataset, motivated by the observation that embeddings of low-frequency addresses are insufficiently trained and thus negating the need for a uniformly large dimension for knowledge representation. Second, ZipZap accelerates the speed through the asymmetric training paradigm: It performs transaction dropping and cross-layer parameter-sharing to expedite the pre-training process, while revert to the standard training paradigm for fine-tuning to strike a balance between efficiency and efficacy, motivated by the observation that the optimization goals of pre-training and fine-tuning are inconsistent. Evaluations on real-world, large-scale datasets demonstrate that ZipZap delivers notable parameter and computational efficiency improvements for training LMs. Our implementation is available at: https://github.com/git-disl/ZipZap.

UAV-enabled data collection is considered a promising paradigm of emergency data transmission for IoT applications when the communication infrastructure is damaged. UAV scheduling for data collection as critical technology has attracted widespread attention. Most studies default to the absolute reliability of UAVs for data collection, yet it is inevitable for UAVs to fail in flight. It is unacceptable if some critical data is lost due to UAV faults. Therefore, we research fault-tolerant scheduling for data collection enabled by heterogeneous UAVs. Firstly, a three-layer data collection motivation scenario is proposed, where the fault tolerance issue is involved for the first time. Then, we propose a utility-based fault tolerance model-UBFT to balance the reliability and efficiency of data collection. The UAV fault-tolerant scheduling is modeled as a multi-objective optimization problem to concurrently optimize the data throughput and load balancing of data collection. Combining the characteristics of optimization objectives, an alternating coordinate optimization method-ACTOR is presented to solve this problem efficiently. Numerous simulation experiments and real-machine experiments demonstrate that ACTOR-UBFT achieves excellent performance in fault tolerance, data throughput, adaptation, algorithm complexity, etc.

Face recognition (FR) technologies have enabled many life-enriching applications but have also opened doors for potential misuse. Governments, private companies, or even individuals can scrape the web, collect facial images, and build a face database to fuel the FR system to identify human faces without their consent. This paper introduces PMask to combat such a privacy threat against unauthorized FR. It provides a holistic approach to enable privacy-preserving sharing of facial images. PMask preprocesses the facial image and hides its unique facial signature through iterative optimization with dual goals: (i) minimizing the amount of noise to ensure high image quality and (ii) minimizing the perception loss between the privacy-protected face and the original face to ensure the face is recognizable to be the same person by humans. Extensive experiments are conducted on eight representative FR models to evaluate PMask against unauthorized FR. The results validate that PMask provides much stronger protection, introduces less perceptible changes to facial images, and runs faster than state-of-the-art methods to provide privacy protection with a better user experience.

Recently, many post hoc graph neural network (GNN) explanation methods have been explored to uncover GNNs' predictive behaviors by analyzing the embeddings produced by the GNN models. However, these methods suffer from explanation ambiguity inherent in learned graph embeddings because aggregation-based embeddings can lead to the loss of unique identifiers for individual graph components and, thus, allow noncausal nodes that are adjacent to true causal patterns to unintentionally embody causal information in their embeddings, hindering the explanations from faithfully representing the true insights of GNNs' predictive reasoning. In this article, we present an embedding ambiguity-sensitive GNN explanation framework (EAGX). EAGX can effectively mitigate the impact of embedding-induced explanation ambiguity by creating edges' ambiguity feature extractor, exploring edges' predictive relevance, and integrating them into the explanation process, thereby capturing each graph component's contribution to the predictions. Specifically, we first propose a centroid-constrained fuzzy c-means algorithm to construct an ambiguity feature extractor. Then, we leverage the ambiguity features for edges to develop the ambiguity-based edge attribution module for assigning a prediction relevance score to each edge. Finally, instead of focusing only on the edges with high influence to the GNN prediction, we introduce a joint optimization strategy to refine the learning process of our edge attribution module, empowering EAGX to capture the subtle interplay of both causal and noncausal subgraphs on model predictions, which further improve the explainability of GNN predictions. Experimental results demonstrate that EAGX outperforms the leading explainers on most evaluation metrics, underscoring its effectiveness in generating reliable and precise explanations for GNNs.

Machine Learning (ML) techniques have been applied to many real-world applications to perform a wide range of tasks. In practice, ML models are typically deployed as the black-box APIs to protect the model owner’s benefits and/or defend against various privacy attacks. In this paper, we present Gradient-Leaks as the first evidence showcasing the possibility of performing membership inference attacks (MIAs), with mere black-box access, which aim to determine whether a data record was utilized to train a given target ML model or not. The key idea of Gradient-Leaks is to construct a local ML model around the given record which locally approximates the target model’s prediction behavior. By extracting the membership information of the given record from the gradient of the substituted local model using an intentionally modified autoencoder, Gradient-Leaks can thus breach the membership privacy of the target model’s training data in an unsupervised manner, without any priori knowledge about the target model’s internals or its training data. Extensive experiments on different types of ML models with real-world datasets have shown that Gradient-Leaks can achieve a better performance compared with state-of-the-art attacks.

Many real-world applications require precise and fast time-series forecasting. Recent trends in time-series forecasting models are shifting from LSTM-based models to Transformer-based models. However, the Transformer-based model has a limited ability to represent sequential relationships in time-series data. In addition, the transformer-based model suffers from slow training and inference speed due to the bottleneck incurred by a deep encoder and step-by-step decoder inference. To address these problems, we propose a time-series forecasting optimized Transformer model, called TS-Fastformer. TS-Fastformer introduces three new optimizations: First, we propose a Sub Window Tokenizer for compressing input in a simple manner. The Sub Window Tokenizer reduces the length of input sequences to mitigate the complexity of self-attention and enables both single and multi-sequence learning. Second, we propose Time-series Pre-trained Encoder to extract effective representations through pre-training. This optimization enables TS-Fastformer to capture both seasonal and trend representations as well as to mitigate bottlenecks of conventional transformer models. Third, we propose the Past Attention Decoder to forecast target by incorporating past long short-term dependency patterns. Furthermore, Past Attention Decoder achieves high performance improvement by removing a trend distribution that changes over a long period. We evaluate the efficiency of our model with extensive experiments using seven real-world datasets and compare our model to six representative time-series forecasting approaches. The results show that the proposed TS-Fastformer reduces MSE by 10.1% compared to state-of-the-art model and demonstrates 21.6% faster training time compared to the existing fastest transformer, respectively.

Deep neural network ensembles combine the wisdom of multiple deep neural networks to improve the generalizability and robustness over individual networks. It has gained increasing popularity to study and apply deep ensemble techniques in the deep learning community. Some mission-critical applications utilize a large number of deep neural networks to form deep ensembles to achieve desired accuracy and resilience, which introduces high time and space costs for ensemble execution. However, it still remains a critical challenge whether a small subset of the entire deep ensemble can achieve the same or better generalizability and how to effectively identify these small deep ensembles for improving the space and time efficiency of ensemble execution. This article presents a novel deep ensemble pruning approach, which can efficiently identify smaller deep ensembles and provide higher ensemble accuracy than the entire deep ensemble of a large number of member networks. Our hierarchical ensemble pruning approach (HQ) leverages three novel ensemble pruning techniques. First, we show that the focal ensemble diversity metrics can accurately capture the complementary capacity of the member networks of an ensemble team, which can guide ensemble pruning. Second, we design a focal ensemble diversity based hierarchical pruning approach, which will iteratively find high quality deep ensembles with low cost and high accuracy. Third, we develop a focal diversity consensus method to integrate multiple focal diversity metrics to refine ensemble pruning results, where smaller deep ensembles can be effectively identified to offer high accuracy, high robustness and high ensemble execution efficiency. Evaluated using popular benchmark datasets, we demonstrate that the proposed hierarchical ensemble pruning approach can effectively identify high quality deep ensembles with better classification generalizability while being more time and space efficient in ensemble decision making. We have released the source codes on GitHub at https://github.com/git-disl/HQ-Ensemble .

Data Poisoning is a dominating threat in the distributed learning-as-a-service API, where the mediator has limited control over the distributed client contributing to the joint model. Through an in-depth characterization of data poisoning risks in federated learning, this paper presents a comprehensive study towards demystifying data poisoning attacks from three perspectives.First, we formally define the targeted dirty-label data poisoning attack, which aims to cause the trained global model to only misclassify the input from a specific victim class with a designated malicious behavior. Then, we demonstrate theoretical statistical robustness in the eigenvalues of the covariance in the gradient update shared from the client to server when under the data poisoning attack.Second, we study the impact of attack timing and identify the most detrimental attack entry point during the federated training.Last, we examine several existing defenses against data poisoning in addition to the robust statistic detection. Through formal analysis and extensive empirical evidence, we investigate under what conditions the statistical robustness of data poisoning can serve as the forensic evidence for attack mitigation in federated-learning-as-a-service, at what attack timing the attack is most detrimental, and how the attack reacts in the presence of the existing defenses.

Ships as moving vehicles have intermittent connectivity and are connected to one another via ad-hoc networks with directional antennas. We refer to such Internet of Ships as IoSNET. Neighbor discovery (ND) is a fundamental requirement and a building block in an IoSNET. However, existing directional antenna-based ND protocols are known to be vulnerable under multi-end wormhole attacks. By analyzing such vulnerabilities, we propose a Directional Antennas powered SEcure Neighbor Discovery protocol, coined asDASEND, for protecting IoSNET against two-end and multi-end wormhole attacks. TheDASENDprotocol is novel in two aspects. First, we construct two timing constraints for detecting the wormhole attacks using a linear clock model without strict time synchronization and propose a message authentication mechanism for ensuring the security of the time stamps. Second, we propose a virtual grid structure to model the relative locations of transmitters jointly using their direction-of-arrival (DOA) and received signal strength (RSS) values. Based on this model, we propose a beam state-based steering approach to shorten the discovery latency of the ND process. Extensive experiments through IoSNET simulations demonstrate the effectiveness of ourDASENDprotocol for delivering secure neighbor discovery as a service in the Internet of Ships environment under different movement scenarios.

The process model repository has demonstrated unprecedented success in a variety of industrial and process as a service scenarios. With the rapid increase of massive business process-related data under Industry 4.0, effectively retrieval of process models from large process model repositories becomes a critical challenge for process mining, process deployment and process model acquisition. To accelerate the retrieval of process models from a large process repository, existing retrieval methods rely solely on building single dimension process model indices. In this article we show that this single dimension indexing approach is not only inefficient but also cumbersome for supporting high performance retrieval services over large process model repositories. We propose a new business process model indexing and retrieval with structure and behavior fusion. In the indexing stage, we propose a process model index generation paradigm method with two novel features. First, our index algorithm can transform thetrace equivalent process model(TEPM) with complex structures into a process tree, which can better capture process sequence semantics than the existing approach based on block structured process model. Second, we improve the method for computing the process tree edit distance for measuring process model similarity by introducing the process tree similarity method, which can distinguish leaf nodes and non-leaf nodes and improve the limitations of the traditional edit distance algorithm. Extensive experiments using real world process repositories demonstrate that the proposed methods are under polynomial time in both the model index generation and model querying stages, and offer superior retrieval performance compared to existing process model retrieval methods in terms of efficiency, search capability and scope.

Federated Learning (FL) has been gaining popularity as a collaborative learning framework to train deep learning-based object detection models over a distributed population of clients. Despite its advantages, FL is vulnerable to model hijacking. The attacker can control how the object detection system should misbehave by implanting Trojaned gradients using only a small number of compromised clients in the collaborative learning process. This paper introduces STDLens, a principled approach to safeguarding FL against such attacks. We first investigate existing mitigation mechanisms and analyze their failures caused by the inherent errors in spatial clustering analysis on gradients. Based on the insights, we introduce a three-tier forensic framework to identify and expel Trojaned gradients and reclaim the performance over the course of FL. We consider three types of adaptive attacks and demonstrate the robustness of STDLens against advanced adversaries. Extensive experiments show that STDLens can protect FL against different model hijacking attacks and outperform existing methods in identifying and removing Trojaned gradients with significantly higher precision and much lower false-positive rates. The source code is available at https://github.com/git-disl/STDLens.

Federated learning (FL) is an attractive distributed learning paradigm supporting real-time continuous learning and client privacy by default. In most FL approaches, all edge clients are assumed to have sufficient computation capabilities to participate in the learning of a deep neural network (DNN) model. However, in real-life applications, some clients may have severely limited resources and can only train a much smaller local model. This paper presents ScaleFL, a novel FL approach with two distinctive mechanisms to handle resource heterogeneity and provide an equitable FL framework for all clients. First, ScaleFL adaptively scales down the DNN model along width and depth dimensions by leveraging early exits to find the best-fit models for resource-aware local training on distributed clients. In this way, ScaleFL provides an efficient balance of preserving basic and complex features in local model splits with various sizes for joint training while enabling fast inference for model deployment. Second, ScaleFL utilizes self-distillation among exit predictions during training to improve aggregation through knowledge transfer among subnetworks. We conduct extensive experiments on benchmark CV (CIFAR-10/100, ImageNet) and NLP datasets (SST-2, AgNews). We demonstrate that ScaleFL outperforms existing representative heterogeneous FL approaches in terms of global/local model performance and provides inference efficiency, with up to 2x latency and 4x model size reduction with negligible performance drop below 2%.

Federated learning (FL) is a distributed learning framework that inherently provides data privacy and parallel computation capability over a set of participating devices (clients). In real-life applications, these clients can have a great variety in terms of resources (storage, RAM, CPU/GPU speed, network speed, etc.). However, most previous FL studies do not consider this scenario with system heterogeneity and assume that all clients can operate on the same full-size deep neural network (DNN) model. In this work, we demonstrate a scalable FL approach, ScaleFL, which tackles system heterogeneity through hierarchically downscaling the DNN model for clients with limited resources. ScaleFL utilizes early exits to form multi-exit DNN models by injecting early exit networks into the given DNN. During FL, the model is adaptively split along depth (exits) and width (hidden dimensions) based on the resource budget of each participating client. A proof-of-concept demonstration is provided with interactive features, demonstrating the system flow on image classification and NLP benchmark workloads.

Deep neural network ensembles hold the potential of improving generalization performance for complex learning tasks. This paper presents formal analysis and empirical evaluation to show that heterogeneous deep ensembles with high ensemble diversity can effectively leverage model learning heterogeneity to boost ensemble robustness. We first show that heterogeneous DNN models trained for solving the same learning problem, e.g., object detection, can significantly strengthen the mean average precision (mAP) through our weighted bounding box ensemble consensus method. Second, we further compose ensembles of heterogeneous models for solving different learning problems, e.g., object detection and semantic segmentation, by introducing the connected component labeling (CCL) based alignment. We show that this two-tier heterogeneity driven ensemble construction method can compose an ensemble team that promotes high ensemble diversity and low negative correlation among member models of the ensemble, strengthening ensemble robustness against both negative examples and adversarial attacks. Third, we provide a formal analysis of the ensemble robustness in terms of negative correlation. Extensive experiments validate the enhanced robustness of heterogeneous ensembles in both benign and adversarial settings. The appendix and source codes are available on GitHub at https://github.com/git-disl/HeteRobust.

Gradient leakage attacks are dominating privacy threats in federated learning, despite the default privacy that training data resides locally at the clients. Differential privacy has been the de facto standard for privacy protection and is deployed in federated learning to mitigate privacy risks. However, much existing literature points out that differential privacy fails to defend against gradient leakage. The paper presents ModelCloak, a principled approach based on differential privacy noise, aiming for safe-sharing client local model updates. The paper is organized into three major components. First, we introduce the gradient leakage robustness trade-off, in search of the best balance between accuracy and leakage prevention. The trade-off relation is developed based on the behavior of gradient leakage attacks throughout the federated training process. Second, we demonstrate that a proper amount of differential privacy noise can offer the best accuracy performance within the privacy requirement under a fixed differential privacy noise setting. Third, we propose dynamic differential privacy noise and show that the privacy-utility trade-off can be further optimized with dynamic model perturbation, ensuring privacy protection, competitive accuracy, and leakage attack prevention simultaneously.

Blockchain provides the unique and accountable channel for financial forensics by mining its open and immutable transaction data. A recent surge has been witnessed by training machine learning models with cryptocurrency transaction data for anomaly detection, such as money laundering and other fraudulent activities. This paper presents a holistic applied data science approach to fraud detection in the Bitcoin network with two original contributions. First, we contribute the Elliptic++dataset, which extends the Elliptic transaction dataset to include over 822k Bitcoin wallet addresses (nodes), each with 56 features, and 1.27M temporal interactions. This enables both the detection of fraudulent transactions and the detection of illicit addresses (actors) in the Bitcoin network by leveraging four types of graph data: (i) the transaction-to-transaction graph, representing the money flow in the Bitcoin network, (ii) the address-to-address interaction graph, capturing the types of transaction flows between Bitcoin addresses, (iii) the address-transaction graph, representing the bi-directional money flow between addresses and transactions (BTC flow from input address to one or more transactions and BTC flow from a transaction to one or more output addresses), and (iv) the user entity graph, capturing clusters of Bitcoin addresses representing unique Bitcoin users. Second, we perform fraud detection tasks on all four graphs by using diverse machine learning algorithms. We show that adding enhanced features from the address-to-address and the address-transaction graphs not only assists in effectively detecting both illicit transactions and illicit addresses, but also assists in gaining in-depth understanding of the root cause of money laundering vulnerabilities in cryptocurrency transactions and the strategies for fraud detection and prevention. The Elliptic++ dataset is released at https://www.github.com/git-disl/EllipticPlusPlus.

Federated learning (FL) is vulnerable to backdoor attacks due to its distributed computing nature. Existing defense solution usually requires larger amount of computation in either the training or testing phase, which limits their practicality in the resource-constrain scenarios. A more practical defense, i.e., neural network (NN) pruning based defense has been proposed in centralized backdoor setting. However, our empirical study shows that traditional pruning-based solution suffers \textit{poison-coupling} effect in FL, which significantly degrades the defense performance.This paper presents Lockdown, an isolated subspace training method to mitigate the poison-coupling effect. Lockdown follows three key procedures. First, it modifies the training protocol by isolating the training subspaces for different clients. Second, it utilizes randomness in initializing isolated subspacess, and performs subspace pruning and subspace recovery to segregate the subspaces between malicious and benign clients. Third, it introduces quorum consensus to cure the global model by purging malicious/dummy parameters. Empirical results show that Lockdown achieves \textit{superior} and \textit{consistent} defense performance compared to existing representative approaches against backdoor attacks. Another value-added property of Lockdown is the communication-efficiency and model complexity reduction, which are both critical for resource-constrain FL scenario. Our code is available at \url{https://github.com/git-disl/Lockdown}.

As various forms of fraud proliferate on Ethereum, it is imperative to safeguard against these malicious activities to protect susceptible users from being victimized. While current studies solely rely on graph-based fraud detection approaches, it is argued that they may not be well-suited for dealing with highly repetitive, skew-distributed and heterogeneous Ethereum transactions. To address these challenges, we propose BERT4ETH, a universal pre-trained Transformer encoder that serves as an account representation extractor for detecting various fraud behaviors on Ethereum. BERT4ETH features the superior modeling capability of Transformer to capture the dynamic sequential patterns inherent in Ethereum transactions, and addresses the challenges of pre-training a BERT model for Ethereum with three practical and effective strategies, namely repetitiveness reduction, skew alleviation and heterogeneity modeling. Our empirical evaluation demonstrates that BERT4ETH outperforms state-of-the-art methods with significant enhancements in terms of the phishing account detection and de-anonymization tasks. The code for BERT4ETH is available at: https://github.com/git-disl/BERT4ETH.

With the continuous expansion of the Internet of Things (IoT) application scope, there are growing IoT scenarios that lack the coverage of wireless communication networks have the demand for data offloading. Efficient data transmission has been the main concern of these applications. Thus, data offloading within the limited communication environment has become a hotspot in both industry and academia. Since unmanned aerial vehicles (UAVs) can move across regions to make up for the communication gap caused by the loss of wireless communication networks, a lot of in-depth studies on UAV-enabled data offloading have been conducted. Nevertheless, few studies to date consider the uncertain user status and the heterogeneous UAV capabilities, which is however more practical and needs more attention. In this article, we propose an innovative framework to dynamically estimate user status information and determine the UAV scheduling strategy. On this basis, the heterogeneous UAV-enabled data offloading is modeled as a constrained multiobjective optimization problem, whose purpose is to lower the user data queue length while extending the working time of the UAV. Moreover, a differential evolution-based dynamic objective approximation method—RUDDER is proposed to solve the constrained multiobjective optimization problem. Through rigorous mathematical proof, we prove that RUDDER can consistently guide the population to approach the optimization solution with polynomial-level time complexity. To verify the effectiveness of the proposed RUDDER, extensive experiments are conducted to compare it with five comparison algorithms. The experimental results demonstrate the superiority of the RUDDER in terms of energy saving, time efficiency, and adaptability.

The choice of learning rate (LR) functions and policies has evolved from a simple fixed LR to the decaying LR and the cyclic LR, aiming to improve the accuracy and reduce the training time of Deep Neural Networks (DNNs). This article presents a systematic approach to selecting and composing an LR policy for effective DNN training to meet desired target accuracy and reduce training time within the pre-defined training iterations. It makes three original contributions. First, we develop an LR tuning mechanism for auto-verification of a given LR policy with respect to the desired accuracy goal under the pre-defined training time constraint. Second, we develop an LR policy recommendation system (LRBench) to select and compose good LR policies from the same and/or different LR functions through dynamic tuning, and avoid bad choices, for a given learning task, DNN model, and dataset. Third, we extend LRBench by supporting different DNN optimizers and show the significant mutual impact of different LR policies and different optimizers. Evaluated using popular benchmark datasets and different DNN models (LeNet, CNN3, ResNet), we show that our approach can effectively deliver high DNN test accuracy, outperform the existing recommended default LR policies, and reduce the DNN training time by 1.6-6.7× to meet a targeted model accuracy.

Deep learning shows great promise in providing more intelligence to the cross-platform web. However, insufficient infrastructure, heavy models, and intensive computation limit the use of deep learning with low-performing web browsers. We propose DeepAdapter, an integrated cloud-edge-device framework that ties the edge, the remote cloud, with the device by cross-platform web technology for adaptive deep learning services towards lower latency, lower mobile energy, and higher system throughput. DeepAdapter consists of context-aware pruning, service updating, and online scheduling. First, the offline pruning module provides a context-aware pruning algorithm that incorporates the latency, the network condition, and the device's computing capability to fit various contexts. Second, the service updating module optimizes branch model cache on the edge for massive mobile users and updates the new model pruning requirements. Third, the online scheduling module matches optimal branch models for mobile users. Also, a two-stage DRL-based online scheduling method named DeepScheduler can handle high concurrent requests between edge centers and remote cloud by designing the reward prediction model. Extensive experiments show that DeepAdapter can decrease average latency by 1.33x, reduce average mobile energy by 1.4x, and improve system throughput by 2.1x with considerable accuracy.

This paper presents a holistic approach to gradient leakage resilient distributed Stochastic Gradient Descent (SGD).First, we analyze two types of strategies for privacy-enhanced federated learning: (i) gradient pruning with random selection or low-rank filtering and (ii) gradient perturbation with additive random noise or differential privacy noise. We analyze the inherent limitations of these approaches and their underlying impact on privacy guarantee, model accuracy, and attack resilience.Next, we present a gradient leakage resilient approach to securing distributed SGD in federated learning, with differential privacy controlled noise as the tool. Unlike conventional methods with the per-client federated noise injection and fixed noise parameter strategy, our approach keeps track of the trend of per-example gradient updates. It makes adaptive noise injection closely aligned throughout the federated model training.Finally, we provide an empirical privacy analysis on the privacy guarantee, model utility, and attack resilience of the proposed approach. Extensive evaluation using five benchmark datasets demonstrates that our gradient leakage resilient approach can outperform the state-of-the-art methods with competitive accuracy performance, strong differential privacy guarantee, and high resilience against gradient leakage attacks.

Metric learning as a service (MLaaS) represents one of the main learning streams to handle complex datasets in service computing research communities and industries. A common approach for dealing with high-dimensional and complex datasets is employing a feature embedding algorithm to compress data through dimension reduction while optimizing intra-class distance. To create generalizable MLaaS for high-performance artificial intelligence applications with high-dimensional Big Data, a robust and meaningful embedding space representation by efficiently optimizing both intra-class and inter-class relationships is required. We developed a novel MLaaS methodology that incorporates covariance to signify the direction of the linear relationship between data points in an embedding space. Our covariance-based feature embedding architecture introduces three different yet complementary mapping functions: inner-class mapping, intra-class with semi-inter-class mapping, and intra- and inter-class mapping. Unlike conventional metric learning, our covariance-embedding-enhanced approach is more expressive and explainable for computing similar or dissimilar measures and can capture positive, negative, or neutral relationships. Our MLaaS framework ensures efficient, composable, and extensible metric learning by supporting the selection of dimension reduction and data compression methods. Experiments conducted using various benchmark datasets demonstrate that the proposed model can obtain higher-quality, more separable, and more expressive embedding representations than existing models.

The emergence of edge computing and 5G networks has fueled the growth of mobile Web AR. Although efforts have been made to improve the edge system efficiency for Web AR applications, efficient edge-assisted mobile Web AR services remain technically challenging. This paper presents EARNet, a distributed edge system orchestration approach for mobile Web AR in 5G networks. The design of EARNet makes three novel contributions. First, EARNet manages the edge network dynamics with respect to user mobility and their Web AR service requests by employing landmarks and grid index based edge node localization mechanisms. Second, EARNet takes into account both request serving performance and offloading cost in managing workload balance and quality of service and leverages dynamic hash and max heap mechanisms for efficient Web AR service lookup and AR computations. Third, EARNet designs the service migration schemes by optimizing several performance factors, such as message efficiency, scheduling latency, request density and locality of mobile users and edge nodes, and accuracy of Web AR services after migration. Experimental evaluations are conducted using the real base station deployment data in the Melbourne Central Business District (CBD) area. The results shows the effectiveness of the EARNet edge orchestration approach compared to several baseline approaches.

In most computing services, imperfect event logs with missing events are generated for a variety of reasons. Because missing events in imperfect event logs adversely affect the results of process mining analysis, it is essential to handle them effectively. Most existing process mining studies focus on methodologies for generation of good process models, very few methodologies, in fact, have been developed to deal with missing events. To the best of our knowledge, there is a lack of high-performance methods for restoration of missing events in actual event log data. In this paper, we propose a new categorical event imputation method that can restore missing categorical events by learning the structural features between observed events in the event log. We evaluated the proposed method by way of comparative experiments with previous studies using six real datasets, and the results demonstrate that the restoration performance was greatly improved and that thereby, our proposed method can significantly improve both the quality of event logs (specifically by restoring missing events in imperfect event logs) and the overall quality of process mining analysis.

Most cloud providers offer Deep Learning as a Service (DLaaS) for different business, science and engineering domains. However, it is known that deep neural networks (DNNs) are vulnerable to adversarial examples, which can cause well-trained DNN models to misbehave by injecting human-imperceptible perturbations to the query input data. Securing deep learning as a service becomes a critical challenge in mitigating such adversarial input perturbations, and enhancing the robustness of DNNs. In this article, we report two important facts: First, most adversarial perturbations are high frequency signals or are added to high frequency signals. Second, due to Frequency Principle that neural networks overly pay attention to fit the low frequency signals during training, the models could be easily misled by the high frequency signals of adversarial examples. These facts consequently contribute to the vulnerability of DNNs service in the Cloud. We conjecture that the more robust the neural networks are in learning from high frequency signals, the more resilient these neural networks are against adversarial perturbed examples. We propose a novel method for generating high-frequency-enhanced adversarial examples, which is achieved by a high-pass filter in the frequency domain via Fourier Transform. This method enhances the learning ability for high frequency signals and ameliorates to over-fit useless low frequency signals. In order to improve the robustness of DNNs service under such signal frequency attacks, we propose a multi-modal collaborative adversarial training framework, named as MMCAT, which uses the multi-modal information of the input images for cross-modal collaborative training, delivering excellent extension for effectively learning of multi-modal image information. Extensive experiments show that under strong adaptive frequency attacks, the DNNs service trained with the proposed MMCAT method achieve superior performance and robustness over the state-of-the-art adversarial training approaches.

Point-of-interest (POI) recommendation service has drawn growing attention with the widespread popularity of location- based social networks (LBSNs). Recent research methods on POI recommendation based on graph embedding have mainly focused on explicit interactions of LBSN objects such as user's check-ins on POIs and social relationships, while neglecting implicit relationship that cannot be directly observed but may notably contribute to the POI recommendation. This paper presents VirHpoi, a heterogeneous hypergraph embedding method for POI recommendation in LBSNs with three original contributions. First, we model the LBSNs as a hypergraph to capture the complex interactions in LBSNs and learn the hypergraph by preserving homophily and interaction attribute affinity of the LBSNs. Second, we introduce the notion of “virtual hyperedges” to capture the intrinsic correlations of POIs. Virtual hyperedges incorporate implicit yet informative connections of the check-in patterns in LBSNs in terms of geographical and semantic characteristics. Third, we propose techniques to learn heterogenous hypergraph embedding on the complex LBSN graph with both homogenous edges and heterogenous hyperedges with dual objectives: we aim to preserve the homophily of objects intra domain by maximizing the co-occurrence probability of all homogenous edges, and we want to learn the interaction attribute affinity across domains by maximizing the probability of predicting the target object in the hyperedges. As a result, our approach can preserve both the intra domain homophily of objects and the interaction attribute affinity across domains by learning low-dimensional embeddings of LBSN objects and then make more effective recommendations based on the embeddings. Extensive experiments on four real-world datasets show the effectiveness and superiority of VirHpoi compared with the state-of-the-art methods.

Efficiently managing database instances over cloud-scale clusters is significant for increasing service quality and reducing operational cost, especially confronting the growing cluster size and heterogeneous application services. Alibaba Cloud provides a large-scale Relational Database Service (RDS) for millions of users including enterprises from start-ups to large international corporations. To manage tremendous amount of RDS instances in the Cloud with the goal of reducing cost while guaranteeing service level agreement(SLA), YISHAN, an intelligent database instance management system, is designed to dynamically manage the placement of instances using machine learning techniques. YISHAN collects historical performance data to analyze patterns of the resource utilization of instances and hosts. By learning “good packings” in which instances colocate harmoniously, YISHAN is able to optimize the instances placement to provide better quality of service and improve the efficiency of CPU, memory, and disk resources. We deploy and run YISHAN in Alibaba Cloud RDS. The running logs show that YISHAN successfully saves 17% of the resources in hosts and efficiently reduces the burdens and crash risks of RDS instances.

Location-based queries enable fundamental services for mobile users. While the benefits of location-based services (LBS) are numerous, exposure of mobile users’ locations to untrusted LBS providers may lead to privacy concerns. This article proposesStarCloak, a utility-aware and attack-resilient location anonymization service for privacy-preserving LBS usage.StarCloakcombines several desirable properties. First, unlike conventional approaches which are indifferent to underlying road network structure,StarCloakuses the concept ofstars and proposescloaking graphsfor effective location cloaking on road networks. Second,StarCloaksupports user-specified$k$-user anonymity and$l$-segment indistinguishability, for enabling personalized privacy protection and for serving users with varying privacy preferences. Third,StarCloakachieves strong attack-resilience against replay and query injection attacks through randomized star selection and pruning. Finally, to enable efficient query processing with high throughput and low bandwidth overhead,StarCloakmakes cost-aware star selection decisions by considering query evaluation and network communication costs. We evaluateStarCloakon two datasets using real-world road networks, under various privacy and utility constraints. Results show thatStarCloakachieves improved query success rate and throughput, reduced anonymization time and network usage, and higher attack-resilience in comparison toXStar, its most relevant competitor.

Neural network ensemble is a collaborative learning paradigm that utilizes multiple neural networks to solve a complex learning problem. Constructing predictive models with high generalization performance is an important and yet most challenging goal for robust intelligence systems in the presence of dirty data. Given a target learning task, popular approaches have been dedicated to find the top performing model. However, it is difficult in general to estimate the best model when available data is finite, possibly dirty, and insufficient for the problem. In this keynote, I will give an overview of a diversity-centric ensemble learning framework developed at Georgia Tech, including methodologies and algorithms for measuring, enforcing, and combining multiple neural networks by improving generalization performance of the overall system and maximizing ensemble utility and resilience to dirty data.

Interactive microservices expose API endpoints to be invoked by users. For such applications, precisely estimating the resources required to serve specific API traffic is challenging. This is because an API request can interact with different components and consume different resources for each component. The notion of API traffic is vital to application owners since the API endpoints often reflect business logic, e.g., a customer transaction. The existing systems that simply rely on historical resource utilization are not API-aware and thus cannot estimate the resource requirement accurately. This paper presents DeepRest, a deep learning-driven resource estimation system. DeepRest formulates resource estimation as a function of API traffic and learns the causality between user interactions and resource utilization directly in a production environment. Our evaluation shows that DeepRest can estimate resource requirements with over 90% accuracy, even if the API traffic to be estimated has never been observed (e.g., 3× more users than ever or unseen traffic shape). We further apply resource estimation for application sanity checks. DeepRest identifies system anomalies by verifying whether the resource utilization is justifiable by how the application is being used. It can successfully identify two major cyber threats: ransomware and cryptojacking attacks.

Popularity prediction is to predict the number of social network users involved in information diffusion. Recently, deep learning methods for popularity prediction advance traditional approaches that rely on hand-crafted features. However, existing approaches ignore the multi-source cascade that consists of multiple sub-cascades with different content but under the same topic. Different from single-source cascade, more cascading information can be observed from multi-source cascade and they are potentially correlated. How to correlate the diverse information and take advantage of them from both temporal and spatial aspects is critical for prediction. To this end, we propose a novel framework, called HEterogeneous Recurrent Integrated Graph Convolutional Neural Network (HERI-GCN). Specifically, we construct a heterogeneous cascade graph to model the multi-source cascade where time intervals are treated as heterogeneous time nodes. Besides, we propose a heterogeneous GCN to learn rich features from the multi-source cascade. RNN is organically integrated into the heterogeneous GCN to overcome the limited learning ability toward temporal and spatial data. We evaluate HERI-GCN through comparative experiments on three datasets. The experimental evaluation shows that HERI-GCN outperforms the state-of-the-art baseline methods.

Object detection has played a pivotal role in numerous mission-critical applications. This paper presents a focal error diversity framework, called EDI, for strengthening the robustness of object detection ensembles under benign and adversarial scenarios. We introduce an ensemble pruning method for object detection using a novel focal error diversity measure as the robustness synergy indicator. Given a base model pool, it recommends top sub-ensembles with a smaller ensemble size yet achieving equivalent or even better mAP performance than using all available object detection models as a large ensemble. This is made possible by our negative sampling methods for object detection to capture the degree of negative correlations and the focal error diversity to measure the failure independence of component detection models in an ensemble. Extensive experiments on three object detection benchmark datasets validate that EDI effectively selects space-time efficient object detection ensembles with high mAP performance.

This paper presents a novel approach to the problem of event-dense text and image cross-modal retrieval where the text contains the descriptions of numerous events. It is known that modality alignment is crucial for retrieval performance. However, due to the lack of event sequence information in the image, it is challenging to perform the fine-grain alignment of the event-dense text with the image. Our proposed approach incorporates the event-oriented features to enhance the cross-modal alignment, and applies the event-dense text-image retrieval to the food domain for empirical validation. Specifically, we capture the significance of each event by Transformer, and combine it with the identified key event elements, to enhance the discriminative ability of the learned text embedding that summarizes all the events. Next, we produce the image embedding by combining the event tag jointly shared by the text and image with the visual embedding of the event-related image regions, which describes the eventual consequence of all the events and facilitates the event-based cross-modal alignment. Finally, we integrate text embedding and image embedding with the loss optimization empowered with the event tag by iteratively regulating the joint embedding learning for cross-modal retrieval. Extensive experiments demonstrate that our proposed event-oriented modality alignment approach significantly outperforms the state-of-the-art approach with a 23.3% improvement on top-1 Recall for image-to-recipe retrieval on Recipe1M 10k test set.

Spectral clustering (SC) is an unsupervised learning method that has been widely used in industrial product analysis. Compressive spectral clustering (CSC) effectively accelerates clustering by leveraging graph filter and random sampling techniques. However, CSC suffers from two major problems. First, the direct use of the dichotomy and eigencount techniques for estimating Laplacian matrix's$k$kth eigenvalue is expensive. Second, the computation of interpolation is time-consuming because it requires to repeat matrix-vector product for every cluster in each iteration. To address these problems, we propose a new method calledfast compressive spectral clustering(FCSC). Our method addresses the first problem by assuming that the eigenvalues approximately satisfy local uniform distribution, and addresses the second problem by recalculating the pairwise similarity between nodes with low-dimensional representation to reconstruct denoised laplacian matrix. The time complexity of reconstruction is linear with the number of non-zeros in Laplacian matrix. As experimentally demonstrated on both artificial and real-world datasets, our method significantly reduces the computation time while preserving high clustering accuracy comparable to previous designs, demonstrating the effectiveness of FCSC.

Multi-user mobile Augmented Reality (AR) has been successfully used in various fields as a novel visual interaction technology. But current mainstream wearable device-based and app-based solutions are still facing cross-platform, real-time communication, and intensive computing requirements. Mobile Web technology is envisioned to be a promising supporting technology for cross-platform application of mobile AR especially in 5G networks, which provide pervasive communication and computing resources thereby forming a formidable framework for the practical application of multi-user mobile Web AR. However, the problem of how to use these new techniques properly to achieve efficient communication and computing collaboration is obviously paramount in order for multi-user mobile Web AR to be realized in 5G networks. In this article, we propose the first edge-assisted multi-user collaborative framework for mobile Web AR in the 5G era. First, we propose a heuristic mechanism BA-CPP for efficient communication planning, which allows multi-user interaction synchronization to be achieved. Second, we introduce a motion-aware key frame selection mechanism called Mo-KFP to optimize the computational efficiency of the edge system, and simultaneously alleviate the initialization problem by collaborating with nearby mobile devices using the Device-to-Device (D2D) communication technique. Experiments are conducted in a real-world 5G network, and the results demonstrate the superiority of our proposed collaborative framework.

Local Differential Privacy (LDP) is a popular standard for privacy-preserving data collection. Numerous LDP protocols have been proposed in the literature which differ in how they provide higher utility in different settings. Yet, few have engaged in analyzing the privacy relationships of these protocols under varying settings, and consequently, it is non-trivial to select which LDP protocol is best to use in a newly emerging application. In this paper, we present an adversarial approach to protocol analysis and selection and make three original contributions. First, we introduce a Bayesian adversary to analyze the privacy relationships of LDP protocols under varying settings. We show that different protocols have substantially different responses to the attack effectiveness of the Bayesian adversary, measured in terms of Adversarial Success Rate (ASR). Second, we provide a formal and empirical analysis on a set of privacy and utility-critical factors, including encoding parameters, privacy budget, data domain, adversarial knowledge, and statistical distribution. We show that different settings of these factors have significant effects on the ASRs of LDP protocols, and no protocol provides consistently low ASR across all settings. Third, we design and develop LDPLens, a prototype implementation of our proposed framework. Given a data collection scenario with various factors and constraints, LDPLens enables optimized selection of a desirable LDP protocol for the given scenario. We evaluate the effectiveness of LDPLens using three case studies with real-world datasets. Results show that LDPLens recommends a different protocol in each case study, and the protocol recommended by LDPLens can yield up to 1.5–2 fold reduction in utility loss, ASR or privacy budget compared to a randomly selected protocol.

Gradient leakage attacks are considered one of the wickedest privacy threats in deep learning as attackers covertly spy gradient updates during iterative training without compromising model training quality, and yet secretly reconstruct sensitive training data using leaked gradients with high attack success rate. Although deep learning with differential privacy is a defacto standard for publishing deep learning models with differential privacy guarantee, we show that differentially private algorithms with fixed privacy parameters are vulnerable against gradient leakage attacks. This paper investigates alternative approaches to gradient leakage resilient deep learning with differential privacy (DP).First, we analyze existing implementation of deep learning with differential privacy, which use fixed noise variance to injects constant noise to the gradients in all layers using fixed privacy parameters. Despite the DP guarantee provided, the method suffers from low accuracy and is vulnerable to gradient leakage attacks.Second, we present a gradient leakage resilient deep learning approach with differential privacy guarantee by using dynamic privacy parameters. Unlike fixed-parameter strategies that result in constant noise variance, different dynamic parameter strategies present alternative techniques to introduce adaptive noise variance and adaptive noise injection which are closely aligned to the trend of gradient updates during differentially private model training.Finally, we describe four complementary metrics to evaluate and compare alternative approaches. Extensive experiments on six benchmark datasets show that differentially private deep learning with dynamic privacy parameters outperforms the deep learning using fixed DP parameters, and existing adaptive clipping approaches in all aspects: compelling accuracy performance, strong differential privacy guarantee, and high attack resilience.

Enabling deep learning technology on the mobile web can improve the user’s experience for achieving web artificial intelligence in various fields. However, heavy DNN models and limited computing resources of the mobile web are now unable to support executing computationally intensive DNNs when deploying in a cloud computing platform. With the help of promising edge computing, we propose a lightweight collaborative deep neural network for the mobile web, named LcDNN, which contributes to three aspects: (1) We design a composite collaborative DNN that reduces the model size, accelerates inference, and reduces mobile energy cost by executing a lightweight binary neural network (BNN) branch on the mobile web. (2) We provide a jointly training method for LcDNN and implement an energy-efficient inference library for executing the BNN branch on the mobile web. (3) To further promote the resource utilization of the edge cloud, we develop a DRL-based online scheduling scheme to obtain an optimal allocation for LcDNN. The experimental results show that LcDNN outperforms existing approaches for reducing the model size by about 16x to 29x. It also reduces the end-to-end latency and mobile energy cost with acceptable accuracy and improves the throughput and resource utilization of the edge cloud.

This article introduces a two-phase deep feature engineering framework for efficient learning of semantics enhanced joint embedding, which clearly separates the deep feature engineering in data preprocessing from training the text-image joint embedding model. We use the Recipe1M dataset for the technical description and empirical validation. In preprocessing, we perform deep feature engineering by combining deep feature engineering with semantic context features derived from raw text-image input data. We leverage LSTM to identify key terms, deep NLP models from the BERT family, TextRank, or TF-IDF to produce ranking scores for key terms before generating the vector representation for each key term by using Word2vec. We leverage Wide ResNet50 and Word2vec to extract and encode the image category semantics of food images to help semantic alignment of the learned recipe and image embeddings in the joint latent space. In joint embedding learning, we perform deep feature engineering by optimizing the batch-hard triplet loss function with soft-margin and double negative sampling, taking into account also the category-based alignment loss and discriminator-based alignment loss. Extensive experiments demonstrate that our SEJE approach with deep feature engineering significantly outperforms the state-of-the-art approaches.

Currently, cloud computing has received great attention in commerce and scientific research due to its flexibility and strong data processing capability. However, in view of the fact that the types of tasks display an upward trend as the growth of service demands, and the different types of tasks arrive at the system without regularity. Moreover, the resources deployed in cloud are insufficiency to be flexibly provisioned in the face of obvious workload fluctuations. In this article, we present a method of elastic resource provisioning using date clustering in cloud service platform. The framework of proposed method consists of three core components: tasks clustering, the amount of tasks prediction in cluster, dynamic resource provisioning and scheduling. In workload classification, we propose a clustering ensemble method, which utilizes a novel distance decision-making method to obtain the final results. Our method can effectively partition the arriving tasks into several clusters based on similarity among tasks. For each cluster, we forecast the amount of tasks arriving at next moment by prediction model based on time-series to provide reference for the follow-up resource provisioning. Afterwards, an energy-saving resource provisioning method is designed to dynamically provide resources for tasks in each cluster to meet their performance requirements. We implement the experiments in Google cloud traces dataset and the results show that our method achieves 92.3, 91.2 percent, and 3679.2 kW$ \cdot$·h respectively in terms of guarantee ratio, resource utilization and total energy consumption, which demonstrates the effectiveness of proposed method for dynamic resource provisioning.

Fueled by the advances in the Internet of Things, and the growing capacity of smart mobile devices at the edge of the Internet, we have witnessed a growing trend in research and development for edge computing and edge storage, which extends the abilities of single mobile device on the edge through on-demand collaboration among multiple geographically distributed mobile devices. In this article, we address several technical challenges that are unique for collaborative storage at the edge due to the unique characteristics of mobile devices. First, we formalize the collaborative storage problem as an optimization problem. Second, we design an Acceleration Algorithm for Collaborative Storage, called A2CS, based on the architecture of Alternating Direction Method of Multipliers (ADMM). Specifically, we use the Nesterov’s Acceleration strategy and the step size rules in the process of updating variables and determining the optimal speed of convergence. We develop a novel collaborative storage policy in order to guide the whole lifecycle of collaborative storage. Finally, we conduct a series of experiments for acceleration performance analysis and validation. We show that A2CS delivers a better convergence performance with different step size rules, compared with two existing approaches: the ADMM baseline and the ADMM-OR (ADMM with Over-Relaxation), achieving the acceleration percentage by at least 25.33 percent and at most 64.01 percent. In addition, by conducting the utility performance comparison analysis with the existing Average Distribution Strategy (ADS) and the existing Distance Preferred Distribution Strategy (DPDS), we show the advantage of A2CS over both ADS and DPDS with respect to the total utility and energy consumption.

Numerous research efforts have been proposed for efficient processing of range queries in high-dimensional space by either redesigning R-tree access structure for exploring massive parallelism on single GPU or exploring distributed framework of R-tree. However, none of the existing efforts explores the integration of the parallelization of the R-tree on a single GPU with a distributed framework for the R-tree. The problem of designing an efficient multi-GPU indexing method, which can effectively combine the parallelism maximization with distributed processing of the R-tree, remains an open challenge. In this article, we present a novel multi-GPU efficient parallel and distributed indexing method, called LBPG-tree. The rationale of the LBPG-tree is to combine the advantages of an instruction pipeline in CPU with the massive parallel processing potential of multiple GPUs by introducing two new optimization strategies: First, we exploit the GPU L2 cache for accelerating both index search and index node access on GPUs. Second, we further improve utilization of L2 cache on GPUs by compacting and sorting candidate nodes called Compact-and-Sort. Our experimental results show that the LBPG-tree outperforms G-tree, the previous representative GPU index method and effectively support multiple GPUs for providing efficient high dimensional range query service.

Web-based Deep Neural Networks (DNNs) enhance the ability of object recognition and has attracted considerable attention in mobile Web AR and other services. However, neither performing the DNN inference on mobile Web browsers locally nor offloading computations to the cloud can strike a balance between accuracy and efficiency; generally, rude methods are often accompanied by unsatisfactory accuracy. Collaborative approaches seem to fill this gap by coordinating the distributed hierarchical computing resources, especially in the 5G era, but it still faces challenges in the current solutions, such as the lack of (1) full use of 5G resources for the one point DNN computation partitioning schemes; (2) fine-grained branching mechanism; (3) efficient partitioning method; and (4) multi-objective optimization. To this end, we present the fine-grained elastic computation partitioning mechanism for distributed DNN in 5G networks. First, we elaborate two collaborative scenarios. Second, we study the DNN branching mechanism at layer granularity. Next, we propose a DNN computation partitioning algorithm based on deep reinforcement learning. Finally, we develop a mobile Web AR application as a proof of concept. The experiments were conducted in an actually deployed 5G trial network, and the results show the superiority of this collaborative approach. The common theme is, under the premise that Quality of Service (QoS) is satisfied, to balance multiple interests by orchestrating computations across heterogeneous computing platforms.

Research grants have played an important role in seeding and promoting fundamental research projects worldwide. There is a growing demand for developing and delivering scientific influence analysis as a service on research grant repositories. Such analysis can provide insight on how research grants help foster new research collaborations, encourage cross-organization collaborations, influence new research trends, and identify technical leadership. This article presents the design and development of a grant-based scientific influence analysis service, coined asGImpact. It takes a graph-theoretic approach to design and develop the scientific influence analysis algorithms over a real research-grant repository with three original contributions. First, we model the scientific influence analysis problem as a graph-based analysis problem by constructing heterogeneous graphs from the grants dataset, including mining the dataset to identify and extract important features and represent such features as a research grants information network. Second, we develop the scientific influence analysis algorithms over the research grants information network, which compute the overall scientific influence score by integrating self-influence score and multiple co-influence scores. The self-influence score reflects the grant-based research collaborations among institutions, and the co-influence scores reflect various types of cross-institution collaborations in terms of disciplines and keywords (subject areas). Third, we leverage the cluster analysis on the institution graph as an example application of scientific influence analysis service. By partitioning the institution graph into$K$clusters, with$K$as one of the service interface parameters, we show how different disciplines and different keywords are co-related through the grant-based influence analysis. We evaluateGImpactusing a real grants dataset, consisting of 2512 institutions and their grants received over a period of 14 years. Our experimental results show that theGImpactinfluence analysis approach can effectively identify the grant-based research collaboration groups and provide valuable insight on an in-depth understanding of the scientific influence of research grants on research programs, institution leadership, and future collaboration opportunities.

Software bugs remain frequent in the life cycle of software development and maintenance. Automatic localization of buggy source code files is critical for timely bug fixing and improving the efficiency of software quality assurance. Various bug localization techniques have been proposed using different dimensions of features. Recent studies have shown that different dimensions of features may play different roles in bug localization. Unfortunately, how to effectively merge these dimensions of features for improving bug localization has rarely been investigated. This article presents a Multi-Dimension Convolutional Neural Network (MD-CNN) model for bug localization automatically based on a bug report. Our approach has dual-novelty. First, we identify and extract five statistical dimensions of features. Second, we design a Convolutional Neural Network (CNN) model that takes our five statistical dimensions of features as the input and iteratively learns the complex and non-linear relationship between the features and the bug locations. The MD-CNN bug localization model is verified using six large-scale open source projects. The experimental results show that our MD-CNN outperforms the existing representative bug localization techniques in terms of the Mean Average Precision (MAP) and the number of bugs successfully localized in the top 1, 5, and 10 matched source code files.

Big data powered Deep Learning (DL) and its applications have blossomed in recent years, fueled by three technological trends: a large amount of digitized data openly accessible, a growing number of DL software frameworks in open source and commercial markets, and a selection of affordable parallel computing hardware devices. However, no single DL framework, to date, dominates in terms of performance and accuracy even for baseline classification tasks on standard datasets, making the selection of a DL framework an overwhelming task. This paper takes a holistic approach to conduct empirical comparison and analysis of four representative DL frameworks with three unique contributions.First, given a selection of CPU-GPU configurations, we show that for a specific DL framework, different configurations of its hyper-parameters may have a significant impact on both performance and accuracy of DL applications.Second, to the best of our knowledge, this study is the first to identify the opportunities for improving the training time performance and the accuracy of DL frameworks by configuring parallel computing libraries and tuning individual and multiple hyper-parameters.Third, we also conduct a comparative measurement study on the resource consumption patterns of four DL frameworks and their performance and accuracy implications, including CPU and memory usage, and their correlations to varying settings of hyper-parameters under different configuration combinations of hardware, parallel computing libraries. We argue that this measurement study provides in-depth empirical comparison and analysis of four representative DL frameworks, and offers practical guidance for service providers to deploying and delivering DL as a Service (DLaaS) and for application developers and DLaaS consumers to select the right DL frameworks for the right DL workloads.

It is widely acknowledged that learning joint embeddings of recipes with images is challenging due to the diverse composition and deformation of ingredients in cooking procedures. We present a Multi-modal Semantics enhanced Joint Embedding approach (MSJE) for learning a common feature space between the two modalities (text and image), with the ultimate goal of providing high-performance cross-modal retrieval services. Our MSJE approach has three unique features. First, we extract the TFIDF feature from the title, ingredients and cooking instructions of recipes. By determining the significance of word sequences through combining LSTM learned features with their TFIDF features, we encode a recipe into a TFIDF weighted vector for capturing significant key terms and how such key terms are used in the corresponding cooking instructions. Second, we combine the recipe TFIDF feature with the recipe sequence feature extracted through two-stage LSTM networks, which is effective in capturing the unique relationship between a recipe and its associated image(s). Third, we further incorporate TFIDF enhanced category semantics to improve the mapping of image modality and to regulate the similarity loss function during the iterative learning of cross-modal joint embedding. Experiments on the benchmark dataset Recipe1M show the proposed approach outperforms the state-of-the-art approaches.

Reachability queries ask whether a vertex can reach another vertex on large directed graphs. It is one of the most fundamental graph operators and has attracted researchers in both academics and industry to study it. The main technical challenge is to support fast reachability queries by efficient managing the three main costs: the index construction time, the index size and the query processing time on large/small and sparse/dense graphs. As real world graphs grow bigger in size, these problems remain open challenges that demand high performance solutions. In this article, we propose aMulti-Dimensional Graph Labelingapproach (calledMGTag) to supporting fast reachability queries.MGTagis novel in three aspects. First, it recursively partitions a graph into multiple subgraphs with disjoint vertex sets, called non-shared graphs, and several inter-partition edges, called cross-edges. Second, we build a four-dimensional label – one dimension of layer, one dimension of non-shared graph and two dimensions of interval for each vertex in non-shared graphs. Finally, with the four-dimensional labeling scheme, we design algorithms to answer reachability queries efficiently. The extensive experiments on 28 large/small and dense/sparse graphs show that building the high dimensional index is quickly and the index size is also competitive compared with most of the state-of-the-art approaches. The results also show that our approach is more scalable and efficient than the state-of-the-art approaches in answering reachability queries on large/small and sparse/dense graphs.

Healthcare blockchains provide an innovative way to store healthcare information, execute healthcare transactions, and build trust for healthcare data sharing and data integration in a decentralized open healthcare network environment. Although the healthcare blockchain technology has attracted broad interests and attention in industry, government and academia, the security and privacy concerns remain the focus of debate when deploying blockchains for information sharing in the healthcare sector from business operation to research collaboration. This article focuses on the security and privacy requirements for medical data sharing using blockchain, and provides a comprehensive analysis of the security and privacy risks and requirements, accompanied by technical solution techniques and strategies. First, we discuss the security and privacy requirements and attributes required for electronic medical data sharing by deploying the healthcare blockchain. Second, we categorize existing efforts into three reference blockchain usage scenarios for electronic medical data sharing, and discuss the technologies for implementing these security and privacy properties in the three categories of usage scenarios for healthcare blockchain, such as anonymous signatures, attribute-based encryption, zero-knowledge proofs, verification techniques for smart contract security. Finally, we discuss other potential blockchain application scenarios in healthcare sector. We conjecture that this survey will help healthcare professionals, decision makers, and healthcare service developers to gain technical and intuitive insights into the security and privacy of healthcare blockchains in terms of concepts, risks, requirements, development and deployment technologies and systems.

Real-world networks, such as phone-call networks and social networks, are often not static but temporal. Mining cohesive subgraphs from static graphs is a fundamental task in network analysis and has been widely investigated in the past decades. However, the concepts of cohesive subgraphs shift from static to temporal graphs raise many important problems. For instance, how to detect stable cohesive subgraphs on temporal networks such that the nodes in the subgraph are densely and stably connected over time. To address this problem, we resort to the conventional quasi-clique and propose a new model, called maximal$\rho $-stable ($\delta, \gamma $)-quasi-clique, to capture both the cohesiveness and the stability of a subgraph. We show that the problem of enumerating all maximal$\rho $-stable ($\delta, \gamma $)-quasi-cliques is NP-hard. To efficiently tackle our problem, we first devise a novel temporal graph reduction algorithm to significantly reduce the temporal graph without losing any maximal$\rho $-stable ($\delta, \gamma $)-quasi-clique. Then, on the reduced temporal graph, we propose an effective branch and bound enumeration algorithm, namedBB&SCM, with four carefully designed pruning techniques to accomplish the enumeration process. Finally, we conduct extensive experiments on seven real-world temporal graphs, and the results demonstrate that the temporal graph reduction algorithm can safely reduce 98% nodes of the temporal graph (with millions of nodes and edges) andBB&SCMis at least two orders of magnitude faster than the baseline algorithms. Moreover, we also evaluate the effectiveness of our model against other baseline models.

In this paper, we introduce DLB, a Deep Learning based load Balancing mechanism, to effectively address the data skew problem. The key idea of DLB is to replace hash functions in the load balancing mechanisms with deep learning models, which are trained to be able to map different distributions of workloads and data to the servers in a uniformed manner. We implemented DLB and deployed it on a practical Cloud environment using CloudSim. Experimental results using both synthetic and real-world data sets show that compared with traditional hash function based load balancing methods, DLB is able to achieve more balanced mappings, especially when the workload is highly skewed.

This paper presents a transparent Container Network Memory storage device, coined as CNetMem, aiming to address the open problem of unpredictable performance degradation of containers when the working set of an application no longer fits in container memory. First, CNetMem will enable application tenants running in a container to park their working set memory/file to a faster network memory storage by organizing a group of remote memory nodes as remote memory donors. This allows CNetMem to take advantage of remote idle memory on a cluster before resorting to a slow local I/O subsystem like local disk without any modification of host OS or application. Second, CNetMem provides a hybrid batching technique to remove or alleviate performance bottlenecks in the I/O performance critical path for remote memory read/write with replication or disk backup for fault tolerance. Third, CNetMem introduces a rank-based node selection algorithm to find the optimal node for placing remote memory blocks across cluster. This helps CNetMem to reduce the performance impact due to remote memory eviction. Extensive experiments are conducted on three big data applications and four machine learning workloads. The results show that CNetMem achieves up to 172× throughput improvements compared to vanilla Linux and up to 5.9× completion time improvements over existing approaches in big data and ML workload.

An efficient approach to managing big data workloads is to enable applications to work directly with huge pages. This can effectively avoid or reduce the memory fragmentation problem due to high frequent memory allocation and deallocation and significantly minimize the performance degradation of big data applications. This paper presents XPage, a huge page memory management framework, with three novel features. First, XPage by design can provide automated huge page managements with transparency to both OS and applications. Second, Xpage represents a memory management redesign that brings performance and memory saving to memory intensive applications by supporting dynamic huge page memory management without resorting to splitting huge pages for memory fragmentation. Third but not the least, XPage can efficiently minimize the internal fragmentation without impacting performance of applications. We conduct extensive experiments to evaluate the effectiveness of XPage in minimizing internal memory fragmentation in the presence of dynamic memory intensive big data workloads, by comparing XPage with vanilla Linux using 4KB base page and Linux with 2MB huge page.

This paper presents a three-tier modality alignment approach to learning text-image joint embedding, coined as JEMA, for cross-modal retrieval of cooking recipes and food images. The first tier improves recipe text embedding by optimizing the LSTM networks with term extraction and ranking enhanced sequence patterns, and optimizes the image embedding by combining the ResNeXt-101 image encoder with the category embedding using wideResNet-50 with word2vec. The second tier modality alignment optimizes the textual-visual joint embedding loss function using a double batch-hard triplet loss with soft-margin optimization. The third modality alignment incorporates two types of cross-modality alignments as the auxiliary loss regularizations to further reduce the alignment errors in the joint learning of the two modality-specific embedding functions. The category-based cross-modal alignment aims to align the image category with the recipe category as a loss regularization to the joint embedding. The cross-modal discriminator-based alignment aims to add the visual-textual embedding distribution alignment to further regularize the joint embedding loss. Extensive experiments with the one-million recipes benchmark dataset Recipe1M demonstrate that the proposed JEMA approach outperforms the state-of-the-art cross-modal embedding methods for both image-to-recipe and recipe-to-image retrievals.

Neural network ensembles are gaining popularity by harnessing the complementary wisdom of multiple base models. Ensemble teams with high diversity promote high failure independence, which is effective for boosting the overall ensemble accuracy. This paper provides an in-depth study on how to design and compute ensemble diversity, which can capture the complementary decision capacity of ensemble member models. We make three original contri-butions. First, we revisit the ensemble diversity metrics in the literature and analyze the inherent problems of poor correlation between ensemble diversity and ensemble ac-curacy, which leads to the low quality ensemble selection using such diversity metrics. Second, instead of computing diversity scores for ensemble teams of different sizes using the same criteria, we introduce focal model based ensemble diversity metrics, coined as FQ-diversity metrics. Our new metrics significantly improve the intrinsic correlation between high ensemble diversity and high ensemble accuracy. Third, we introduce a diversity fusion method, coined as the EQ-diversity metric, by integrating the top three most representative FQ-diversity metrics. Comprehensive experiments on two benchmark datasets (CIFAR-10 and ImageNet) show that our FQ and EQ diversity metrics are effective for selecting high diversity ensemble teams to boost overall ensemble accuracy.

Federated learning(FL) is an emerging distributed learning paradigm with default client privacy because clients can keep sensitive data on their devices and only share local training parameter updates with the federated server. However, recent studies reveal that gradient leakages in FL may compromise the privacy of client training data. This paper presents a gradient leakage resilient approach to privacy-preserving federated learning with per training example-based client differential privacy, coined as Fed-CDP. It makes three original contributions. First, we identify three types of client gradient leakage threats in federated learning even with encrypted client-server communications. We articulate when and why the conventional server coordinated differential privacy approach, coined as Fed-SDP, is insufficient to protect the privacy of the training data. Second, we introduce Fed-CDP, the per example-based client differential privacy algorithm, and provide a formal analysis of Fed-CDP with the (∊,δ) differential privacy guarantee, and a formal comparison between Fed-CDP and Fed-SDP in terms of privacy accounting. Third, we formally analyze the privacy-utility tradeoff for providing differential privacy guarantee by Fed-CDP and present a dynamic decay noise-injection policy to further improve the accuracy and resiliency of Fed-CDP. We evaluate and compare Fed-CDP and Fed-CDP(decay) with Fed-SDP in terms of differential privacy guarantee and gradient leakage resilience over five benchmark datasets. The results show that the Fed-CDP approach outperforms conventional Fed-SDP in terms of resilience to client gradient leakages while offering competitive accuracy performance in federated learning.

Modern distributed machine learning (ML) systems leverage large-scale computing infrastructures to achieve fast model training. For many servers jointly training a model, failure recovery becomes an important challenge when a training task could be accomplished in minutes rather than days. The state-of-the-art checkpointing mechanism cannot meet the need of efficient recovery for large-scale ML, because its high cost prevents timely checkpointing and a server failure will likely cause a substantial loss of intermediate results when the checkpointing intervals are comparable to the entire training times. This paper proposes FreeLauncher (FLR), a lossless recovery mechanism for large-scale ML which performs ultralight replication (instead of checkpointing) to guarantee all intermediate training results (parameters) to be timely replicated. Our key insight is that in the parameter-server (PS) architecture there already exist multiple copies for each intermediate result not only in the server but also in the workers, most of which are qualified for failure recovery. FLR addresses the challenges of parameter sparsity (e.g., when training LDA) and staleness (e.g., when adopting relaxed consistency) by selectively replicating the latest copies of the sparse/stale parameters to ensure at least k up-to-date copies to be existent, which can handle any k-1 failures by re-launching the failed servers with recovered parameters from workers. We implement FLR on Tensorflow. Evaluation results show that FLR achieves lossless failure recovery (almost requiring no recomputation) at little cost.

Deep neural network ensembles have become attractive learning techniques with better generalizability over individual models. Some mission critical applications may require a large number of deep neural networks to achieve desirable accuracy and generalizability, making the ensemble execution costly with respect to runtime and space. This paper proposes a novel hierarchical ensemble pruning approach, which can effectively examine a given pool of M base models and identify smaller high quality deep ensembles of size $S(\ll M)$ with higher ensemble accuracy than the entire deep ensemble of all M models. Our hierarchical pruning approach, coined as HQ, combines three novel techniques. First, we show that the focal diversity metrics is innovative and can accurately capture the negative correlation among the member models of an ensemble, and the use of focal diversity metrics can boost ensemble accuracy. Second, we introduce a focal-diversity based hierarchical pruning algorithm to progressively identify low-cost ensembles with high ensemble diversity and accuracy. Third, we design a focal diversity consensus method to find smaller deep ensembles with low negative correlation. We demonstrate such ensembles offer high accuracy and high robustness while being more time and space efficient in ensemble decision making. Evaluated using two benchmark datasets, we show that the proposed focal diversity powered hierarchical pruning can find significantly smaller ensembles of deep neural network models while achieving the same or better classification generalizability.

This paper introduces a two-phase deep feature calibration framework for efficient learning of semantics enhanced text-image cross-modal joint embedding, which clearly separates the deep feature calibration in data preprocessing from training the joint embedding model. We use the Recipe1M dataset for the technical description and empirical validation. In preprocessing, we perform deep feature calibration by combining deep feature engineering with semantic context features derived from raw text-image input data. We leverage LSTM to identify key terms, NLP methods to produce ranking scores for key terms before generating the key term feature. We leverage wideResNet50 to extract and encode the image category semantics to help semantic alignment of the learned recipe and image embeddings in the joint latent space. In joint embedding learning, we perform deep feature calibration by optimizing the batch-hard triplet loss function with soft-margin and double negative sampling, also utilizing the category-based alignment loss and discriminator-based alignment loss. Extensive experiments demonstrate that our SEJE approach with the deep feature calibration significantly outperforms the state-of-the-art approaches.

Multi-device interaction has attracted a growing interest in both mobile communication industry and mobile computing research community as mobile devices enabled social media and social networking continue to blossom. However, due to the stringent low latency requirements and the complexity and intensity of computation, implementing efficient multi-device interaction for real-time video streaming analysis and sharing is still in its infancy. Unlike previous approaches that rely on high network bandwidth and high availability of cloud center with GPUs to support intensive computations for multi-device interaction and for improving the service experience, we propose MIRSA, a novel edge centric multi-device interaction framework with a lightweight end-to-end DNN for on-device visual odometry (VO) streaming analysis by leveraging edge computing optimizations with three main contributions. First, we design MIRSA to migrate computations from the cloud to the device side, reducing the high overhead for large transmission of video streaming while alleviating the server load of the cloud. Second, we design a lightweight VO network by utilizing temporal shift module to support on-device pose estimation. Third, we provide on-device resource-aware scheduling algorithm to optimize the task allocation. Extensive experiments show MIRSA provides real-time high quality pose estimation as an interactive service and outperforms baseline methods.

Deep neural network (DNN) based object detection has become an integral part of numerous cyber-physical systems, perceiving physical environments and responding proactively to real-time events. Recent studies reveal that well-trained multi-task learners like DNN-based object detectors perform poorly in the presence of deception. This paper presents FUSE, a deception-resilient detection fusion approach with three novel contributions. First, we develop diversity-enhanced fusion teaming mechanisms, including diversity-enhanced joint training algorithms, for producing high diversity fusion detectors. Second, we introduce a three-tier detection fusion framework and a graph partitioning algorithm to construct fusion-verified detection outputs through three mutually reinforcing components: objectness fusion, bounding box fusion, and classification fusion. Third but not least, we provide a formal analysis of robustness enhancement by FUSE-protected systems. Extensive experiments are conducted on eleven detectors from three families of detection algorithms on two benchmark datasets. We show that FUSE guarantees strong robustness in mitigating the state-of-the-art deception attacks, including adversarial patches - a form of physical attacks using confined visual distortion.

Continuous Data Protection (CDP) is becoming instrumental in recovering applications from crypto-ransomware attacks. It enables fine-grained recovery through journaling, allowing the applications (its volumes) to recover to any previous state. While zero data loss can be achieved during recovery with CDP, the timestamp of the desired restore point, i.e., the one just prior to the attack, needs to be provided to reconstruct the volume. Such information is often unavailable in practice, and system administrators can only adopt a trial-and-error strategy to narrow down the time range of desired restore points by making multiple time-consuming recovery attempts. The recovery systems offer little guidance in pointing to the restore points containing a valid application state and reducing data loss. To address this problem, we equip the CDP-based recovery with machine intelligence. This demonstration showcases Smart Recovery Advisor (SRA), which offers interpretable, data-driven, and feedback-aware restore point recommendations that reduce the number of recovery attempts while minimizing data loss.

Local Differential Privacy (LDP) is popularly used in practice for privacy-preserving data collection. Although existing LDP protocols offer high utility for large user populations (100,000 or more users), they perform poorly in scenarios with small user populations (such as those in the cybersecurity domain) and lack perturbation mechanisms that are effective for both ordinal and non-ordinal item sequences while protecting sequence length and content simultaneously. In this paper, we address the small user population problem by introducing the concept of Condensed Local Differential Privacy (CLDP) as a specialization of LDP, and develop a suite of CLDP protocols that offer desirable statistical utility while preserving privacy. Our protocols support different types of client data, ranging from ordinal data types in finite metric spaces (numeric malware infection statistics), to non-ordinal items (OS versions, transaction categories), and to sequences of ordinal and non-ordinal items. Extensive experiments are conducted on multiple datasets, including datasets that are an order of magnitude smaller than those used in existing approaches, which show that proposed CLDP protocols yield high utility. Furthermore, case studies with Symantec datasets demonstrate that our protocols accurately support key cybersecurity-focused tasks of detecting ransomware outbreaks, identifying targeted and vulnerable OSs, and inspecting suspicious activities on infected machines.

Deep neural network (DNN) models are known to be vulnerable to maliciously crafted adversarial examples and to out-of-distribution inputs drawn sufficiently far away from the training data. How to protect a machine learning model against deception of both types of destructive inputs remains an open challenge. This article presents XEnsemble, a diversity ensemble verification methodology for enhancing the adversarial robustness of DNN models against deception caused by either adversarial examples or out-of-distribution inputs. XEnsemble by design has three unique capabilities. First, XEnsemble builds diverse input denoising verifiers by leveraging different data cleaning techniques. Second, XEnsemble develops a disagreement-diversity ensemble learning methodology for guarding the output of the prediction model against deception. Third, XEnsemble provides a suite of algorithms to combine input verification and output verification to protect the DNN prediction models from both adversarial examples and out of distribution inputs. Evaluated using 11 popular adversarial attacks and two representative out-of-distribution datasets, we show that XEnsemble achieves a high defense success rate against adversarial examples and a high detection success rate against out-of-distribution data inputs, and outperforms existing representative defense methods with respect to robustness and defensibility.

Machine learning techniques have been widely applied to various applications. However, they are potentially vulnerable to data poisoning attacks, where sophisticated attackers can disrupt the learning procedure by injecting a fraction of malicious samples into the training dataset. Existing defense techniques against poisoning attacks are largely attack-specific: they are designed for one specific type of attacks but do not work for other types, mainly due to the distinct principles they follow. Yet few general defense strategies have been developed. In this paper, we propose De-Pois, an attack-agnostic defense against poisoning attacks. The key idea of De-Pois is to train a mimic model the purpose of which is to imitate the behavior of the target model trained by clean samples. We take advantage of Generative Adversarial Networks (GANs) to facilitate informative training data augmentation as well as the mimic model construction. By comparing the prediction differences between the mimic model and the target model, De-Pois is thus able to distinguish the poisoned samples from clean ones, without explicit knowledge of any ML algorithms or types of poisoning attacks. We implement four types of poisoning attacks and evaluate De-Pois with five typical defense methods on different realistic datasets. The results demonstrate that De-Pois is effective and efficient for detecting poisoned data against all the four types of poisoning attacks, with both the accuracy and F1-score over 0.9 on average.

Recommender systems (RS) have become an essential component of web services due to their excellent performance. Despite their great success, RS have proved to be vulnerable to data poisoning attacks, which inject well-crafted fake profiles into RS, so that the target items can be maliciously recommended. In this paper, we first reveal that existing poisoning attacks in RS can be detected effortlessly, as the features of the generated fake profiles cannot be inconsistent with those of normal profiles all the time. We further propose RecUP, a poisoning attack in RS that can generate plausible profiles whose features stay almost the same as the normal ones, based on Generative Adversarial Networks (GAN). To tailor GAN for poisoning in RS, we develop HRGAN and devise a loss function to guide the training of the generator, along with a masking operation with selected potentially powerful profiles, so that the final generated profiles can perform malicious recommendations as expected. Evaluations against various defense methods using three real-world datasets show that, RecUP can generate the most plausible profiles while maintaining comparable attacking performance compared with state-of-the-art attacks.

Membership inference attacks seek to infer membership of individual training instances of a model to which an adversary has black-box access through a machine learning-as-a-service API. In providing an in-depth characterization of membership privacy risks against machine learning models, this paper presents a comprehensive study towards demystifying membership inference attacks from two complimentary perspectives. First, we provide a generalized formulation of the development of a black-box membership inference attack model. Second, we characterize the importance of model choice on model vulnerability through a systematic evaluation of a variety of machine learning models and model combinations using multiple datasets. Through formal analysis and empirical evidence from extensive experimentation, we characterize under what conditions a model may be vulnerable to such black-box membership inference attacks. We show that membership inference vulnerability is data-driven and corresponding attack models are largely transferable. Though different model types display different vulnerabilities to membership inference, so do different datasets. Our empirical results additionally show that (1) using the type of target model under attack within the attack model may not increase attack effectiveness and (2) collaborative learning exposes vulnerabilities to membership inference risks when the adversary is a participant. We also discuss countermeasure and mitigation strategies.

Distributed graph computation is central to applications ranging from language processing to social networks. However, natural graphs tend to have skewed power-law distrbutions where a small subset of the vertices have a large number of neighbors. Existing graph-parallel systems suffer from load imbalance, high communication cost, and inefficient processing. To address this problem, in this paper we present GraphA, an adaptive scheme for efficient large-scale graph computation. At the core of GraphA is an adaptive and uniform graph partitioning algorithm, which partitions the datasets by using an incremental number of mapping functions. GraphA further improves and leverages the ART index structure to realize fine-grained and low-cost graph storage. We have implemented GraphA both on Spark and on GraphLab. Extensive evaluation shows that GraphA significantly outperforms state-of-the-art graph-parallel systems (GraphX and PowerLyra) in ingress time, execution time and storage cost, for both real-world and synthetic graphs. GraphA achieves up to 7.1 x performance improvement over GraphX and 19.7 percent improvement over PowerLyra.

Deep learning shows great promise in providing more intelligence to the mobile web, but insufficient infrastructure, heavy models, and intensive computation limit the use of deep learning in mobile web applications. In this paper, we present DeepAdapter, a collaborative framework that ties the mobile web with an edge server and a remote cloud server to allow executing deep learning on the mobile web with lower processing latency, lower mobile energy, and higher system throughput. DeepAdapter provides a context-aware pruning algorithm that incorporates the latency, the network condition and the computing capability of the mobile device to fit the resource constraints of the mobile web better. It also provides a model cache update mechanism improving the model request hit rate for mobile web users. At runtime, it matches an appropriate model with the mobile web user and provides a collaborative mechanism to ensure accuracy. Our results show that DeepAdapter decreases average latency by 1.33x, reduces average mobile energy consumption by 1.4x, and improves system throughput by 2.1x with a considerable accuracy. Its contextaware pruning algorithm also improves inference accuracy by up to 0.3% with a smaller and faster model.

This article presents XMemPod, a hierarchical disaggregated memory orchestration system. XMemPod virtualizes cluster wide memory to scale large memory workloads in virtualized clouds. It makes three novel contributions: (1) XMemPod offers efficient, transparent, and dynamic sharing of available memory that is disaggregated across VMs on the same host or in the cluster. (2) XMemPod provides a hierarchical memory expansion framework, which enables memory-intensive workloads on a VM to expand its memory demand over virtualized host memory first, and remote memory next, before resorting to external disk. (3) XMemPod provides a suite of optimization techniques to further improve the utilization and access latency of disaggregated memory. XMemPod is deployed on a virtualized RDMA cluster without any modifications to user applications and the OSes. Evaluated with multiple workloads on unmodified Spark, Apache Hadoop, Memcached, Redis and VoltDB, using XMemPod, throughputs of these applications improve by 11× to 612× over conventional Linux, and by 1.7× to 14× over the existing representative remote memory paging systems, and yet the total amount of network traffic consumed by XMemPod is only 24 percent of the existing approaches.

ACM Transactions on Internet Technology (TOIT)is the flagship ACM transaction journal dedicated to featuring advancements in research and development of Internet technology.I became the editor-in-chief of TOIT in January 2019.In 2019, we celebrated the 50 th anniversary of the Internet.The Internet is a network of computer networks connecting computing servers and computing clients across geographical boundaries all over the world.In its first 30 years, the Internet technology grew into the dominating cyber communication technology platform.In the most recent 20 years, Internet technology continued to evolve, enabling and enhancing human-to-machine communication (H2M), humanto-human communication (H2H), and machine-to-machine communication (M2M).This evolution started from communication technology and expanded and continues to advance with storage and cognitive computing technologies at both the Edge of the Internet and in the core Clouds of the Internet.(1) Starting Point: Internet as a communication technology platform.Since the birth of ARPANET in 1969, the Internet has grown to be a dominating communication technology platform through many revolutionary advances.Here are some example milestones:• In 1974, Telenet became the first commercial version of ARPANET with the first Internet Service Provider (ISP); • In 1979, Usenet was launched for hosting news and discussion groups, creating the first appearance of "social network"; • In 1981, an NSF grant was made for establishing the computer science network (CSNET) to provide network services to university computer scientists, which led to the release of NSFNET online in 1986, connecting supercomputer centers with a network of 2K computers;

The advent of social networks and activity networks affords us an opportunity of utilizing explicit social information and activity information to improve the quality of recommendation in the presence of data sparsity. In this article, we present a social-influence-based collaborative filtering (SICF) framework over heterogeneous information networks with three unique features. First, we integrate different types of entities, links, attributes, and activities from rating networks, social networks, and activity networks into a unified social-influence-based collaborative filtering model through the intra-network and inter-network social influence. Second, we propose three social-influence propagation models to capture three kinds of information propagation within heterogeneous information networks: user-based influence propagation on user rating networks, item-based influence propagation on user-rating activity networks, and term-based influence propagation on user-review activity networks, respectively. We compute three kinds of social-influence-based user similarity scores based on three social-influence propagation models, respectively. Third, a unified social-influence-based CF prediction model is proposed to infer rating tastes by incorporating three kinds of social-influence-based similarity measures with different weighting factors. We design a weight-learning algorithm, SICF, to refine the prediction result by quantifying the contribution of each kind of information propagation to make a good balance between prediction accuracy and data sparsity. Extensive evaluation on real datasets demonstrates that SICF outperforms existing representative collaborative filtering methods.

Deep neural networks (DNNs) have demonstrated impressive performance on many challenging machine learning tasks. However, DNNs are vulnerable to adversarial inputs generated by adding maliciously crafted perturbations to the benign inputs. As a growing number of attacks have been reported to generate adversarial inputs of varying sophistication, the defense-attack arms race has been accelerated. In this paper, we present MODEF, a cross-layer model diversity ensemble framework. MODEF intelligently combines unsupervised model denoising ensemble with supervised model verification ensemble by quantifying model diversity, aiming to boost the robustness of the target model against adversarial examples. Evaluated using eleven representative attacks on popular benchmark datasets, we show that MODEF achieves remarkable defense success rates, compared with existing defense methods, and provides a superior capability of repairing adversarial inputs and making correct predictions with high accuracy in the presence of black-box attacks.

We are entering an exciting era where human intelligence is being enhanced by machine intelligence through big data fueled artificial intelligence (AI) and machine learning (ML). However, recent work shows that DNN models trained privately are vulnerable to adversarial inputs. Such adversarial inputs inject small amount of perturbations to the input data to fool machine learning models to misbehave, turning a deep neural network against itself. As new defense methods are proposed, more sophisticated attack algorithms are surfaced. This arms race has been ongoing since the rise of adversarial machine learning. This keynote provides a comprehensive analysis and characterization of the most representative attacks and their defenses. As more and more mission critical systems are incorporating machine learning and AI as an essential component in their real-world big data applications and their big data service provisioning platforms or products, understanding and ensuring the verifiable robustness of deep learning becomes a pressing challenge in the presence of adversarial attacks. This includes (1) the development of formal metrics to quantitatively evaluate and measure the robustness of a DNN prediction with respect of intentional and unintentional artifacts and deceptions, (2) the comprehensive understanding of the blind spots and the invariants in the DNN trained models and the DNN training process, and (3) the statistical measurement of trust and distrust that we can place on a deep learning algorithm to perform reliably and truthfully. In this keynote talk, I will use empirical analysis and evaluation of our cross-layer strategic teaming defense framework and techniques to illustrate the feasibility of ensuring robust deep learning.

Learning Rate (LR) is an important hyper-parameter to tune for effective training of deep neural networks (DNNs). Even for the baseline of a constant learning rate, it is non-trivial to choose a good constant value for training a DNN. Dynamic learning rates involve multi-step tuning of LR values at various stages of the training process and offer high accuracy and fast convergence. However, they are much harder to tune. In this paper, we present a comprehensive study of 13 learning rate functions and their associated LR policies by examining their range parameters, step parameters, and value update parameters. We propose a set of metrics for evaluating and selecting LR policies, including the classification confidence, variance, cost, and robustness, and implement them in LRBench, an LR benchmarking system. LRBench can assist end-users and DNN developers to select good LR policies and avoid bad LR policies for training their DNNs. We tested LRBench on Caffe, an open source deep learning framework, to showcase the tuning optimization of LR policies. Evaluated through extensive experiments, we attempt to demystify the tuning of LR policies by identifying good LR policies with effective LR value ranges and step sizes for LR update schedules.

Memory usage imbalance has been consistently observed in many virtualized Clouds and production datacenters. Such temporal memory utilization variance is a major root cause for excessive paging and thrashing on virtual servers even though there are sufficient idle memory on the same node or in the Cloud cluster. Memory disaggregation is an emerging research and development endeavor towards addressing these memory usage imbalance problems. This paper first defines and characterizes the concept of memory disaggregation, and discusses the demands and challenges of efficient memory disaggregation in cloud datacenters. It then examines some promising research issues, design choices and directions to overcome some of the challenges posed by memory disaggregation. Specifically, it proposes two major new research challenges and solution directions for enabling elastic, on-demand disaggregated memory orchestration: (1) virtual server memory and node level memory co-design and (2) local memory and remote memory co-design. A brief description of two ongoing research projects is provided for both solution directions. The paper ends with a brief discussion of other advanced and emerging memory and storage technologies and potential opportunities for memory disaggregation.

Lightweight and precise recognition is a key component of web-based augmented reality (Web AR) applications. Although edge-based distributed deep learning approach is now possible to achieve satisfactory recognition for Web AR applications, it puts significant pressure on the computation and energy consumption of the mobile web browser, especially the app-based embedded browser. Thus, reducing the model size and accelerating the inference are regarded as the two fundamental challenges to enable this edge-based collaborative recognition system efficiently. In this paper, we propose a lightweight collaborative recognition system (LCRS) for Web AR applications. LCRS contributes to three aspects: (1) we design a composite deep neural network for reducing the model size and inference latency by introducing binary convolutional neural network; (2) we provide a joint training method to co-train the general branch and the binary branch; (3) we develop a JavaScript library for the mobile web browser to execute and accelerate inference of the binary branch, which also provides a collaborative mechanism between the mobile web browser and the edge server. We have conducted extensive experiments using several well-known networks and datasets. The experimental results have shown that the proposed system outperforms the existing approaches in terms of reducing the model size by about 16x to 29x, and it also reduces end-to-end latency and outpaces the existing state-of-the-art approaches by over 3x to 60x when applying it in practical Web AR cases.

RDF data is big and continues to grow rapidly. RDF data sets are typically viewed as heterogeneous graph data sets with complex correlations and multifaceted heterogeneity. Although there have been flurry of research on processing and analyzing RDF data, efficient storage, computation, analysis of big and growing RDF data continue to challenge multiple computer science disciplines, ranging from systems, network computing, data management to data analytics. In this paper, we make use of strawman arguments on big RDF data challenges with respect to storage, computation, analysis, focusing on identifying the grand challenges in developing high performance RDF storage, parallel computation, efficient distribution, and smart RDF analytics. We attempt to answer a number of important and frequently asked questions: (1) Do we really need RDF-specific storage techniques and algorithms for building efficient and high performance RDF stores? (2) How hard can it be to effectively parallelize big RDF data for high performance storage, computation, and mining? (3) What type of distribution models can we employ to scale RDF processing models and algorithms for real-time querying (subgraph pattern matching) and mining of big RDF data? (4) Can machine learning algorithms be leveraged to effectively mine and dive into RDF data? We attempt to answer these questions by sharing some of our results, our unique experiences, and our lessons learned from a collection of research projects in RDF systems research and development. We will also discuss our ongoing research endeavors, potential applications and avenues for future work.

Ensemble learning is a methodology that integrates multiple DNN learners for improving prediction performance of individual learners. Diversity is greater when the errors of the ensemble prediction is more uniformly distributed. Greater diversity is highly correlated with the increase in ensemble accuracy. Another attractive property of diversity optimized ensemble learning is its robustness against deception: an adversarial perturbation attack can mislead one DNN model to misclassify but may not fool other ensemble DNN members consistently. In this paper we first give an overview of the concept of ensemble diversity and examine the three types of ensemble diversity in the context of DNN classifiers. We then describe a set of ensemble diversity measures, a suite of algorithms for creating diversity ensembles and for performing ensemble consensus (voted or learned) for generating high accuracy ensemble output by strategically combining outputs of individual members. This paper concludes with a discussion on a set of open issues in quantifying ensemble diversity for robust deep learning.

GPS-dependent localization, navigation and air traffic control (ATC) applications have had a significant impact on the modern aviation industry. However, the lack of encryption and authentication makes GPS vulnerable to spoofing attacks with the purpose of hijacking aerial vehicles or threatening air safety. In this paper, we propose GPS-Probe, a GPS spoofing detection algorithm that leverages the ATC messages that are periodically broadcasted by aerial vehicles. By continuously analyzing the received signal strength indicator (RSSI) and the timestamps at server (TSS) of the ATC messages, which are monitored by multiple ground sensors, GPS-Probe constructs a machine learning enabled framework to estimate the real position of the target aerial vehicle and to detect whether or not the position data is compromised by GPS spoofing attacks. Unlike existing techniques, GPS-Probe neither requires any updates of the GPS infrastructure nor updates of the GPS receivers. More importantly, it releases the requirement on time synchronization of the ground sensors distributed around the world. Using the real-world ATC data crowdsourced by the OpenSky Network, our experiment results show that GPS-Probe can achieve the detection accuracy and precision, of 81.7% and 85.3% respectively on average, and up to 89.7% and 91.5% respectively at the best.

Trajectory clustering aims at grouping similar trajectories into one cluster. It is an efficient way of finding the representative path or common trend shared by different moving objects, and also provides a foundation for movement pattern mining, anomaly detection and other applications. Existing trajectory clustering studies mainly rely on feature selection and similarity measurement based on their geographical and spatial properties. However, one obstacle hindering their wide usage is the problem of clustering accuracy in the presence of noisy or incomplete sensing data, due to limited sensory device quantity, communication errors, sensor failures, and sensor vacancy. This paper proposes an error-tolerant trajectory clustering approach by incorporating denoising methods.We propose the Robust Deep Attention Auto-encoders model (called Robust DAA) to learn the representations of low-dimensional denoising trajectories with three novel features. First, we present the deep attention auto-encoders by integrating the attention mechanism into the classical deep auto-encoder, which is capable of enhancing feature propagation and feature selection. Second, we train the deep attention auto-encoder by applying proximal method, back propagation and the Alternating Direction of Method of Multipliers (ADMM). As a result, our Robust DAA can reduce the negative influence of the noise on trajectory data. Finally, we perform clustering over the low-dimensional denoising representations using traditional clustering algorithms and demonstrates the quality of the clustering results by comparing our approach with existing representative methods. Extensive experiments are conducted on both synthetic datasets and real datasets. The results show that our approach outperforms the existing models in terms of accuracy, precision, recall and f1-score.

Deep learning techniques based on neural networks have shown significant success in a wide range of AI tasks. Large-scale training datasets are one of the critical factors for their success. However, when the training datasets are crowdsourced from individuals and contain sensitive information, the model parameters may encode private information and bear the risks of privacy leakage. The recent growing trend of the sharing and publishing of pre-trained models further aggravates such privacy risks. To tackle this problem, we propose a differentially private approach for training neural networks. Our approach includes several new techniques for optimizing both privacy loss and model accuracy. We employ a generalization of differential privacy called concentrated differential privacy(CDP), with both a formal and refined privacy loss analysis on two different data batching methods. We implement a dynamic privacy budget allocator over the course of training to improve model accuracy. Extensive experiments demonstrate that our approach effectively improves privacy loss accounting, training efficiency and model quality under a given privacy budget.

Mobile augmented reality (Mobile AR) is gaining increasing attention from both academia and industry. Hardware-based Mobile AR and App-based Mobile AR are the two dominant platforms for Mobile AR applications. However, hardware-based Mobile AR implementation is known to be costly and lacks flexibility, while the App-based one requires additional downloading and installation in advance and is inconvenient for cross-platform deployment. In comparison, Web-based AR (Web AR) implementation can provide a pervasive Mobile AR experience to users thanks to the many successful deployments of the Web as a lightweight and cross-platform service provisioning platform. Furthermore, the emergence of 5G mobile communication networks has the potential to enhance the communication efficiency of Mobile AR dense computing in the Web-based approach. We conjecture that Web AR will deliver an innovative technology to enrich our ways of interacting with the physical (and cyber) world around us. This paper reviews the state-of-the-art technology and existing implementations of Mobile AR, as well as enabling technologies and challenges when AR meets the Web. Furthermore, we elaborate on the different potential Web AR provisioning approaches, especially the adaptive and scalable collaborative distributed solution which adopts the osmotic computing paradigm to provide Web AR services. We conclude this paper with the discussions of open challenges and research directions under current 3G/4G networks and the future 5G networks. We hope that this paper will help researchers and developers to gain a better understanding of the state of the research and development in Web AR and at the same time stimulate more research interest and effort on delivering life-enriching Web AR experiences to the fast-growing mobile and wireless business and consumer industry of the 21st century.

In social networks, the identification of key elements in a topic propagation network, including key messages, paths, and users, is important in network public opinion mining and control. In view of the three key elements, this paper proposes a social hotspot tracking scheme termed MPURank, which is based on tripartite graphing and a multimessage-driven iterative method. First, the topic is regarded as a network of simultaneous multimessage dissemination, and retweeting topology graphing is used to determine the propagation path; then, a topic tripartite graph is established to elucidate the interrelationship among the nodes of messages, paths, and users. Second, an iterative scoring algorithm based on hyperlink-induced topic search is proposed to rank the key elements based on the tripartite graph. Using the initial score vectors of different elements and weight matrices, the algorithm obtains the final score sequences of the elements. This algorithm circumvents the influence of the multimessage network and propagation path on the evolution of public opinion, overcoming the issues of node attributes and topics, and the influence of multimessage and multipath complexity in source tracing, thus identifying the key elements. Finally, we conducted an experiment based on real-world data, which verifies the utility of the algorithm.

As more business workflow systems are being deployed in modern enterprises and organizations, more employee-activity log data are being collected and analyzed. In this paper, we develop a latent ability model (LAM) as a generative probabilistic learning framework for workforce analytics over employee-activity logs. The LAM development is novel in three aspects. First, we introduce the concept of latent ability variables to model hidden relations between employees and activities in terms of job performance, such as the set of skills provided by an employee and the set of skills required by an activity, and how well they matchup in employee-activity assignment. Second, we construct the latent ability model by learning latent ability parameters from the employee-activity log data using expectation-maximization and gradient descent. Finally, we leverage LAM to build inference and prediction models for employee performance prediction, employee ability comparison, and employee-activity matchup quality estimation. We evaluate the accuracy and efficiency of our approach using real log datasets collected from a workflow system deployed in the government of the city of Hangzhou, China, which consists of 5,287,621 log records over two years involving 744 activities and 1,725 employees. We show that LAM approach outperforms existing representative methods in both accuracy and efficiency.

The universal popularity of GPS-enabled mobile devices and traffic navigation services has fueled the growth of trajectory data, as evidenced by Uber Movement and NYC taxi data release. Although trajectory data can generate valuable insights and value-added services for many, publishing this data while respecting mobile users' privacy has been a long-standing challenge. In this paper, we present DP-Star, a methodical framework for publishing trajectory data with differential privacy guarantee as well as high utility preservation. DP-Star relies on a novel combination of several components. First, DP-Star's normalization algorithm uses the Minimum Description Length metric to summarize raw trajectories using their representative points, thereby achieving a desirable trade-off between the preciseness and conciseness of their information content. Second, DP-Star constructs a density-aware grid which ensures spatial densities can be preserved despite the noise added to satisfy differential privacy. Third, DP-Star preserves the correlations between trajectories' end points through a private trip distribution, and intermediate points through a private Markov mobility model. Finally, DP-Star estimates users' trip lengths using a median length estimation method, and generates synthetic trajectories that preserve both differential privacy and high utility. Our experimental comparison shows that DP-Star significantly outperforms existing approaches in terms of trajectory utility and accuracy.

As accessing remote cloud via cellular network is costly due to the lower bandwidth, higher wide area network (WAN) latency, and higher energy consumption, mobile cloudlet that formed by several edge mobile devices has become an emerging computing paradigm and attracted increasing attention recently. Being different from the existing studies that mainly focus on the issue of computation offloading among the peers, this paper investigates the problem of cooperative data sharing among peers to overcome the data dissymmetry, especially with the presence of dynamic network context. First, a publish/subscribe-based data sharing model is designed to cope with the unpredictable communication condition. Then, the data transmission scheduling within cooperative mobile devices is formulated as a utility maximization optimization considering the limited channel capacity, heterogeneous quality of experience (QoE) requirements, and incentive mechanism for participation. To encourage cooperation among mobile devices, a data downloading/uploading queuing mechanism is elegantly designed. Furthermore, an online algorithm without predicting the future information on request arrivals and network changes is developed to simultaneously optimize data transmission and communication interface selection in the long run. Theoretical analysis shows that the proposed algorithm is able to obtain a utility arbitrarily close to the offline optimum and guarantee the delay bound. Simulations demonstrate the effectiveness and the superiority of the proposed algorithm over some existing typical strategies.

Distributed RAM storage aggregates the RAM of servers in data center networks (DCN) to provide extremely high I/O performance for large-scale cloud systems. For quick recovery of storage server failures, MemCube [53] exploits the proximity of the BCube network to limit the recovery traffic to the recovery servers’ 1-hop neighborhood. However, the previous design is applicable only to the symmetric BCube( n , k ) network with n k +1 nodes and has suboptimal recovery performance due to congestion and contention. To address these problems, in this article, we propose CubeX, which (i) generalizes the “1-hop” principle of MemCube for arbitrary cube-based networks and (ii) improves the throughput and recovery performance of RAM-based key-value (KV) store via cross-layer optimizations. At the core of CubeX is to leverage the glocality (= globality + locality) of cube-based networks: It scatters backup data across a large number of disks globally distributed throughout the cube and restricts all recovery traffic within the small local range of each server node. Our evaluation shows that CubeX not only efficiently supports RAM-based KV store for cube-based networks but also significantly outperforms MemCube and RAMCloud in both throughput and recovery time.

With the widespread use of GPS-equipped smartphones and Internet of Things devices, a huge amount of data with location information is being generated at an unprecedented rate. To gain a deeper insight into such a plethora of spatial data, scientists and engineers are widely using spatial queries for their big data applications. However, because of not only the massive spatial data size but also the complexity of spatial query processing, they are struggling to efficiently process the spatial queries. In this paper, we propose lightweight and scalable indexing and querying services for big spatial data stored in distributed storage systems or graph-based systems. Our spatial services have several advantages over existing approaches. First, our services can be easily applied to existing storage systems or graph-based models without modifying the internal implementation of existing systems/models. Second, our services achieve high pruning power by efficiently selecting only relevant spatial objects based on a simple yet effective filter. Third, our services support a customizable and easy-to-use control of index data size by adjusting the precision of indexed geometries. Lastly, our services support efficient updates of spatial data. Our experimental results using real-world datasets validate the effectiveness and efficiency of our spatial services.

Container technique is gaining increasing attention in recent years and has become an alternative to traditional virtual machines. Some of the primary motivations for the enterprise to adopt the container technology include its conveniency to encapsulate and deploy applications, lightweight operations, as well as efficiency and flexibility in resources sharing. However, there still lacks an in-depth and systematic comparison study on how big data applications, such as Spark jobs, perform between a container environment and a virtual machine environment. In this paper, by running various Spark applications with different configurations, we evaluate the two environments from many interesting aspects, such as how convenient the execution environment can be set up, what are makespans of different workloads running in each setup, how efficient the hardware resources, such as CPU and memory, are utilized, and how well each environment can scale. The results show that compared with virtual machines, containers provide a more easy-to-deploy and scalable environment for big data workloads. The research work in this paper can help practitioners and researchers to make more informed decisions on tuning their cloud environment and configuring the big data applications, so as to achieve better performance and higher resources utilization.

Many big data applications are memory-intensive workloads and perform iterative analytics algorithms. When the dataset used in each iteration of the analytic job exceeds the physical memory of their allocation, this type of workloads suffers from serious performance degradation or experience out of memory error. Existing proposals focus on estimating working set size for accurate resource allocation of executors, but lack of desired efficiency and transparency. This paper presents an efficient shared-memory based memory paging service, called FastSwap. The design of FastSwap makes a number of original contributions. First, FastSwap improves VM memory swapping performance by leveraging idle host memory and redirecting the VM swapping traffic to the host-guest compressed shared memory swap area. Second, FastSwap develops a compressed swap page table as an efficient index structure to provide high utilization of shared memory swap area through supporting multi-granularity of compression factors. Third, FastSwap provides hybrid swap-out and proactive swap-in to further improve the performance of shared memory swapping. Finally, FastSwap is by design light-weighted and non-intrusive. We evaluate FastSwap using a set of well-known big data analytics workloads and benchmarks, such as Spark, Redis, HiBench, SparkBench and YCSB. The results show that FastSwap offers up to two orders of magnitude performance improvements over existing memory swapping methods and more than four orders of magnitude faster than conventional disk based VM swapping facility.

Big Data has fueled the wide deployment of Deep Learning (DL) in many fields, such as image classification, voice recognition and NLP. The growing number of open source DL software frameworks has put forward high demands on comparative study of their efficiency with respect to both runtime performance and accuracy. This paper presents a brief overview of our empirical evaluation of four representative DL frameworks: TensorFlow, Caffe, Torch and Theano through a comparative analysis and characterization. First, we show that the complex interactions among neural networks (NN), hyper-parameters, their specific runtime implementations and datasets are latent factors for the uncertainty of runtime performance and accuracy. Second, we characterized the CPU/GPU resource usage patterns under different configurations for different frameworks to obtain an in-depth understanding of the impact of different batch sizes. Third, we describe the data loading process of ImageNet for TensorFlow and present an experimental characterization of TensorFlow with respect to its data loading process when the dataset is too large to fit into the main memory of the CPU server. We conjecture that our experimental characterization and analysis can offer empirical guidance for users and application developers to select the right DL frameworks and configurations for their domain-specific learning tasks and datasets.

While location-based services and applications are increasing in popularity, there are growing concerns over users' location privacy. Although there exist general purpose mobile permission systems and cloaking techniques, these techniques suffer from several problems when applied to continuous location and GPS access, as they are often rigid, coarse-grained, not sufficiently personalizable, and unaware of road network semantics. This paper proposes PrivacyZone, a novel system for constructing personalized fine-grained privacy quarantine regions and protecting users' privacy within these regions. PrivacyZone allows users to seamlessly enter their privacy specifications under spatial, temporal, and semantic customization. Novel challenges arise from having to enforce privacy zones for large volume and variety of users with frequent location updates. We show that naive privacy zone processing techniques are inefficient and cause excessive energy consumption. We therefore develop advanced processing techniques based on the concept of safe hibernation. We empirically evaluate our techniques to demonstrate their trade-offs with respect to hibernation time, computation effort, and network bandwidth usage. Our results show that PrivacyZone is efficient, scalable, and flexible, while preserving users' location privacy.

As mobile devices and location-based services become increasingly ubiquitous, the privacy of mobile users' location traces continues to be a major concern. Traditional privacy solutions rely on perturbing each position in a user's trace and replacing it with a fake location. However, recent studies have shown that such point-based perturbation of locations is susceptible to inference attacks and suffers from serious utility losses, because it disregards the moving trajectory and continuity in full location traces. In this paper, we argue that privacy-preserving synthesis of complete location traces can be an effective solution to this problem. We present AdaTrace, a scalable location trace synthesizer with three novel features: provable statistical privacy, deterministic attack resilience, and strong utility preservation. AdaTrace builds a generative model from a given set of real traces through a four-phase synthesis process consisting of feature extraction, synopsis learning, privacy and utility preserving noise injection, and generation of differentially private synthetic location traces. The output traces crafted by AdaTrace preserve utility-critical information existing in real traces, and are robust against known location trace attacks. We validate the effectiveness of AdaTrace by comparing it with three state of the art approaches (ngram, DPT, and SGLT) using real location trace datasets (Geolife and Taxi) as well as a simulated dataset of 50,000 vehicles in Oldenburg, Germany. AdaTrace offers up to 3-fold improvement in trajectory utility, and is orders of magnitude faster than previous work, while preserving differential privacy and attack resilience.

The cloud computing technology offers consistent access to large-scale computing capabilities, thereby bringing convenience to life. However, the virtualized cloud systems are still too vulnerable to maintain performance scalability and service agility once a task burst surges in without any warning. A mounting account of research has been conducted on proper strategies for accurate workload prediction as well as effective resource reservation and arrangement, but commonly cloud providers seek help to strategies that deploy excessive resources, adding overhead cost and sacrificing the cloud's advantage of scalability, or otherwise fail to reconfigure timely and properly, causing dissatisfaction and even financial loss, which are not expected by both cloud providers and clients.

Blockchain is one of the technology innovations for sharing data across organizations through a peer to peer overlay network. Many blockchain- based data sharing applications, such as sharing Electronic Health Records (EHRs) among different Care Delivery Organizations (CDOs), require privacy preserving verification services with dual capabilities. On one hand, the users want to verify the authenticity of EHR data as well as the identity of the signer. On the other hand, the signer wants to keep his real identity private such that others cannot trace and infer his identity information. However, typical blockchain systems that use pseudonyms as public keys, such as Bitcoin's blockchain, cannot support such privacy-preserving verification. In such systems, it is hard to verify the authenticity of signer's identity, and adversaries or curious parties can guess the real identity from the series of statements and actions taken with a specific pseudonym through inference attacks, such as by transaction graph analysis. In this paper, we propose a decentralized attribute- based signature scheme for healthcare blockchain, which provides efficient privacy-preserving verification of authenticity of EHR data and signer's identity. We also describe a holistic on-chain and off- chain collaborative storage system for efficient storage and verification EHR data. The analysis and experiments show that our scheme is effective and deployable.

With increasing number of open-source deep learning (DL) software tools made available, benchmarking DL software frameworks and systems is in high demand. This paper presents design considerations, metrics and challenges towards developing an effective benchmark for DL software frameworks and illustrate our observations through a comparative study of three popular DL frameworks: TensorFlow, Caffe, and Torch. First, we show that these deep learning frameworks are optimized with their default configurations settings. However, the default configuration optimized on one specific dataset may not work effectively for other datasets with respect to runtime performance and learning accuracy. Second, the default configuration optimized on a dataset by one DL framework does not work well for another DL framework on the same dataset. Third, experiments show that different DL frameworks exhibit different levels of robustness against adversarial examples. Through this study, we envision that unlike traditional performance-driven benchmarks, benchmarking deep learning software frameworks should take into account of both runtime and accuracy and their latent interaction with hyper-parameters and data-dependent configurations of DL frameworks.

Memory is increasingly becoming a bottleneck for big data and latency-sensitive applications in virtualized systems. Memory efficiency is critical for high-performance execution of virtual machines (VMs). Mechanisms proposed for improving memory utilization often rely on an accurate estimation of VM working set size at runtime, which is difficult under changing workloads. This paper explores opportunities for improving memory efficiency and their impacts on the performance of VM executions. First, we show that if each VM is initialized with an application-specified lower bound memory, then by maintaining a shared memory region across VMs in the presence of temporal memory usage variations on the host, those VMs under high memory pressure can minimize their performance loss by opportunistically and transparently harvesting idle memory on other VMs. Second, we show that by enabling on-demand VM memory allocation and deallocation in the presence of changing workloads, VM performance degradation due to memory swapping can be reduced effectively, compared to the conventional VM configuration scenario, in which all VMs are allocated with the upper-bound of memory requested by their applications. Third, we show that by providing shared memory pipes between co-located VMs, the inter-VM communication can speed up by avoiding unnecessary overhead of communication via the network. We develop MemLego, a lightweight shared memory based system, to achieve all these benefits without requiring any modification to user applications and the OSes. We demonstrate the effectiveness of these opportunities through extensive experiments on unmodified Redis and MemCached. Using MemLego, the throughput of Redis and Memcached improves by up to 4x over the native system without MemLego, up to 2 orders of magnitude when the applications working set size does not fit in memory.

NgramCNN is a deep convolutional neural network developed for classification of graphs based on common substructure patterns and their latent relationships in the collection of graphs. Our NgramCNN deep learning framework consists of three novel components: (1) The concept of n-gram graph block to transform each raw graph object into a sequence of n-gram blocks connected through overlapping regions. (2) The diagonal convolution layer to extract local patterns and connectivity features hidden in the n-gram blocks by performing n-gram normalization before conducting deep learning through the network of convolution layers. (3) The extraction of deeper global patterns based on the local patterns and the ways that they respond to overlapping regions by building a n-gram deep convolutional neural network. Extensive evaluation of NgramCNN using five real graph repositories from bioinformatics and social networks domains show the effectiveness of NgramCNN over the existing state of art methods with high accuracy and comparable performance.

Reachability query asks whether a vertex can reach another vertex on large directed graphs. It is one of the most fundamental graph operators and has attracted many researchers to study it. Although there are many approaches solving this problem, it still remains a challenging problem when it comes to leverage the three main costs: the index construction time, the index size, and the query time on large and dense graphs. In this paper, we propose a High Dimension Graph Labeling approach to answer reachability queries. First, we recursively partition a graph into disjoint non-shared graphs, among which there are no common vertices, and cross edges. Second, we build a four dimensional label - one dimension of layer, one dimension of sub-graph and two dimensions of interval for each vertex. With the layer label and the sub-graph label, we can determine the positions (non-shared graphs) of any two vertices. Two dimensional interval label is used to assist to answer the reachability queries for vertex pair in those non-shared graphs. Finally, we design algorithms to answer researchability queries efficiently. In order to speed up query answering, we also build two directional labels: up and down labels to filter vertices quickly. The extensive experiments on 28 large/small and dense/sparse graphs show that building the high dimensional index is quickly and the index size is also competitive compared with most of the state of the art approaches. The results also show that our approach is more scalable and efficient than the state-of-the-art approaches in answering reachability queries.

In Big data and IoT environments, process execution generates huge-sized data some of which is subsequently obtained by sensors. The main issue in such areas has been the necessity of analyzing data in order to suggest enhancements to processes. In this regard, evaluation of process model conformance to the execution log is of great importance. For this purpose, previous reports on process mining approaches have advocated conformance checking by fitness measure, which is a process that uses token replay and node-arc relations based on Petri net. However, fitness measure so far has not considered statistical significance, but just offers a numeric ratio. We herein propose a statistical verification method based on the Kolmogorov–Smirnov (K–S) test to judge whether two different log datasets follow the same process model. Our method can be easily extended to determinations that process execution actually follows a process model, by playing out the model and generating event log data from it. Additionally, in order to solve the problem of the trade-off between model abstraction and process conformance, we also propose the new concepts of Confidence Interval of Abstraction Value (CIAV) and Maximum Confidence Abstraction Value (MCAV). We showed that our method can be applied to any process mining algorithm (e.g. heuristic mining, fuzzy mining) that has parameters related to model abstraction. We expect that our method will come to be widely utilized in many applications dealing with business process enhancement involving process-model and execution-log analyses.

The problem of the curse of dimensionality for processing large high-dimensional datasets has been an open challenge. Numerous research efforts have been proposed for improving query performance in high-dimensional space through hierarchical indexing using the R-tree or its variants and exploring parallel processing of the R-tree on GPUs. Despite these existing efforts, the curse of dimensionality remains to be a grand challenge since the existing methods deteriorate drastically as the dimensionality of datasets increases. To cope with this problem, we present a novel GPU-aware parallel indexing method called G-tree, which offers consistent and stable performance in high-dimensional space. The rationale of the G-tree is to combine the efficiency of the R-tree in low-dimensional space with the massive parallel processing potential of GPUs by introducing a new data structure and three new optimization techniques to better utilize the GPU memory structure for accelerating both index search and index node access on GPUs. The first two optimizations promote effective parallelism utilization in GPU memory access. We dedicate the third optimization to further speed up the G-tree index by conducting progressive filtering using our dimension filters. We evaluate the validity of the G-tree approach by extensive experiments on high-dimensional datasets, showing that the G-tree outperforms the existing state-of-the-art techniques.

The unique characteristics of underwater channels, such as the mobility of sensor nodes, require robust and efficient security mechanisms for secure and efficient data collection in mobile underwater sensor networks (UWSNs). In this paper, we propose a novel three-tier framework for securing data collections in mobile UWSNs, called SenseVault. First, to support both the authentication of intra-cluster nodes and the authentication of inter-cluster nodes, we propose a cubic cluster formation approach to adapt to dynamic environments with flexible split and merge operations, and use cryptographic hash functions to derive secret keys for the authentication of nodes moving among the clusters. Second, we design a lightweight node revocation and authentication key update mechanism based on higher-order polynomial. To provide strong resilience against misbehavior of the compromised cluster head, we further introduce a Tripartite Cooperation Update (TCU) scheme by leveraging the cooperation among surface stations, safe neighboring cluster heads, and safe intra-cluster members. Third, we propose a deterministic quantization approach, called Virtual Phase Shift (VPS) for the secure generation of secret keys at the physical layer. Extensive experimental results demonstrate the effectiveness of the proposed framework under different scenarios.

Conventional private data publication schemes are targeted at publication of sensitive datasets either after the k -anonymization process or through differential privacy constraints. Typically these schemes are designed with the objective of retaining as much utility as possible for the aggregate queries while ensuring the privacy of the individual records. Such an approach, though suitable for publishing aggregate information as public datasets, is inapplicable when users have different levels of access to the same data. We argue that existing schemes either result in increased disclosure of private information or lead to reduced utility when some users have more access privileges than the others. In this article, we present an anonymization framework for publishing large datasets with the goals of providing different levels of utility to the users based on their access privilege levels. We design and implement our proposed multilevel utility-controlled anonymization schemes in the context of large association graphs considering three levels of user utility, namely, (1) users having access to only the graph structure, (2) users having access to the graph structure and aggregate query results, and (3) users having access to the graph structure, aggregate query results, and individual associations. Our experiments on real large association graphs show that the proposed techniques are effective and scalable and yield the required level of privacy and utility for each user privacy and access privilege level.

In the cloud, multiple tenants share the resource of datacenters and their applications compete with each other for scarce network bandwidth. Current studies have shown that the lack of bandwidth guarantee causes unpredictable network performance, leading to poor application performance. To address this issue, several virtual network abstractions have been proposed which allow the tenants to reserve virtual clusters with specified bandwidth between the Virtual Machines (VMs) in the datacenters. However, all these existing proposals require the tenants to deterministically characterize the bandwidth demands in the abstractions, which can be difficult and result in inefficient bandwidth reservation due to the demand uncertainty. In this paper, we explore a virtual cluster abstraction with stochastic bandwidth characterization to address the bandwidth demand uncertainty. We propose Stochastic Virtual Cluster (SVC), which models the bandwidth demand between VMs in a probabilistic way. Based on SVC, we develop a stochastic framework for virtual cluster allocation, in which the admitted virtual cluster's bandwidth demands are satisfied with a high probability. Efficient VM allocation algorithms are proposed to implement the framework while reducing the possibility of link congestion through minimizing the maximum bandwidth occupancy of a virtual cluster on physical links. Using simulations, we show that SVC achieves the trade-off between the job concurrency and the average job running time, and demonstrate its effectiveness for accommodating cloud application workloads with highly volatile bandwidth demands and its improvement to work-conserving bandwidth enforcement.

Outsourcing medical data and their search services to a third party cloud have been a popular trend for many medical practices, because using healthcare cloud services can help cut down the cost of Electronic Health Records (EHR) systems in terms of front-end ownership cost and IT maintenance burdens. Healthcare cloud applications need searchable encryption with the following two capabilities for protecting data privacy and access privacy: (1) the healthcare providers need to share the encrypted data with authorized users and enable querying over encrypted data, and (2) they also need to keep the query keywords and associated search operations private such that healthcare data hosting service providers cannot gain access to unauthorized content or trace and infer sensitive data stored in the healthcare cloud. This survey paper describes the notion of searchable encryption (SE) in the context of healthcare applications and characterize the SE use cases into four scenarios in healthcare. Then we provide a comprehensive overview of the four representative SE techniques: searchable symmetric encryption (SSE), public key encryption with keyword search (PEKS), attribute-based encryption with keyword search (ABKS), and proxy re-encryption with keyword search (PRES) according to different EHR retrieving scenarios and requirements. We categorize and compare the different SE schemes in terms of their security, efficiency, and functionality. The survey is designed to benefit both experienced researchers in the computer science (CS) field and non-specialists who are domain scientists or healthcare professionals with limited CS and information security background. Thus, we are in favor of technological overview of the state of art searchable encryption models and the underlying key techniques, instead of detailed proofs and constructions of the respective SE algorithms. We describe how the existing SE schemes relate to and differ from one another, and point out the connections between the SE techniques and the security and privacy requirements of healthcare applications and the open research problems.

High-performance analysis of big data demands more computing resources, forcing similar growth in computation cost. So, the challenge to the HPC system designers is providing not only high performance but also high performance at lower cost. For high performance yet cost effective cyberinfrastructure, we propose a new system model augmenting Amdahl's second law for balanced system to optimize price-performance-ratio. We express the optimal balance among CPU-speed, I/O-bandwidth and DRAM-size (i.e., Amdahl's I/O-and memory-number) in terms of application characteristics and hardware cost. Considering Xeon processor and recent hardware prices, we showed that a system needs almost 0.17GBPS I/O-bandwidth and 3GB DRAM per GHz CPU-speed to minimize the price-performance-ratio for data-and compute-intensive applications. To substantiate our claim, we evaluate three different cluster architectures: 1) SupermikeII, a traditional HPC cluster, 2) SwatIII, a regular datacenter, and 3) CeresII, a MicroBrick-based novel hyperscale system. CeresII with 6-Xeon-D1541 cores (2GHz/core), 1-NVMe SSD (2GBPS I/O-bandwidth) and 64GB DRAM per node, closely resembles the optimum produced by our model. Consequently, in terms of price-performance-ratio CeresII outperformed both SupermikeII (by 65-85%) and SwatIII (by 40-50%) for data-and compute-intensive Hadoop benchmarks (TeraSort and WordCount) and our own benchmark genome assembler based on Hadoop and Giraph.

This paper presents a cloud-based positioning method that leverages visualized signal images through visual analytics and deep learning. At a mobile client, such as a smart phone, this approach transforms multidimensional signals captured at a known location and at a given time into a signal image and transmits such visual signal images to the Cloud. By collecting and storing many such visual signal images as fingerprints, we can build a visual signal image cloud and produce a signal image map for the geographical region of interest and utilize such signal image map to serve the positioning requests of mobile clients on the move. When a user Alice wants to know her current position, her mobile client will generate a signal image from the multiple signals it receives with timestamp and send this query image to a Cloud server. The server searches the existing signal images stored in the cloud to find those that are similar to this query signal image and estimates the positioning of Alice based on the locations of those similar signal images collected by the server. We evaluate our visual signal images based positioning system on the entire two floors of a large department store and on the street and shops outside the department store, with the signal images collected over 30 minutes before serving positioning queries. The mean error of up to 4 meters is observed. To further verify the applicability of the proposed method, extensive experiments were conducted to distinguish whether a user is indoor or outdoor by applying a deep learning algorithm with 60% of signal images collected for training and 40% signal images for testing. This experiment shows that the proposed method is able to distinguish indoor and outdoor with accuracy of about 95%.

Key resources are important, influential and powerful performers in a social network structure. Identifying them in the social network of a business process activity is beneficial and rewarding. One of the most effective centrality measures for identification of the key nodes in a social network is to rank resources based on a selection of criteria. PageRank is a representative example of such algorithms, which was first utilized in the Google search engine in 1998. However, the PageRank approach merely assumes a single link as a vote, which allows one originator to link or transfer his work to others more than once in a handover work scenario. We argue that this problem can lead to inaccurate influence based ranking in the context of business processes for resources in a social network. In this paper, we propose f-PageRank, a new approach specifically designed to identify the key resources in a social network generated from a business process activity. We evaluate our proposed method by comparing it with the existing approaches in process mining tools, such as degree centrality, betweenness centrality, BaryRanker, and HITS. The experimental results show that our approach can obtain a satisfying outcome.

The introduction of verifiable computation came as a result of the increasingly common phenomenon of "outsourcing" computation to untrusted servers and also to the growing desire of weak clients to outsource computational tasks to more powerful computation services like in cloud computing. Verifiable computation enables a computer to offload the computation of some function, to other perhaps untrusted cloud servers, while maintaining verifiable results. The servers evaluate the function and return the result with a proof that the computation of the function was carried out correctly. In the previous setting of verifiable computation, there is only one data provider. But in practice, there exist scenarios such as a network of sensors where each sensor collects data (e.g. air temperature in a certain area of a city) and stores them on servers. A control unit performs computation (e.g. the average air temperature of the city in certain period) on the outsourced data on the cloud (e.g. Amazon Cloud). When the control unit receives the results, it wants to verify the correctness of the computation results returned by the servers. For this scenario, we define a novel two-server multiclient verifiable computation service framework for outsourced data. An efficient construction is proposed, whose security is based on the existence of one-way functions. There are two advantages in our construction: (1) The size of the proof vouching for the correctness of computation result is independent with the number of data providers. (2) The verification only needs two equality tests executed by one who wants to get the computation result. We also experimentally analyze our construction and show our construction is very efficient in practice.

On a crowdsourcing platform, in order to cheat for rewards or sabotage the crowdsourcing processes, spam workers submit numerous random and erroneous answers to the tasks published by honest requesters. This type of behaviours extremely reduces the enthusiasm of honest users, which may lead a crowdsourcing platform to a failure. To defend the threats from spam workers, reputation-based defense mechanisms and verification-based defense mechanisms have been proposed in crowdsourcing environments. However, reputation-based defense fails to indicate the trust level of a worker who boosts its reputation by transacting with his/her accomplices. In addition, verification-based defense is costly and ineffective when facing a large number of spam workers with "good" reputations. Thus, it is a challenging problem to effectively defend the threats from spam workers. In this paper, we propose a new trust vector-based threat defense model CrowdDefense. Firstly, we build a Crowdsourcing Trust Network (CTN) consisting of requesters, workers and their transaction relations. Then, we analyze three threat patterns of spam workers. Based on the analysis, we infer the trust relation between a worker and a requester who are indirectly linked with each other. Moreover, we compute a worker's trust relations with different requesters, and present them in a new Worker Trust Vector (WTV) that indicates the worker's global trust. As spam workers always succeed in the transactions with their accomplices, they cannot obtain comprehensively good trust scores in their WTVs and thus are prevented from participating in tasks. The experiments demonstrate that CrowdsDefense significantly outperforms the state-of-the-art approaches in terms of preventing spam workers from participating in the tasks published by honest requesters.

Outsource encrypted data has attracted attentions from industry and academics for storing sensitive data in third party clouds. Many cloud applications need privacy preserving multiple keywords search services over encrypted data with dual capabilities. On one hand, they need to keep the query keywords and associated search operations private such that data hosting service providers cannot trace and infer sensitive data stored in the third party data hosting servers. On the other hand, they need to support multiple keywords search to significantly improve the search efficiency. However, current keyword search protocols for encrypted data are not practical with poor privacy and low efficiency. In this paper, we propose a new oblivious multiple keywords search (OMKS) service, which provides privacy for both users and cloud storage service provider and supports efficient multiple keywords search. Compared to previous oblivious keyword search (OKS) protocols, our protocols maintain strong privacy, i.e., database security and query privacy, and effectively support disjunctive and conjunctive keywords search. The analysis and experiments show that OMKS protocols significantly reduce the storage and communication overhead. Moreover, the computation overhead of conjunctive search is not increased with the number of query keywords such that it can performs highly efficient conjunctive keywords search.

With the development of next-generation sequencing (NGS), DNA/RNA sequencing has become cheaper and more efficient. Today, a whole human genome can be sequenced under $1,000, providing opportunities for large-scale bioinformatic analysis on big datasets. However, most of existing bioinformatic analysis tools are programmed for single server based computing platform and not suitable to process such big datasets. As Hadoop MapReduce and Spark are gaining popularity as cluster computing based big data processing platform, more and more bioinformatic applications start to explore cluster computing platform for large scale data analysis. In this paper we present an in-depth experimental study on deploying Spark clusters for high performance bioinformatic short sequence reconstruction. Our experimental results enable us to answer a number of challenging and yet most frequently asked questions regarding efficient management of bioinformatic data analysis services on Spark systems. Example questions include how to best split big dataset into multiple partitions, and how to distribute data partitions and bioinformatic analysis tasks on a Spark cluster for carrying out a high performance distributed analysis job? What types of memory models are effective for bioinformatic data analysis services on a Spark cluster? Why do different bioinformatic data analysis operations exhibit different throughput performance on the same Spark cluster? We conjecture that this experimental study not only demonstrates the feasibility of high performance bioinformatic data analysis on Spark platform, but also will help bioinformatic application developers to make more informed decisions on both design and configuration of Spark Cluster, managing and tuning parameters of Spark runtime system for enhancing the performance of large scale big data analytics.

Service optimization and energy conservation requires a thorough understanding of the performance impact of different hardware configurations. In this paper we focus on the configuration of memory and investigate the impact of memory dynamic voltage and frequency scaling (DVFS) on the performance of services/applications. We propose a quantitative metric called frequency sensitivity (FS) and study memory FS of various benchmarks. Our experiments yield several insights for memory DVFS based performance tuning.

Ballooning is a popular solution for dynamic memory balancing. However, existing solutions may perform poorly in the presence of heavy guest swapping. Furthermore, when the host has sufficient free memory, guest virtual machines (VMs) under memory pressure is not be able to use it in a timely fashion. Even after the guest VM has been recharged with sufficient memory via ballooning, the applications running on the VM are unable to utilize the free memory in guest VM to quickly recover from the severe performance degradation. To address these problems, we present MemFlex, a shared memory swapper for improving guest swapping performance in virtualized environment with three novel features: (1) MemFlex effectively utilizes host idle memory by redirecting the VM swapping traffic to the host-guest shared memory area. (2) MemFlex provides a hybrid memory swapping model, which treats a fast but small shared memory swap partition as the primary swap area whenever it is possible, and smoothly transits to the conventional disk-based VM swapping on demand. (3) Upon ballooned with sufficient VM memory, MemFlex provides a fast swap-in optimization, which enables the VM to proactively swap in the pages from the shared memory using an efficient batch implementation. Instead of relying on costly page faults, this optimization offers just-in-time performance recovery by enabling the memory intensive applications to quickly regain their runtime momentum. Performance evaluation results are presented to demonstrate the effectiveness of MemFlex when compared with existing swapping approaches.

Most of mobile object trajectory clustering analysis to date has been focused on clustering the location points or sub-trajectories extracted from trajectory data. This paper presents TraceMob, a systematic approach to clustering whole trajectories of mobile objects traveling in road networks. TraceMob as a whole trajectory clustering framework has three unique features. First, we design a quality measure for the distance between two whole trajectories. By quality, we mean that the distance measure can capture the complex characteristics of trajectories as a whole including their varying lengths and their constrained movement in the road network space. Second, we develop an algorithm that transforms whole trajectories in a road network space into multidimensional data points in a euclidean space while preserving their relative distances in the transformed metric space. This transformation enables us to effectively shift the clustering task for whole mobile object trajectories in the complex road network space to the traditional clustering task for multidimensional data in a euclidean space. Third, we develop a cluster validation method for evaluating the clustering quality in both the transformed metric space and the road network space. Extensive experimental evaluation with trajectories generated on real road network maps of different cities shows that TraceMob produces higher quality clustering results and outperforms existing approaches by an order of magnitude.

Convolutional Neural Network (CNN) has gained attractions in image analytics and speech recognition in recent years. However, employing CNN for classification of graphs remains to be challenging. This paper presents the Ngram graph-block based convolutional neural network model for classification of graphs. Our Ngram deep learning framework consists of three novel components. First, we introduce the concept of n-gram block to transform each raw graph object into a sequence of n-gram blocks connected through overlapping regions. Second, we introduce a diagonal convolution step to extract local patterns and connectivity features hidden in these n-gram blocks by performing n-gram normalization. Finally, we develop deeper global patterns based on the local patterns and the ways that they respond to overlapping regions by building a n-gram deep learning model using convolutional neural network. We evaluate the effectiveness of our approach by comparing it with the existing state of art methods using five real graph repositories from bioinformatics and social networks domains. Our results show that the Ngram approach outperforms existing methods with high accuracy and comparable performance.

With the wide deployment of cloud computing in many business enterprises as well as science and engineering domains, high quality security services are increasingly critical for processing workflow applications with sensitive intermediate data. Unfortunately, most existing worklfow scheduling approaches disregard the security requirements of the intermediate data produced by workflows, and overlook the performance impact of encryption time of intermediate data on the start of subsequent workflow tasks. Furthermore, the idle time slots on resources, resulting from data dependencies among workflow tasks, have not been adequately exploited to mitigate the impact of data encryption time on workflows' makespans and monetary cost. To address these issues, this paper presents a novel task-scheduling framework for security sensitive workflows with three novel features. First, we provide comprehensive theoretical analyses on how selectively duplicating a task's predecessor tasks is helpful for preventing both the data transmission time and encryption time from delaying task's start time. Then, we define workflow tasks' latest finish time, and prove that tasks can be completed before tasks' latest finish time by using cheapest resources to reduce monetary cost without delaying tasks' successors' start time and workflows' makespans. Based on these analyses, we devise a novel scheduling approach with selective tasks duplication, named SOLID, incorporating two important phases: 1) task scheduling with selectively duplicating predecessor tasks to idle time slots on resources; and 2) intermediate data encrypting by effectively exploiting tasks' laxity time. We evaluate our solution approach through rigorous performance evaluation study using both randomly generated workflows and some real-world workflow traces. Our results show that the proposed SOLID approach prevails over existing algorithms in terms of makespan, monetary costs and resource efficiency.

As advanced computing and communication technologies penetrate every aspect of our life, we have witnessed the persistent growth of open systems where entities interact with one another without prior knowledge or experiences. Trust becomes an important metric in such open systems. This paper presents a dependable trust management scheme-GroupTrust, and a working system to support GroupTrust. It makes three original contributions. First, we identify a set of vulnerabilities that are common in existing reputation based trust models. We show that reputation trust built solely on direct experiences or by combining direct experiences with uniform trust propagation can be vulnerable. Second, we develop GroupTrust, a dependable trust management scheme to provide reliable trust management in the presence of dishonest ratings, malicious camouflage, and malicious collusive behaviors. The GroupTrust scheme is novel in two aspects: (i) we develop a pairwise similarity based feedback credibility to enhance the resilience of trust computation in the presence of dishonest ratings; (ii) we propose to propagate trust based on a Susceptible-Infected-Recovered (SIR) model, which defines trust propagation threshold to control how trust should be propagated. Finally, we evaluate the effectiveness of GroupTrust against fourthreat models using both simulated and real world datasets. Our experimental results show that feedback credibility based local trust computation can effectively constrain strategically malicious participants from taking advantages of their dishonest ratings. SIR-based trust propagation control enables safe trust propagation and blocks irrational trust propagation. We show that GroupTrust scheme significantly outperforms other trust models in terms of both performance and attack resilience in the presence of dishonest feedbacks, sparse feedbacks, and strategically malicious participants against four representative threat models.

With the globalization of service, organizations continuously produce large volumes of data that need to be analysed over geo-dispersed locations. Traditionally central approach that moving all data to a single cluster is inefficient or infeasible due to the limitations such as the scarcity of wide-area bandwidth and the low latency requirement of data processing. Processing big data across geo-distributed datacenters continues to gain popularity in recent years. However, managing distributed MapReduce computations across geo-distributed datacenters poses a number of technical challenges: how to allocate data among a selection of geo-distributed datacenters to reduce the communication cost, how to determine the Virtual Machine (VM) provisioning strategy that offers high performance and low cost, and what criteria should be used to select a datacenter as the final reducer for big data analytics jobs. In this paper, these challenges is addressed by balancing bandwidth cost, storage cost, computing cost, migration cost, and latency cost, between the two MapReduce phases across datacenters. We formulate this complex cost optimization problem for data movement, resource provisioning and reducer selection into a joint stochastic integer nonlinear optimization problem by minimizing the five cost factors simultaneously. The Lyapunov framework is integrated into our study and an efficient online algorithm that is able to minimize the long-term time-averaged operation cost is further designed. Theoretical analysis shows that our online algorithm can provide a near optimum solution with a provable gap and can guarantee that the data processing can be completed within pre-defined bounded delays. Experiments on WorldCup98 web site trace validate the theoretical analysis results and demonstrate that our approach is close to the offline-optimum performance and superior to some representative approaches.

Presents the introductory editorial for this issue of the publication.

Scheduling real-time workflows running in the Cloud often need to deal with uncertain task execution time and minimize uncertainty propagation during the workflow runtime. Efficient scheduling approaches can minimize the operational cost of Cloud providers and provide higher guarantee of the quality of services (QoS) for Cloud consumers. However, most of the existing workflow scheduling approaches are designed for the individual workflow runtime environments that are deterministic. Such static workflow schedulers are inadequate for multiple and dynamic workflows, each with possibly uncertain task execution time. In this paper, we address the problem of minimizing uncertainty propagation in real-time workflow scheduling. We first introduce an uncertainty-aware scheduling architecture to mitigate the impact of uncertainty factors on the quality of workflow schedules. Then we present a dynamic workflow scheduling algorithm (PRS) that can dynamically exploit proactive and reactive scheduling methods. Finally, we conduct extensive experiments using real-world workflow traces and our experimental results show that PRS outperforms two representative scheduling algorithms in terms of costs (up to 60%), resource utilization (up to 40%) and deviation (up to 70%).

Cloud computing has successfully enabled large scale computing to be offered as pay-as-you-go services to many enterprise and individual tenants. However, the trust on public cloud services has been a sensitive issue for both cloud tenants and cloud service providers (CSPs). Tenants tend to worry about losing the total control over their codes and data hosted on remote servers. Public cloud providers often fear that the applications uploaded by their tenants may carry vicious codes, which may cause serious violations of security and privacy on their cloud platforms. This trust issue has slowed down the wide deployment of public clouds and hindered the promises of cloud computing for both CSPs and Cloud consumers. In this paper, we present Ta-TCS, a novel system framework for two-phase tenants attested trust validation and trust management over their remote VMs and cloud service executions. At the CSP end, we build a Minimal Trusted Environment (MTE) in VMM and an Integrity Verification & Report Service (IVRS) hosted in the control domain Dom0. At the tenant end, we deploy an Integrity Configuration and Attestation Service (ICAS) in new framework. With Ta-TCS, tenants can configure and attest the integrity of their services, and Cloud providers can verify codes running on a guest VM by introspection. Tenants can also check whether the basic platform of Dom0 is trusted or not. This two phase trust establishment increases the level of mutual trust between tenants and its CSP. We implement the first prototype system of Ta-TCS on Xen platform, and most of our implementation mechanisms can be deployed to some open-source virtualization platforms such as KVM. Our evaluation results show that Ta-TCS is effective with negligible performance overhead.

With the proliferation of mobile devices and the improvement of wireless communication technology, an increasing number of mobile devices are utilized for emergency management and healthcare monitoring. Redundant data upload to the cloud datacenters is gaining growing interest and attraction. One of the main challenges for redundant data upload in the cooperative mobile cloud is the optimization problem of how to provide high utility and high energy efficiency for data upload in the presence of intermittent connectivity and unpredictable bandwidth of wireless and mobile network. In this paper, we formulate the problem of redundant data upload in the cooperative mobile cloud as an energy-constrained utility maximization problem that aims at maximizing the amount of effective data uploaded under the energy consumption constraints. We propose an online distributed approach to enabling mobile devices to optimally make upload decisions without depending on the current state information of other devices and the prior knowledge of its own future context. We provide a rigorous theoretical analysis and an extensive suite of simulation experiments to demonstrate the effectiveness and superiority of our approach.

Summary form only given. Big graphs are finding increasing applications in many science and engineering domains, such as computational biology, cybermanufacturing and social media. Graphs provide a very flexible mathematical abstraction for describing relationships between entities in complex systems. Real world graphs are characterized by high connectivity and high irregularity. Such non-uniform characteristics increase the mismatch between the vertex centric parallel computation model and the computer hardware resources. Another problem with the vertex-centric computation model is that it treats vertices symmetrically and this uniform assumption breaks when graphs exhibit high irregularity and graph algorithms reveal non-uniform workloads. In this keynote, I will advocate a fundamental revisit of graph computation models and promotes a methodical framework for support high performance graph parallel abstractions that are resource aware, composable and programmable. I will discuss a suite of graph optimization techniques that explore workload characteristics of graph algorithms and irregularity hidden in graph structures. I will conclude the talk by presenting some interesting research problems and unique opportunities for big graph analytics.

Clouds have become an important platform to deliver services for various applications. Task scheduling and resource provisioning are key components to improve system performance under provisioned resources and satisfy users' demands for quality of service (QoS). To address the diversity of cloud services and applications, much of recent research and development efforts have been engaged in designing and implementing scheduling strategies and algorithms for specific tasks, such as dependent or independent tasks, fault-tolerant tasks with real-time deadlines or energy-efficient tasks. However, these task scheduling and resource provisioning schemes, though optimized with specific objectives, suffer from several inherent problems in cloud execution environments. In this paper, we propose a general framework for task scheduling and resource provisioning in cloud computing systems with dynamic customizability. By utilizing software engineering framework as the design guideline, we incorporate multiple scheduling objectives and multiple types of tasks to be processed under varied resource constraints to enable cloud applications to dynamically select and assemble scheduling strategies and algorithms according to different runtime QoS requirements. We illustrate the flexibility and customizability of our framework through two example scheduling algorithms: EASU and RAS. We validate the ffectiveness of our proposed framework through experimental evaluation of the effectiveness of our proposed algorithms using both simulation and in real cloud platforms.

Outsource encrypted data is a popular trend for storing sensitive data in third party clouds. Many cloud applications need privacy preserving data encryption services with two capabilities: On one hand, they need querying over encrypted data in Web based data hosting services. On the other hand, they also need to keep the query keywords and associated search operations private such that data hosting service providers cannot gain access to unauthorized content or trace and infer sensitive data stored in the third party data hosting servers. In this paper we present a novel service oriented framework for verifiable searchable asymmetric encryption, called PVSAE. PVSAE offers strong support for outsourced encrypted data with two formal security properties in terms of IND-CKA security and search pattern privacy. Our framework supports two concrete PVSAE schemes. The first scheme l-PVSAE is based on the l-dimensional vectors and achieves strong security notions, namely statistical IND-CKA security and statistical search pattern privacy. The second scheme 3-PVSAE is a light-weight version based on 3-dimensional vectors. 3-PVSAE maintains the strong security properties and offers higher efficiency for search over encrypted data compared with existing verifiable searchable asymmetric encryption schemes. We experimentally evaluate the proposed PVSAE schemes and show that they not only offer strong security but also are practical and deployable.

NoSQL systems are deployed as the core components for delivering big data Web services today. With growing main memory capacity, we witness the growing interest and deployment of in-memory NoSQL services (IM-NoSQL), which are designed to maximize the utilization of DRAM for ultra low latency services. To address the volatility of DRAM for in-memory computing services, persistence and failure recovery are important functionality for IM-NoSQL. In this paper we report an extensive measurement study on the performance of persistence and recovery for IM-NoSQL. We evaluate the performance and effectiveness of several common mechanisms used for persistence and recovery in the presence of server crashes, such as snapshot and logging based approaches. Through this study, we are able to answer some of the most frequently asked questions in provisioning of IM-NoSQL services: (i) Can an IM-NoSQL system work effectively when the available memory is insufficient to load the whole dataset? (ii) What is the overhead of maintaining snapshot compared to logging? (iii) How fast an IM-NoSQL system can recover in the presence of failure? And (iv) how does an IM-NoSQL system respond to the different persistence models? We report our comprehensive measurement results on execution, persistence and recovery performance of Redis, a representative implementation of IM-NoSQL services.

The design of maintenance mechanisms ofdistributed hash tables (DHTs) is usually specific totheir initial graphs, and thus it is complicated anderror-prone. Zhang and Liu propose in [4] the "distributedline graphs" (DLG) mechanism, a universaltechnique for designing DHTs based on arbitraryregular graphs while preserving the main features ofthe initial graphs. However, two important propertiesof DLG, the expandability and fidelity, have notbeen studied with detailed explanations or analysis.In this paper, we study the above two properties ofDLG transformations, and prove that (i) the DLGtransformations are incrementally expandable, and(ii) the DLG transformations from Gi to Gi+1 keepfidelity.

Dynamic VM memory management via the balloon driver is a common strategy to manage the memory resources of VMs under changing workloads. However, current approaches rely on kernel instrumentation to estimate the VM working set size, which usually result in high run-time overhead. Thus system administrators have to tradeoff between the estimation accuracy and the system performance. This paper presents iBalloon, a light-weight, accurate and transparent prediction based mechanism to enable more customizable and efficient ballooning policies for rebalancing memory resources among VMs. Experiment results from well known benchmarks such as Dacapo and SPECjvm show that iBalloon is able to quickly react to the VM memory demands, provide up to 54% performance speedup for memory intensive applications running in the VMs, while incurring less than 5% CPU overhead on the host machine as well as the VMs.

Utilizing graph analysis models and algorithms to exploit complex interactions over a network of entities is emerging as an attractive network analytic technology. In this paper, we show that traditional column or row-based trace analysis may not be effective in deriving deep insights hidden in the storage traces collected over complex storage applications, such as complex spatial and temporal patterns, hotspots and their movement patterns. We propose a novel graph analytics framework, GraphLens, for mining and analyzing real world storage traces with three unique features. First, we model storage traces as heterogeneous trace graphs in order to capture multiple complex and heterogeneous factors, such as diverse spatial/temporal access information and their relationships, into a unified analytic framework. Second, we employ and develop an innovative graph clustering method that employs two levels of clustering abstractions on storage trace analysis. We discover interesting spatial access patterns and identify important temporal correlations among spatial access patterns. This enables us to better characterize important hotspots and understand hotspot movement patterns. Third, at each level of abstraction, we design a unified weighted similarity measure through an iterative dynamic weight learning algorithm. With an optimal weight assignment scheme, we can efficiently combine the correlation information for each type of storage access patterns, such as random versus sequential, read versus write, to identify interesting spatial/temporal correlations hidden in the traces. Some optimization techniques on matrix computation are proposed to further improve the efficiency of our clustering algorithm on large trace datasets. Extensive evaluation on real storage traces shows GraphLens can provide broad and deep trace analysis for better storage strategy planning and efficient data placement guidance. GraphLens can be applied to both a single PC with multiple disks and a distributed network across a cluster of compute nodes to offer a few opportunities for optimization of storage performance.

This paper presents the design and implementation of MemPipe, a dynamic shared memory management system for high performance network I/O among virtual machines (VMs) located on the same host. MemPipe delivers efficient inter-VM communication with three unique features. First, MemPipe employs an inter-VM shared memory pipe to enable high throughput data delivery for both TCP and UDP workloads among co-located VMs. Second, instead of static allocation of shared memories, MemPipe manages its shared memory pipes through a demand driven and proportional memory allocation mechanism, which can dynamically enlarge or shrink the shared memory pipes based on the demand of the workloads in each VM. Third but not the least, MemPipe employs a number of optimizations, such as time-window based streaming partitions and socket buffer redirection, to further optimize its performance. Extensive experiments show that MemPipe improves the throughput of conventional (native) inter VM communication by up to 45 times, reduces the latency by up to 62 percent, and achieves up to 91 percent shared memory utilization.

Road network-aware spatial alarms extend the concept of time-based alarms to spatial dimension and remind us when we travel on spatially constrained road networks and enter some predefined locations of interest in the future. This paper argues that road network-aware spatial alarms need to be processed by taking into account spatial constraints on road networks and mobility patterns of mobile subscribers. We show that the Euclidian distance-based spatial alarm processing techniques tend to incur high client energy consumption due to unnecessarily frequent client wakeups. We design and develop a road network-aware spatial alarm processing system, called ROADALARM, with three unique features. First, we introduce the concept of road network-based spatial alarms using road network distance measures. Instead of using a rectangular region, a road network-aware spatial alarm is a star-like subgraph with an alarm target as the center of the star and border points as the scope of the alarm region. Second, we describe a baseline approach for spatial alarm processing by exploiting two types of filters. We use subscription filter and Euclidean lower bound filter to reduce the amount of shortest path computations required in both computing alarm hibernation time and performing alarm checks at the server. Last but not the least, we develop a suite of optimization techniques using motion-aware filters, which enable us to further increase the hibernation time of mobile clients and reduce the frequency of wakeups and alarm checks, while ensuring high accuracy of spatial alarm processing. Our experimental results show that the road network-aware spatial alarm processing significantly outperforms existing Euclidean space-based approaches, in terms of both the number of wakeups and the hibernation time at mobile clients and the number of alarm checks at the server.

Querying over encrypted data is gaining increasing popularity in cloud-based data hosting services. Security and efficiency are recognized as two important and yet conflicting requirements for querying over encrypted data. In this article, we propose an efficient private keyword search (EPKS) scheme that supports binary search and extend it to dynamic settings (called DEPKS ) for inverted index--based encrypted data. First, we describe our approaches of constructing a searchable symmetric encryption (SSE) scheme that supports binary search. Second, we present a novel framework for EPKS and provide its formal security definitions in terms of plaintext privacy and predicate privacy by modifying Shen et al.’s security notions [Shen et al. 2009]. Third, built on the proposed framework, we design an EPKS scheme whose complexity is logarithmic in the number of keywords. The scheme is based on the groups of prime order and enjoys strong notions of security, namely statistical plaintext privacy and statistical predicate privacy. Fourth, we extend the EPKS scheme to support dynamic keyword and document updates. The extended scheme not only maintains the properties of logarithmic-time search efficiency and plaintext privacy and predicate privacy but also has fewer rounds of communications for updates compared to existing dynamic search encryption schemes. We experimentally evaluate the proposed EPKS and DEPKS schemes and show that they are significantly more efficient in terms of both keyword search complexity and communication complexity than existing randomized SSE schemes.

Large scale iterative graph computation presents an interesting systems challenge due to two well known problems: (1) the lack of access locality and (2) the lack of storage efficiency. This paper presents PathGraph, a system for improving iterative graph computation on graphs with billions of edges. First, we improve the memory and disk access locality for iterative computation algorithms on large graphs by modeling a large graph using a collection of tree-based partitions. This enables us to use path-centric computation rather than vertex-centric or edge-centric computation. For each tree partition, we re-label vertices using DFS in order to preserve consistency between the order of vertex ids and vertex order in the paths. Second, a compact storage that is optimized for iterative graph parallel computation is developed in the PathGraph system. Concretely, we employ delta-compression and store tree-based partitions in a DFS order. By clustering highly correlated paths together as tree based partitions, we maximize sequential access and minimize random access on storage media. Third but not the least, our path-centric computation model is implemented using a scatter/gather programming model. We parallel the iterative computation at partition tree level and perform sequential local updates for vertices in each tree partition to improve the convergence speed. To provide well balanced workloads among parallel threads at tree partition level, we introduce the concept of multiple stealing points based task queue to allow work stealings from multiple points in the task queue. We evaluate the effectiveness of PathGraph by comparing with recent representative graph processing systems such as GraphChi and X-Stream etc. Our experimental results show that our approach outperforms the two systems on a number of graph algorithms for both in-memory and out-of-core graphs. While our approach achieves better data balance and load balance, it also shows better speedup than the two systems with the growth of threads.

Clouds are becoming an important platform for scientific workflow applications. However, with many nodes being deployed in clouds, managing reliability of resources becomes a critical issue, especially for the real-time scientific workflow execution where deadlines should be satisfied. Therefore, fault tolerance in clouds is extremely essential. The PB (primary backup) based scheduling is a popular technique for fault tolerance and has effectively been used in the cluster and grid computing. However, applying this technique for real-time workflows in a virtualized cloud is much more complicated and has rarely been studied. In this paper, we address this problem. We first establish a real-time workflow fault-tolerant model that extends the traditional PB model by incorporating the cloud characteristics. Based on this model, we develop approaches for task allocation and message transmission to ensure faults can be tolerated during the workflow execution. Finally, we propose a dynamic fault-tolerant scheduling algorithm, FASTER, for realtime workflows in the virtualized cloud. FASTER has three key features: 1) it employs a backward shifting method to make full use of the idle resources and incorporates task overlapping and VM migration for high resource utilization, 2) it applies the vertical/horizontal scaling-up technique to quickly provision resources for a burst of workflows, and 3) it uses the vertical scaling-down scheme to avoid unnecessary and ineffective resource changes due to fluctuated workflow requests. We evaluate our FASTER algorithm with synthetic workflows and workflows collected from the real scientific and business applications and compare it with six baseline algorithms. The experimental results demonstrate that FASTER can effectively improve the resource utilization and schedulability even in the presence of node failures in virtualized clouds.

Presents the introductory editorial for this issue of the publication.

Presents the introductory editorial for this issue of the publication.

NoSQL systems have become the vital components to deliver big data services in the Cloud. However, existing NoSQL systems rely on experienced administrators to configure and tune the wide range of configurable parameters in order to achieve high performance. In this paper, we present a policy-driven configuration management system for NoSQL systems, called PCM. PCM can identify workload sensitive configuration parameters and capture the tuned parameters for different workloads as configuration policies. PCM also can be used to analyze the range of configuration parameters that may impact on the runtime performance of NoSQL systems in terms of read and write workloads. The configuration optimization recommended by PCM can enable NoSQL systems such as HBase to run much more efficiently than the default settings for both individual worker node and entire cluster in the Cloud. Our experimental results show that HBase under the PCM configuration outperforms the default configuration and some simple configurations on a range of workloads with offering significantly higher throughput.

Conventional private data publication schemes are targeted at publication of sensitive datasets with the objective of retaining as much utility as possible for statistical (aggregate) queries while ensuring the privacy of individuals' information. However, such an approach to data publishing is no longer applicable in shared multi-tenant cloud scenarios where users often have different levels of access to the same data. In this paper, we present a privacy-preserving data publishing framework for publishing large datasets with the goals of providing different levels of utility to the users based on their access privileges. We design and implement our proposed multi-level utility-controlled data anonymization schemes in the context of large association graphs considering three levels of user utility namely: (i) users having access to only the graph structure (ii) users having access to graph structure and aggregate query results and (iii) users having access to graph structure, aggregate query results as well as individual associations. Our experiments on real large association graphs show that the proposed techniques are effective, scalable and yield the required level of privacy and utility for user-specific utility and access privilege levels.

Shared memory management is widely recognized as an optimization technique in the virtualized cloud. Most current shared memory techniques allocate shared memory resources from guest VMs based on pre-defined system configurations. Such static management of shared memory not only increases the VM memory pressure, but also limits the flexibility to balance shared memory resources across multiple VMs running on a single host. In this paper, we present a dynamic shared memory management framework that enables multiple VMs to dynamically access shared memory resources according to their demands. We illustrate our system design through two case studies: One aims to improve the performance of inter-domain communication while the other aims to improve VM memory swapping efficiency. We demonstrate that the dynamic shared memory mechanism not only improves the utilization of shared memory resources but also significantly enhances the performance of VM applications. Our experimental results show that by using dynamic shared memory management, we can improve the performance of inter-VM communication by up to 45 times, while mitigating the VM memory swapping overhead by up to 58%.

The recent shift towards write-intensive workload on big data (e.g., financial trading, social user-generated data streams)has pushed the proliferation of log-structured key-value stores, represented by Google's BigTable [1], Apache HBase [2] andCassandra [3]. While providing key-based data access with aPut/Get interface, these key-value stores do not support value-based access methods, which significantly limits their applicability in modern web and database applications. In this paper, we present DELI, a DEferred Lightweight Indexing scheme on the log-structured key-value stores. To index intensively updated bigdata in real time, DELI aims at making the index maintenance as lightweight as possible. The key idea is to apply an append-only design for online index maintenance and to collect index garbage at carefully chosen time. DELI optimizes the performance of index garbage collection through tightly coupling its execution with a native routine process called compaction. The DELI's system design is fault-tolerant and generic (to most key-valuestores), we implemented a prototype of DELI based on HBase without internal code modification. Our experiments show that the DELI offers significant performance advantage for the write-intensive index maintenance.

Iterative computation on large graphs has challenged system research from two aspects: (1) how to conduct high performance parallel processing for both in-memory and out-of-core graphs; and (2) how to handle large graphs that exceed the resource boundary of traditional systems by resource aware graph partitioning such that it is feasible to run large-scale graph analysis on a single PC. This paper presents GraphLego, a resource adaptive graph processing system with multi-level programmable graph parallel abstractions. GraphLego is novel in three aspects: (1) we argue that vertex-centric or edge-centric graph partitioning are ineffective for parallel processing of large graphs and we introduce three alternative graph parallel abstractions to enable a large graph to be partitioned at the granularity of subgraphs by slice, strip and dice based partitioning; (2) we use dice-based data placement algorithm to store a large graph on disk by minimizing non-sequential disk access and enabling more structured in-memory access; and (3) we dynamically determine the right level of graph parallel abstraction to maximize sequential access and minimize random access. GraphLego can run efficiently on different computers with diverse resource capacities and respond to different memory requirements by real-world graphs of different complexity. Extensive experiments show the competitiveness of GraphLego against existing representative graph processing systems, such as GraphChi, GraphLab and X-Stream.

Most existing solutions on location anonymization fail to address the issue of location privacy protection for mobile users traveling on road networks. In this paper, we present a road network-aware privacy model to handle the location privacy problem in mobile systems. We present third party supported anonymization which can be plugged into existing mobile systems without major modifications. As opposed to current road network specific solutions, our system has the following salient features. First, we argue that the graph density of a cloaked subgraph is an important location cloaking quality measure for road network-based location privacy protection. High graph density leads to high utility of the cloaked location with respect to spatial resolution. Second, we devise a suite of graph-based cloaking algorithms which guarantee reciprocity - an important location cloaking property that many existing approaches fail to support - under an enhanced privacy model. Further, we use controlled randomization in the cloaking process to provide higher privacy strength while maintaining the utility of the cloaked location. Last but not the least, our experimental evaluation shows the efficiency and robustness of the approach in terms of privacy, utility and performance of the model.

The emerging need for conducting complex analysis over big RDF datasets calls for scale-out solutions that can harness a computing cluster to process big RDF datasets. Queries over RDF data often involve complex self-joins, which would be very expensive to run if the data are not carefully partitioned across the cluster and hence distributed joins over massive amount of data are necessary. Existing RDF data partitioning methods can nicely localize simple queries but still need to resort to expensive distributed joins for more complex queries. In this paper, we propose a new data partitioning approach that takes use of the rich structural information in RDF datasets and minimizes the amount of data that have to be joined across different computing nodes. We conduct an extensive experimental study using two popular RDF benchmark data and one real RDF dataset that contain up to billions of RDF triples. The results indicate that our approach can produce a balanced and low redundant data partitioning scheme that can avoid or largely reduce the cost of distributed joins even for very complicated queries. In terms of query execution time, our approach can outperform the state-of-the-art methods by orders of magnitude.

On a crowd sourcing platform consisting of task publishers and workers, it is critical for a task publisher to select trustworthy workers to solve human intelligence tasks (HITs). Currently, the prevalent trust evaluation mechanism employs the overall approval rate of HITs, with which dishonest workers can easily succeed in pursuing the maximal profit by quickly giving plausible answers or counterfeiting HITs approval rates. In crowd sourcing environments, a worker's trustworthiness varies in contexts, i.e. It varies in different types of tasks and different reward amounts of tasks. Thus, we propose two classifications based on task types and task reward amount respectively. On the basis of the classifications, we propose a trust evaluation model, which consists of two types of context-aware trust: task type based trust (TaTrust) and reward amount based trust (RaTrust). Then, we model trustworthy worker selection as a multi-objective combinatorial optimization problem, which is NP-hard. For solving this challenging problem, we propose an evolutionary algorithm MOWS_GA based on NSGA-II. The results of experiments illustrate that our proposed trust evaluation model can effectively differentiate honest workers and dishonest workers when both of them have high overall HITs approval rates.

Many popular web service networks are content-rich in terms of heterogeneous types of entities and links, associated with incomplete attributes. Clustering such heterogeneous service networks demands new clustering techniques that can handle two heterogeneity challenges: (1) multiple types of entities co-exist in the same service network with multiple attributes, and (2) links between entities have diverse types and carry different semantics. Existing heterogeneous graph clustering techniques tend to pick initial centroids uniformly at random, specify the number k of clusters in advance, and fix k during the clustering process. In this paper, we propose Service Cluster, a novel heterogeneous service network clustering algorithm with four unique features. First, we incorporate various types of entity, attribute and link information into a unified distance measure. Second, we design a Discrete Steepest Descent method to naturally produce initial k and initial centroids simultaneously. Third, we propose a dynamic learning method to automatically adjust the link weights towards clustering convergence. Fourth, we develop an effective optimization strategy to identify new suitable k and k well-chosen centroids at each clustering iteration. Extensive evaluation on real datasets demonstrates that Service Cluster outperforms existing representative methods in terms of both effectiveness and efficiency.

Meta paths are good mechanisms to improve the quality of graph analysis on heterogeneous information networks. This paper presents a meta path graph clustering framework, VEPATHCLUSTER, that combines meta path vertex-centric clustering with meta path edge-centric clustering for improving the clustering quality of heterogeneous networks. First, we propose an edge-centric path graph model to capture the meta-path dependencies between pairwise path edges. We model a heterogeneous network containing M types of meta paths as M vertex-centric path graphs and M edge-centric path graphs. Second, we propose a clustering-based multigraph model to capture the fine-grained clustering-based relationships between pairwise vertices and between pairwise path edges. We perform clustering analysis on both a unified vertex-centric path graph and each edge-centric path graph to generate vertex clustering and edge clusterings of the original heterogeneous network respectively. Third, a reinforcement algorithm is provided to tightly integrate vertex-centric clustering and edge-centric clustering by mutually enhancing each other. Finally, an iterative learning strategy is presented to dynamically refine both vertex-centric clustering and edge-centric clustering by continuously learning the contributions and adjusting the weights of different path graphs.

The open nature of recommender systems makes them vulnerable to shilling attacks. Biased ratings are introduced in order to affect recommendations, have been shown to cause great harm to collaborative filtering algorithms. Most of previous research focuses on the differences between genuine profiles and attack profiles, ignoring the group characteristics in an attack. There exists class unbalance problems in SVM based detecting methods, that is, the detecting performance is not good when the amount of samples of attack profiles in training set is small. In this paper, we study the use of SVM based method and group characteristics in attack profiles to detect attack profiles. Based on this, a two phase detecting method SVM-TIA is proposed. In the first phase, Borderline-SMOTE method is used to alleviate the class unbalance problem in classification; a rough detecting result is obtained in this phase; the second phase is a fine-tuning phase whereby the target items in the potential attack profiles set are analysed. We conduct experiments on the MovieLens 100K Dataset and compare the performance of SVM-TIA with other shilling detecting methods to demonstrate the effectiveness of the proposed approach.

In recent years, systems researchers have devoted considerable effort to the study of large-scale graph processing. Existing distributed graph processing systems such as Pregel, based solely on distributed memory for their computations, fail to provide seamless scalability when the graph data and their intermediate computational results no longer fit into the memory; and most distributed approaches for iterative graph computations do not consider utilizing secondary storage a viable solution. This paper presents GraphMap, a distributed iterative graph computation framework that maximizes access locality and speeds up distributed iterative graph computations by effectively utilizing secondary storage. GraphMap has three salient features: (1) It distinguishes data states that are mutable during iterative computations from those that are read-only in all iterations to maximize sequential access and minimize random access. (2) It entails a two-level graph partitioning algorithm that enables balanced workloads and locality-optimized data placement. (3) It contains a proposed suite of locality-based optimizations that improve computational efficiency. Extensive experiments on several real-world graphs show that GraphMap outperforms existing distributed memory-based systems for various iterative graph algorithms.

Large-scale real-world graphs are known to have highly skewed vertex degree distribution and highly skewed edge weight distribution. Existing vertex-centric iterative graph computation models suffer from a number of serious problems: (1) poor performance of parallel execution due to inherent workload imbalance at vertex level; (2) inefficient CPU resource utilization due to short execution time for low-degree vertices compared to the cost of in-memory or on-disk vertex access; and (3) incapability of pruning insignificant vertices or edges to improve the computational performance. In this paper, we address the above technical challenges by designing and implementing a scalable, efficient, and provably correct two-tier graph parallel processing system, GraphTwist. At storage and access tier, GraphTwist maximizes parallel efficiency by employing three graph parallel abstractions for partitioning a big graph by slice, strip or dice based partitioning techniques. At computation tier, GraphTwist presents two utility-aware pruning strategies: slice pruning and cut pruning, to further improve the computational performance while preserving the computational utility defined by graph applications. Theoretic analysis is provided to quantitatively prove that iterative graph computations powered by utility-aware pruning techniques can achieve a very good approximation with bounds on the introduced error.

Social influence analysis has shown great potential for strategic marketing decision. It is well known that people influence one another based on both their social connections and the social activities that they have engaged in the past. In this article, we develop an innovative and high-performance social influence based graph clustering framework with four unique features. First, we explicitly distinguish social connection based influence (self-influence) and social activity based influence (co-influence). We compute the self-influence similarity between two members based on their social connections within a single collaboration network, and compute the co-influence similarity by taking into account not only the set of activities that people participate but also the semantic association between these activities. Second, we define the concept of influence-based similarity by introducing a unified influence-based similarity matrix that employs an iterative weight update method to integrate self-influence and co-influence similarities. Third, we design a dynamic learning algorithm, called SI-C luster , for social influence based graph clustering. It iteratively partitions a large social collaboration network into K clusters based on both the social network itself and the multiple associated activity information networks, each representing a category of activities that people have engaged. To make the SI-C luster algorithm converge fast, we transform sophisticated nonlinear fractional programming problem with respect to multiple weights into a straightforward nonlinear parametric programming problem of single variable. Finally, we develop an optimization technique of diagonalizable-matrix approximation to speed up the computation of self-influence similarity and co-influence similarities. Our SI-Cluster-Opt significantly improves the efficiency of SI-Cluster on large graphs while maintaining high quality of clustering results. Extensive experimental evaluation on three real-world graphs shows that, compared to existing representative graph clustering algorithms, our SI-C luster -O pt approach not only achieves a very good balance between self-influence and co-influence similarities but also scales extremely well for clustering large graphs in terms of time complexity while meeting the guarantee of high density, low entropy and low Davies--Bouldin Index.

In emerging information networks, it is crucially important to provide efficient search on distributed documents while preserving their owners' privacy, for which privacy preserving indexes or PPI presents a possible solution. An understudied problem for the PPI techniques is how to provide differentiated privacy preservation in the presence of multi-keyword document search. The differentiation is necessary as terms and phrases bear innate differences in their semantic meanings. In this paper, we present ϵ-MPPI, the first work to provide the distributed document search with quantitatively differentiated privacy preservation. In the design of ϵ-MPPI, we identified a suite of challenging problems and proposed novel solutions. For one, we formulated the quantitative privacy computation as an optimization problem that strikes a balance between privacy preservation and search efficiency. We also addressed the challenging problem of secure ϵ-MPPI construction in the multi-domain information network which lacks mutual trusts between domains. Towards a secure ϵ-MPPIconstruction with practically acceptable performance, we proposed to optimize the performance of secure multi-party computations by making a novel use of secret sharing. We implemented the ϵ-MPPI construction protocol with a functioning prototype. We conducted extensive experiments to evaluate the prototype's effectiveness and efficiency based on a real-world dataset.

Mining trajectory data has been gaining significant interest in recent years. However, existing approaches to trajectory clustering are mainly based on density and Euclidean distance measures. We argue that when the utility of spatial clustering of mobile object trajectories is targeted at road-network aware location-based applications, density and Euclidean distance are no longer the effective measures. This is because traffic flows in a road network and the flow-based density characterization become important factors for finding interesting trajectory clusters. We propose NEAT-a road-network aware approach for fast and effective clustering of trajectories of mobile objects traveling in road networks. Our approach carefully considers the traffic locality characterized by the physical constraints of the road network, the traffic flow among consecutive road segments, and the flow-based density to organize trajectories into spatial clusters in a comprehensive three-phase clustering framework. NEAT discovers spatial clusters as groups of sub-trajectories which describe both dense and highly continuous flows of mobile objects. We perform extensive experiments with mobility traces generated using different scales of real road networks. Experimental results demonstrate the flexibility of the NEAT system and show that NEAT is highly accurate and runs orders of magnitude faster than existing density-based trajectory clustering approaches.

Continuous exposure of location information, even with spatially cloaked resolution, may lead to breaches of location privacy due to statistics-based inference attacks. An alternative and complementary approach to spatial cloaking based location anonymization is to break the continuity of location exposure by introducing techniques, such as mix-zones, where no application can trace user movements. Several factors impact on the effectiveness of mix-zone approach, such as user population, mix-zone geometry, location sensing rate and spatial resolution, as well as spatial and temporal constraints on user movement patterns. However, most of the existing mix-zone proposals fail to provide effective mix-zone construction and placement algorithms that are resilient to timing and transition attacks. This paper presents MobiMix, a road network based mix-zone framework to protect location privacy of mobile users traveling on road networks. It makes three original contributions. First, we provide the formal analysis on the vulnerabilities of directly applying theoretical rectangle mix-zones to road networks in terms of anonymization effectiveness and resilience to timing and transition attacks. Second, we develop a suite of road network mix-zone construction methods that effectively consider the above mentioned factors to provide higher level of resilience to timing and transition attacks, and yield a specified lowerbound on the level of anonymity. Third, we present a set of mix-zone placement algorithms that identify the best set of road intersections for mix-zone placement considering the road network topology, user mobility patterns and road characteristics. We evaluate the MobiMix approach through extensive experiments conducted on traces produced by GTMobiSim on different scales of geographic maps. Our experiments show that MobiMix offers high level of anonymity and high level of resilience to timing and transition attacks, compared to existing mix-zone approaches.

This paper presents a new MapReduce cloud service model, Cura, for provisioning cost-effective MapReduce services in a cloud. In contrast to existing MapReduce cloud services such as a generic compute cloud or a dedicated MapReduce cloud, Cura has a number of unique benefits. First, Cura is designed to provide a cost-effective solution to efficiently handle MapReduce production workloads that have a significant amount of interactive jobs. Second, unlike existing services that require customers to decide the resources to be used for the jobs, Cura leverages MapReduce profiling to automatically create the best cluster configuration for the jobs. While the existing models allow only a per-job resource optimization for the jobs, Cura implements a globally efficient resource allocation scheme that significantly reduces the resource usage cost in the cloud. Third, Cura leverages unique optimization opportunities when dealing with workloads that can withstand some slack. By effectively multiplexing the available cloud resources among the jobs based on the job requirements, Cura achieves significantly lower resource usage costs for the jobs. Cura's core resource management schemes include cost-aware resource provisioning, VM-aware scheduling and online virtual machine reconfiguration. Our experimental results using Facebook-like workload traces show that our techniques lead to more than 80 percent reduction in the cloud compute infrastructure cost with upto 65 percent reduction in job response times.

Presents the introductory editorial for this issue of the publication

Due to the diversity and dynamic nature of cloud services, it is usually hard for potential cloud consumers to select the most suitable cloud service. This paper proposes CCCloud: a context-aware and credible cloud service selection model based on the comparison and aggregation of subjective assessments extracted from ordinary cloud consumers and objective assessments from quantitative performance testing parties. We propose a novel approach to evaluate cloud users' credibility, which not only can accurately evaluate how truthfully they assess cloud services, but also resist user collusion. In addition, in our model, objective assessments are used as benchmarks to filter out potentially biased subjective assessments, and then objective assessments and subjective assessments are aggregated to evaluate the overall performance of a cloud service. Furthermore, our model takes the contexts of objective assessments and subjective assessments into account. By calculating the similarity between different contexts, the benchmark level of objective assessments is dynamically adjusted according to context similarity, and the aggregated final scores of alternative cloud services are weighted by the similarity between the contexts of a potential cloud consumer and every testing party. This makes our cloud service selection model reflect potential cloud consumers' customized requirements more effectively. Finally, our proposed model is evaluated through the experiments conducted under different conditions. The experimental results demonstrate that our model significantly outperforms the existing work, especially in the resistance of user collusion.

Distributed service networks are popular platforms for service providers to offer services to consumers and for service consumers to acquire services from unknown parties. eBay and Amazon are two well-known examples of enabling and hosting such service networks to connect service providers to service consumers. Trust management is a critical component for scaling such distributed service networks to a large and growing number of participants. In this article, we present ServiceTrust ++ , a feedback quality--sensitive and attack resilient trust management scheme for empowering distributed service networks with effective trust management capability. Compared with existing trust models, ServiceTrust ++ has several novel features. First, we present six attack models to capture both independent and colluding attacks with malicious cliques, malicious spies, and malicious camouflages. Second, we aggregate the feedback ratings based on the variances of participants’ feedback behaviors and incorporate feedback similarity as weight into the local trust algorithm. Third, we compute the global trust of a participant by employing conditional trust propagation based on the feedback similarity threshold. This allows ServiceTrust ++ to control and prevent malicious spies and malicious camouflage peers from boosting their global trust scores by manipulating the feedback ratings of good peers and by taking advantage of the uniform trust propagation. Finally, we systematically combine a trust-decaying strategy with a threshold value--based conditional trust propagation to further strengthen the robustness of our global trust computation against sophisticated malicious feedback. Experimental evaluation with both simulation-based networks and real network dataset Epinion show that ServiceTrust ++ is highly resilient against all six attack models and highly effective compared to EigenTrust, the most popular and representative trust propagation model to date.

With more data generated and collected in a geographically distributed manner, combined by the increased computational requirements for large scale data-intensive analysis, we have witnessed the growing demand for geographically distributed Cloud datacenters and hybrid Cloud service provisioning, enabling organizations to support instantaneous demand of additional computational resources and to expand inhouse resources to maintain peak service demands by utilizing cloud resources. A key challenge for running applications in such a geographically distributed computing environment is how to efficiently schedule and perform analysis over data that is geographically distributed across multiple datacenters. In this paper, we first compare multi-datacenter Hadoop deployment with single-datacenter Hadoop deployment to identify the performance issues inherent in a geographically distributed cloud. A generalization of the problem characterization in the context of geographically distributed cloud datacenters is also provided with discussions on general optimization strategies. Then we describe the design and implementation of a suite of system-level optimizations for improving performance of Hadoop service provisioning in a geo-distributed cloud, including prediction-based job localization, configurable HDFS data placement, and data prefetching. Our experimental evaluation shows that our prediction based localization has very low error ratio, smaller than 5%, and our optimization can improve the execution time of Reduce phase by 48.6%.

Hybrid clouds, geo-distributed cloud and continuous upgrades of computing, storage and networking resources in the cloud have driven datacenters evolving towards heterogeneous clusters. Unfortunately, most of MapReduce implementations are designed for homogeneous computing environments and perform poorly in heterogeneous clusters. Although a fair of research efforts have dedicated to improve MapReduce performance, there still lacks of in-depth understanding of the key factors that affect the performance of MapReduce jobs in heterogeneous clusters. In this paper, we present an extensive experimental study on two categories of factors: system configuration and task scheduling. Our measurement study shows that an in-depth understanding of these factors is critical for improving MapReduce performance in a heterogeneous environment. We conclude with five key findings: (1) Early shuffle, though effective for reducing the latency of MapReduce jobs, can impact the performance of map tasks and reduce tasks differently when running on different types of nodes. (2) Two phases in map tasks have different sensitive to input block size and the ratio of sort phase with different block size is different for different type of nodes. (3) Scheduling map or reduce tasks dynamically with node capacity and workload awareness can further enhance the job performance and improve resource consumption efficiency. (4) Although random scheduling of reduce tasks works well in homogeneous clusters, it can significantly degrade the performance in heterogeneous clusters when shuffled data size is large. (5) Phase-aware progress rate estimation and speculation strategy can provide substantial performance gain over the state of art speculation scheduler.

Data outsourcing offers cost-effective computing power to manage massive data streams and reliable access to data. Data owners can forward their data to clouds, and the clouds provide data mirroring, backup, and online access services to end users. However, outsourcing data to untrusted clouds requires data authenticity and query integrity to remain in the control of the data owners and users.

Public storage clouds have become a popular choice for archiving certain classes of enterprise data - for example, application and infrastructure logs. These logs contain sensitive information like IP addresses or user logins due to which regulatory and security requirements often require data to be encrypted before moved to the cloud. In order to leverage such data for any business value, analytics systems (e.g. Hadoop/MapReduce) first download data from these public clouds, decrypt it and then process it at the secure enterprise site. We propose VNCache: an efficient solution for MapReduceanalysis of such cloud-archived log data without requiring an apriori data transfer and loading into the local Hadoop cluster. VNcache dynamically integrates cloud-archived data into a virtual namespace at the enterprise Hadoop cluster. Through a seamless data streaming and prefetching model, Hadoop jobs can begin execution as soon as they are launched without requiring any apriori downloading. With VNcache's accurate pre-fetching and caching, jobs often run on a local cached copy of the data block significantly improving performance. When no longer needed, data is safely evicted from the enterprise cluster reducing the total storage footprint. Uniquely, VNcache is implemented with NO changes to the Hadoop application stack.

The flexibility of the RDF data model has attracted an increasing number of organizations to store their data in an RDF format. With the rapid growth of RDF datasets, we envision that it is inevitable to deploy a cluster of computing nodes to process large-scale RDF data in order to deliver desirable query performance. In this paper, we address the challenging problems of data partitioning and query optimization in a scale-out RDF engine. We identify that existing approaches only focus on using fine-grained structural information for data partitioning, and hence fail to localize many types of complex queries. We then propose a radically different approach, where a coarse-grained structure, namely Rooted Sub-Graph (RSG), is used as the partition unit. By doing so, we can capture structural information at a much greater scale and hence are able to localize many complex queries. We also propose a k-means partitioning algorithm for allocating the RSGs onto the computing nodes as well as a query optimization strategy to minimize the inter-node communication during query processing. An extensive experimental study using benchmark datasets and real dataset shows that our engine, SemStore, outperforms existing systems by orders of magnitudes in terms of query response time.

NoSQL (Not only SQL) data stores become a vital component in many big data computing platforms due to its inherent horizontal scalability. HBase is an open-source distributed NoSQL store that is widely used by many Internet enterprises to handle their big data computing applications (e.g. Facebook h

Service provenance can be defined as a profile of service execution history. Most of the organizations today collect and manage their own service provenance in order to trace service execution failures, locate service bottlenecks, guide resource allocation, detect and prevent abnormal behaviors. T

Spatial queries are widely used in many data mining and analytics applications. However, a huge and growing size of spatial data makes it challenging to process the spatial queries efficiently. In this paper we present a lightweight and scalable spatial index for big data stored in distributed storage systems. Experimental results show the efficiency and effectiveness of our spatial indexing technique for different spatial queries.

In emerging information networks, having a privacy preserving index (or PPI) is critically important for locating information of interest for data sharing across autonomous providers while preserving privacy. An understudied problem for PPI techniques is how to provide controllable privacy preservation, given the innate difference of privacy concerns regarding different data owners. In this paper we present a personalized privacy preserving index, coined ε-PPI, which guarantees quantitative privacy preservation differentiated by personal identities. We devise a new common-identity attack that breaks existing PPI's and propose an identity-mixing protocol against the attack in ε-PPI. The proposed ε-PPI construction protocol is the first without any trusted third party and/or trust relationships between providers. We have implemented our ε-PPI construction protocol by using generic MPC techniques (secure multi-party computation) and optimized the performance to a practical level by minimizing the expensive MPC part.

In the age of big data, key-value data updated by intensive write streams is increasingly common, e.g., in social event streams. To serve such data in a cost-effective manner, a popular new paradigm is to outsource it to the cloud and store it in a scalable key-value store while serving a large user base. Due to the limited trust in third-party cloud infrastructures, data owners have to sign the data stream so that the data users can verify the authenticity of query results from the cloud. In this paper, we address the problem of verifiable freshness for multi-version key-value data. We propose a memory-resident digest structure that utilizes limited memory effectively and can have efficient verification performance. The proposed structure is named IncBM-Tree because it can INCrementally build a Bloom filter-embedded Merkle Tree. We have demonstrated the superior performance of verification under small memory footprints for signing, which is typical in an outsourcing scenario where data owners and users have limited resources.

This paper proposes a novel context-aware cloud service selection model based on the comparison and aggregation of subjective assessment extracted from cloud user feedback and objective assessment from quantitative performance testing. In this model, objective assessment provided by some professional testing parties is used as a benchmark to filter out potentially biased subjective assessment from cloud users, then objective assessment and subjective assessment are aggregated to evaluate the overall performance of cloud services according to potential cloud users' personalized requests. Moreover, our model takes the contexts of objective assessment and subjective assessment into account. By calculating the similarity between different contexts, the benchmark level of objective assessment is dynamically adjusted according to context similarity, which makes the following comparison and aggregation process more accurate and effective. After aggregation, the final results can quantitatively reflect the overall quality of cloud services. Finally, our proposed model is evaluated through the experiments executed in different conditions.

Multi-label classification of heterogeneous information networks has received renewed attention in social network analysis. In this paper, we present an activity-edge centric multi-label classification framework for analyzing heterogeneous information networks with three unique features. First, we model a heterogeneous information network in terms of a collaboration graph and multiple associated activity graphs. We introduce a novel concept of vertex-edge homophily in terms of both vertex labels and edge labels and transform a general collaboration graph into an activity-based collaboration multigraph by augmenting its edges with class labels from each activity graph through activity-based edge classification. Second, we utilize the label vicinity to capture the pairwise vertex closeness based on the labeling on the activity-based collaboration multigraph. We incorporate both the structure affinity and the label vicinity into a unified classifier to speed up the classification convergence. Third, we design an iterative learning algorithm, AEClass, to dynamically refine the classification result by continuously adjusting the weights on different activity-based edge classification schemes from multiple activity graphs, while constantly learning the contribution of the structure affinity and the label vicinity in the unified classifier. Extensive evaluation on real datasets demonstrates that AEClass outperforms existing representative methods in terms of both effectiveness and efficiency.

The continued explosion of Twitter data has opened doors for many applications, such as location-based advertisement and entertainment using smartphones. Unfortunately, only about 0.58 percent of tweets are geo-tagged to date. To tackle the location sparseness problem, this paper presents a methodic

Large scale graph processing represents an interesting challenge due to the lack of locality. This paper presents Path Graph for improving iterative graph computation on graphs with billions of edges. Our system design has three unique features: First, we model a large graph using a collection of tree-based partitions and use an path-centric computation rather than vertex-centric or edge-centric computation. Our parallel computation model significantly improves the memory and disk locality for performing iterative computation algorithms. Second, we design a compact storage that further maximize sequential access and minimize random access on storage media. Third, we implement the path-centric computation model by using a scatter/gather programming model, which parallels the iterative computation at partition tree level and performs sequential updates for vertices in each partition tree. The experimental results show that the path-centric approach outperforms vertex centric and edge-centric systems on a number of graph algorithms for both in-memory and out-of-core graphs.

ABSTRACT Security and privacy are widely recognized as important requirements for access and management of electronic health record (EHR) data. In this paper, we argue that EHR data need to be managed with customizable access control in both spatial and temporal dimensions. We present a role‐based and time‐bound access control (RBTBAC) model that provides more flexibility in both roles (spatial capability) and time (temporal capability) dimensions to control the access of sensitive data. Through algorithmic combination of role‐based access control and time‐bound key management, our RBTBAC model has two salient features. First, we have developed a privacy‐aware and dynamic key structure for role‐based privacy aware access and management of EHR data, focusing on the consistency of access authorization (including data and time interval) with the activated role of user. In addition to role‐based access, a path‐invisible EHR structure is built for preserving privacy of patients. Second, we have employed a time tree method for generating time granule values, offering fine granularity of time‐bound access authorization and control. Our initial experimental results show that tree‐like time structure can improve the performance of the key management scheme significantly, and RBTBAC model is more suitable than existing solutions for EHR data management because it offers high‐efficiency and better security and privacy. Copyright © 2013 John Wiley & Sons, Ltd.

Ontology reuse offers great benefits by measuring and comparing ontologies. However, the state of art approaches for measuring ontologies neglects the problems of both the polymorphism of ontology representation and the addition of implicit semantic knowledge. One way to tackle these problems is to devise a mechanism for ontology measurement that is stable, the basic criteria for automatic measurement. In this paper, we present a graph derivation representation based approach (GDR) for stable semantic measurement, which captures structural semantics of ontologies and addresses those problems that cause unstable measurement of ontologies. This paper makes three original contributions. First, we introduce and define the concept of semantic measurement and the concept of stable measurement. We present the GDR based approach, a three-phase process to transform an ontology to its GDR. Second, we formally analyze important properties of GDRs based on which stable semantic measurement and comparison can be achieved successfully. Third but not the least, we compare our GDR based approach with existing graph based methods using a dozen real world exemplar ontologies. Our experimental comparison is conducted based on nine ontology measurement entities and distance metric, which stably compares the similarity of two ontologies in terms of their GDRs.

In 2013, TSC expanded in a number of dimensions, including scope, editorial board, and contributing communities. First, TSC expanded its scope by introducing the new taxonomy of services computing (http://www.computer.org/ portal/web/tsc/taxonomy). This new taxonomy broadens the scope of TSC from the foundations of services computing to cover cloud computing services, big data technologies, the Internet of Things (IoT), social computing, and business management. This new taxonomy refl ects the continued advancements of services computing from both services science and services engineering perspectives. Second, TSC has expanded its editorial board to 43 associate editors in response to the broadening scope of TSC and the growing community of researchers and practitioners as authors and readers. More importantly, the TSC community continues to expand and attract both technical professionals and business management professionals. In recognition of its important role, TSC has been indexed by major indexing systems, such as Thomson Reuters (ISI), EI Compendex, and Science Citation Index Expanded (SCI-e). TSC welcomes 2014 with a number of new projects on the way. We are introducing survey papers as a featured section in TSC regular issues, in addition to special topic sections where guest editors are invited to solicit submissions on hot and emerging topics of special interests. We are planning to work with the steering committees of the IEEE International Conference on Web Services and the IEEE International Conference on Cloud Computing to create TSC special sessions. Authors of TSC accepted papers are offered an opt-in opportunity to present their work at one of these two conferences. The EiC is working closely with the TSC advisory board to explore new dimensions to attract higher readership and authorship from the business and management community. Finally, The EiC will continue to work with the TSC editorial board through multiple threads to maintain the average review cycle of three months and the average queuing time of less than one year for accepted papers to appear. One new thread of such effort is to continue TSC's participation in the newly introduced Reviewer Appreciation Program by the IEEE Computer Society.

Big data business can leverage and benefit from the Clouds, the most optimized, shared, automated, and virtualized computing infrastructures. One of the important challenges in processing big data in the Clouds is how to effectively partition the big data to ensure efficient distributed processing of the data. In this paper we present a Scalable and yet customizable data PArtitioning framework, called SPA, for distributed processing of big RDF graph data. We choose big RDF datasets as our focus of the investigation for two reasons. First, the Linking Open Data cloud has put forwards a good number of big RDF datasets with tens of billions of triples and hundreds of millions of links. Second, such huge RDF graphs can easily overwhelm any single server due to the limited memory and CPU capacity and exceed the processing capacity of many conventional data processing software systems. Our data partitioning framework has two unique features. First, we introduce a suite of vertexcentric data partitioning building blocks to allow efficient and yet customizable partitioning of large heterogeneous RDF graph data. By efficient, we mean that the SPA data partitions can support fast processing of big data of different sizes and complexity. By customizable, we mean that the SPA partitions are adaptive to different query types. Second, we propose a selection of scalable techniques to distribute the building block partitions across a cluster of compute nodes in a manner that minimizes inter-node communication cost by localizing most of the queries on distributed partitions. We evaluate our data partitioning framework and algorithms through extensive experiments using both benchmark and real datasets. Our experimental results show that the SPA data partitioning framework is not only efficient for partitioning and distributing big RDF datasets of diverse sizes and structures but also effective for processing big data queries of different types and complexity.

Network I/O workloads are dominating in many data centers and cloud computing environments today. One way to improve inter Virtual Machine (VM) communication efficiency is to support co-resident VM communication by using shared memory based approaches and to resort to the traditional TCP/IP for inter-VM communications between VMs that are located on different physical hosts. Although a number of independent efforts are dedicated to improving communication efficiency between co-resident VMs, they differ from one another in terms of how the inter-VM communication optimization is carried out and where in the software stack the shared memory channel is established. In this paper, we provide an in-depth overview of the design choices and techniques for optimizing the performance of the co-resident inter-VM communication, with dual objectives. First, we describe the core design guidelines and key issues for optimizing inter-VM communication by using shared memory based mechanisms. Typical issues include choices of implementation layer in the software stack, seamless agility for VM live migration and VM dynamic deployment support, multilevel transparency. Second, we conduct a comprehensive analysis of representative state-of-the-art research efforts and implementation techniques based on the core design guidelines. We also give an analysis of future requirements in advanced features such as reliability, security and stability. The research reported in this paper not only provides the reference for developing the next generation of inter-VM communication optimization mechanisms, but also offers opportunities for both cloud infrastructure providers and cloud service consumers to improve inter-VM communication efficiency in virtualized platforms.

A known problem for virtualized cloud data centers is the inter-VM communication inefficiency for data transfer between co-resident VMs. Several engineering efforts have been made on building a shared memory based channel between co-resident VMs. The implementations differ in terms of whether user/program transparency, OS kernel transparency or VMM transparency is supported. However, none of existing works has engaged in an in-depth measurement study with quantitative and qualitative analysis on performance improvements as well as tradeoffs introduced by such a residency-aware inter-VM communication mechanism. In this paper we present an extensive experimental study, aiming at addressing a number of fundamental issues and providing deeper insights regarding the design of a shared memory channel for co-resident VMs. Example questions include how much performance gains can a residency-aware shared memory inter-VM communication mechanism provide under different mixtures of local and remote network I/O workloads, what overhead will the residence-awareness detection and communication channel switch introduce over the remote inter-VM communication, what factors may exert significant impact on the throughput and latency performance of such a shared memory channel. We believe that this measurement study not only helps system developers to gain valuable lessons and generate new ideas to further improve the inter-VM communication performance. It also offers new opportunities for cloud service providers to deploy their services more efficiently and for cloud service consumers to improve the performance of their application systems running in the Cloud.

Understanding the performance difference of a multi-tier Cloud application between different provisioning plans and workloads is difficult to achieve. A typical IaaS provider offers a variety of virtual server instances with different performance capacities and rental rates. Such instances are often marked with a high level description of their hardware/software configuration (e.g. 1 or 2 vC-PUs) which provides insufficient information on the performance of the virtual server instances. Furthermore, as each tier of an application can be independently provisioned with different types and numbers of VMs, the number of possible provisioning plans grows exponentially with each additional tier.

With an average of 80% length reduction, the URL shorteners have become the norm for sharing URLs on Twitter, mainly due to the 140-character limit per message. Unfortunately, spammers have also adopted the URL shorteners to camouflage and improve the user click-through of their spam URLs. In this p

Distributed state monitoring plays a critical role in Cloud datacenter management. One fundamental problem in distributed state monitoring is to minimize the monitoring cost while maximizing the monitoring accuracy at the same time. In this paper, we present Volley, a violation likelihood based approach for efficient distributed state monitoring in Cloud datacenters. Volley achieves both efficiency and accuracy with a flexible monitoring framework which uses dynamic monitoring intervals determined by the likelihood of detecting state violations. Volley consists of three unique techniques. It utilizes efficient node-level adaptation algorithms that minimize monitoring cost with controlled accuracy. Volley also employs a distributed scheme that coordinates the adaptation on multiple monitor nodes of the same task for optimal task- level efficiency. Furthermore, it enables multi-task level cost reduction by exploring state correlation among monitoring tasks. We perform extensive experiments to evaluate Volley with system, network and application monitoring tasks in a virtualized datacenter environment. Our results show that Volley can reduce considerable monitoring cost and still deliver user specified monitoring accuracy under various scenarios.

Spatial alarms are one of the fundamental functionalities for many LBSs. We argue that spatial alarms should be road network aware as mobile objects travel on spatially constrained road networks or walk paths. In this software system demonstration, we will present the first prototype system of ROADALARM - a spatial alarm processing system for moving objects on road networks. The demonstration system of ROAD-ALARM focuses on the three unique features of ROADALARM system design. First, we will show that the road network distance-based spatial alarm is best modeled using road network distance such as segment length-based and travel time-based distance. Thus, a road network spatial alarm is a star-like subgraph centered at the alarm target. Second, we will show the suite of ROADALARM optimization techniques to scale spatial alarm processing by taking into account spatial constraints on road networks and mobility patterns of mobile subscribers. Third, we will show that, by equipping the ROADALARM system with an activity monitoring-based control panel, we are able to enable the system administrator and the end users to visualize road network-based spatial alarms, mobility traces of moving objects and dynamically make selection or customization of the ROADALARM techniques for spatial alarm processing through graphical user interface. We show that the ROADALARM system provides both the general system architecture and the essential building blocks for location-based advertisements and location-based reminders.

We present MobiMix, a road network based mix-zone framework to protect location privacy of mobile users traveling on road networks. An alternative and complementary approach to spatial cloaking based location privacy protection is to break the continuity of location exposure by introducing techniques, such as mix-zones, where no applications can trace user movements. However, existing mixzone proposals fail to provide effective mix-zone construction and placement algorithms that are resilient to timing and transition attacks. In MobiMix, mix-zones are constructed and placed by carefully taking into consideration of multiple factors, such as the geometry of the zones, the statistical behavior of the user population, the spatial constraints on movement patterns of the users, and the temporal and spatial resolution of the location exposure. In this demonstration, we first introduce a visualization of the location privacy risks of mobile users traveling on road networks and show how mixzone based anonymization breaks the continuity of location exposure to protect user location privacy. We demonstrate a suite of road network mix-zone construction and placement methods that provide higher level of resilience to timing and transition attacks on road networks. We show the effectiveness of the MobiMix approach through detailed visualization using traces produced by GTMobiSim on different scales of geographic maps.

Location-sensitive information monitoring services are a centerpiece of the technology for disseminating content-rich information from massive data streams to mobile users. The key challenges for such monitoring services are characterized by the combination of spatial and non-spatial attributes being monitored and the wide spectrum of update rates. A typical example of such services is "alert me when the gas price at a gas station within 5 miles of my current location drops to 4 per gallon". Such a service needs to monitor the gas price changes in conjunction with the highly dynamic nature of location information. Scalability of such location sensitive and content rich information monitoring services in the presence of different update rates and monitoring thresholds poses a big technical challenge. In this paper, we present SLIM, a scalable location sensitive information monitoring service framework with two unique features. First, we make intelligent use of the correlation between spatial and non-spatial attributes involved in the information monitoring service requests to devise a highly scalable distributed spatial trigger evaluation engine. Second, we introduce single and multi-dimensional safe value containment techniques to efficiently perform selective distributed processing of spatial triggers to reduce the amount of unnecessary trigger evaluations. Through extensive experiments, we show that SLIM offers high scalability for location-sensitive, content-rich information monitoring services in terms of the number of information sources being monitored, number of users and monitoring requests.

Service network analysis is an essential aspect of web service discovery, search, mining and recommendation. Many popular web service networks are content-rich in terms of heterogeneous types of entities, attributes and links. A main challenge for ranking services is how to incorporate multiple complex and heterogeneous factors, such as service attributes, relationships between services, relationships between services and service providers or service consumers, into the design of service ranking functions. In this paper, we model services, attributes, and the associated entities, such as providers, consumers, by a heterogeneous service network. We propose a unified neighborhood random walk distance measure, which integrates various types of links and vertex attributes by a local optimal weight assignment. Based on this unified distance measure, a reinforcement algorithm, ServiceRank, is provided to tightly integrate ranking and clustering by mutually and simultaneously enhancing each other such that the performance of both can be improved. An additional clustering matching strategy is proposed to efficiently align clusters from different types of objects. Our extensive evaluation on both synthetic and real service networks demonstrates the effectiveness of ServiceRank in terms of the quality of both clustering and ranking among multiple types of entity, link and attribute similarities in a service network.

We propose a new MapReduce cloud service model, Cura, for data analytics in the cloud. We argue that performing MapReduce analytics in existing cloud service models - either using a generic compute cloud or a dedicated MapReduce cloud - is inadequate and inefficient for production workloads. Existing services require users to select a number of complex cluster and job parameters while simultaneously forcing the cloud provider to use those potentially sub-optimal configurations resulting in poor resource utilization and higher cost. In contrast Cura leverages MapReduce profiling to automatically create the best cluster configuration for the jobs so as to obtain a global resource optimization from the provider perspective. Secondly, to better serve modern MapReduce workloads which constitute a large proportion of interactive real-time jobs, Cura uses a unique instant VM allocation technique that reduces response times by up to 65%. Thirdly, our system introduces deadline-awareness which, by delaying execution of certain jobs, allows the cloud provider to optimize its global resource allocation and reduce costs further. Cura also benefits from a number of additional performance enhancements including cost-aware resource provisioning, VMaware scheduling and online virtual machine reconfiguration. Our experimental results using Facebook-like workload traces show that along with response time improvements, our techniques lead to more than 80% reduction in the compute infrastructure cost of the cloud data center.

Social networks continue to grow in size and the type of information hosted. We witness a growing interest in clustering a social network of people based on both their social relationships and their participations in activity based information networks. In this paper, we present a social influence based clustering framework for analyzing heterogeneous information networks with three unique features. First, we introduce a novel social influence based vertex similarity metric in terms of both self-influence similarity and co-influence similarity. We compute self-influence and co-influence based similarity based on social graph and its associated activity graphs and influence graphs respectively. Second, we compute the combined social influence based similarity between each pair of vertices by unifying the self-similarity and multiple co-influence similarity scores through a weight function with an iterative update method. Third, we design an iterative learning algorithm, SI-Cluster, to dynamically refine the K clusters by continuously quantifying and adjusting the weights on self-influence similarity and on multiple co-influence similarity scores towards the clustering convergence. To make SI-Cluster converge fast, we transformed a sophisticated nonlinear fractional programming problem of multiple weights into a straightforward nonlinear parametric programming problem of single variable. Our experiment results show that SI-Cluster not only achieves a better balance between self-influence and co-influence similarities but also scales extremely well for large graph clustering.

As the size and variety of information networks continue to grow in many scientific and engineering domains, we witness a growing demand for efficient processing of large heterogeneous graphs using a cluster of compute nodes in the Cloud. One open issue is how to effectively partition a large graph to process complex graph operations efficiently. In this paper, we present VB-Partitioner -- a distributed data partitioning model and algorithms for efficient processing of graph operations over large-scale graphs in the Cloud. Our VB-Partitioner has three salient features. First, it introduces vertex blocks (VBs) and extended vertex blocks (EVBs) as the building blocks for semantic partitioning of large graphs. Second, VB-Partitioner utilizes vertex block grouping algorithms to place those vertex blocks that have high correlation in graph structure into the same partition. Third, VB-Partitioner employs a VB-partition guided query partitioning model to speed up the parallel processing of graph pattern queries by reducing the amount of inter-partition query processing. We conduct extensive experiments on several real-world graphs with millions of vertices and billions of edges. Our results show that VB-Partitioner significantly outperforms the popular random block-based data partitioner in terms of query latency and scalability over large-scale graphs.

State monitoring is a fundamental building block for Cloud services. The demand for providing state monitoring as services (MaaS) continues to grow and is evidenced by Cloud-Watch from Amazon EC2, which allows Cloud consumers to pay for monitoring a selection of performance metrics with coarse-grained periodical sampling of runtime states. This dissertation research [1], awarded with the SPEC Distinguished Dissertation Award 2012, is dedicated to innovative research and development of an elastic framework for providing state monitoring as a service (MaaS). We analyze limitations of existing techniques, systematically identify the need and the challenges at different layers of a Cloud monitoring service platform, and develop a suite of distributed monitoring techniques to support flexible monitoring infrastructure, cost-effective state monitoring and monitoring-enhanced Cloud management.

Massive volumes of big RDF data are growing beyond the performance capacity of conventional RDF data management systems operating on a single node. Applications using large RDF data demand efficient data partitioning solutions for supporting RDF data access on a cluster of compute nodes. In this paper we present a novel semantic hash partitioning approach and implement a Semantic HAsh Partitioning-Enabled distributed RDF data management system, called Shape. This paper makes three original contributions. First, the semantic hash partitioning approach we propose extends the simple hash partitioning method through direction-based triple groups and direction-based triple replications. The latter enhances the former by controlled data replication through intelligent utilization of data access locality, such that queries over big RDF graphs can be processed with zero or very small amount of inter-machine communication cost. Second, we generate locality-optimized query execution plans that are more efficient than popular multi-node RDF data management systems by effectively minimizing the inter-machine communication cost for query processing. Third but not the least, we provide a suite of locality-aware optimization techniques to further reduce the partition size and cut down on the inter-machine communication cost during distributed query processing. Experimental results show that our system scales well and can process big RDF datasets more efficiently than existing approaches.

The volume of RDF data continues to grow over the past decade and many known RDF datasets have billions of triples. A grant challenge of managing this huge RDF data is how to access this big RDF data efficiently. A popular approach to addressing the problem is to build a full set of permutations of (S, P, O) indexes. Although this approach has shown to accelerate joins by orders of magnitude, the large space overhead limits the scalability of this approach and makes it heavyweight. In this paper, we present TripleBit, a fast and compact system for storing and accessing RDF data. The design of TripleBit has three salient features. First, the compact design of TripleBit reduces both the size of stored RDF data and the size of its indexes. Second, TripleBit introduces two auxiliary index structures, ID-Chunk bit matrix and ID-Predicate bit matrix, to minimize the cost of index selection during query evaluation. Third, its query processor dynamically generates an optimal execution ordering for join queries, leading to fast query execution and effective reduction on the size of intermediate results. Our experiments show that TripleBit outperforms RDF-3X, MonetDB, BitMat on LUBM, UniProt and BTC 2012 benchmark queries and it offers orders of mangnitude performance improvement for some complex join queries.

This paper introduces the concept of monitoring-as-a-service (MaaS), its main components, and a suite of key functional requirements of MaaS in cloud. We argue that MaaS should support not only the conventional state monitoring capabilities, such as instantaneous violation detection, periodical state monitoring, and single tenant monitoring, but also performance-enhanced functionalities that can optimize on monitoring cost, scalability, and the effectiveness of monitoring service consolidation and isolation. In this paper, we present three enhanced MaaS capabilities and show that window-based state monitoring is not only more resilient to noises and outliers, but also saves considerable communication cost. Similarly, violation-likelihood-based state monitoring can dynamically adjust monitoring intensity based on the likelihood of detecting important events, leading to significant gain in monitoring service consolidation. Finally, multitenancy support in state monitoring allows multiple cloud users to enjoy MaaS with improved performance and efficiency at more affordable cost. We perform extensive experiments in an emulated cloud environment with real-world system and network traces. The experimental results suggest that our MaaS framework achieves significant lower monitoring cost, higher scalability, and better multitenancy performance.

With ubiquitous wireless connectivity and technological advances in mobile devices, we witness the growing demands and increasing market shares of mobile intelligent systems and technologies for real-time decision making and location-based knowledge discovery. Spatial alarms are considered as one of the fundamental capabilities for intelligent mobile location-based systems. Like time-based alarms that remind us the arrival of a future time point, spatial alarms remind us the arrival of a future spatial point. Existing approaches for scaling spatial alarm processing are focused on computing Alarm-Free Regions (A fr ) and Alarm-Free Period (A fp ) such that mobile objects traveling within an A fr can safely hibernate the alarm evaluation process for the computed A fp , to save battery power, until approaching the nearest alarm of interest. A key technical challenge in scaling spatial alarm processing is to efficiently compute A fr and A fp such that mobile objects traveling within an A fr can safely hibernate the alarm evaluation process during the computed A fp , while maintaining high accuracy. In this article we argue that on-demand computation of A fr is expensive and may not scale well for dense populations of mobile objects. Instead, we propose to maintain an index for both spatial alarms and empty regions (A fr ) such that for a given mobile user's location, we can find relevant spatial alarms and whether it is in an alarm-free region more efficiently. We also show that conventional spatial indexing methods, such as R-tree family, k -d tree, Quadtree, and Grid, are by design not well suited to index empty regions. We present Mondrian Tree – a region partitioning tree for indexing both spatial alarms and alarm-free regions. We first introduce the Mondrian Tree indexing algorithms, including index construction, search, and maintenance. Then we describe a suite of Mondrian Tree optimizations to further enhance the performance of spatial alarm processing. Our experimental evaluation shows that the Mondrian Tree index, as an intelligent technology for mobile systems, outperforms traditional index methods, such as R-tree, Quadtree, and k -d tree, for spatial alarm processing.

T HE fi eld of services computing has evolved into a well-recognized discipline in computing, bridging information technology services and conventional business services.Services computing covers the science and technology of packaging and delivering computing functionality, spanning hardware, software, and applications layers as a suite of self-describing and autonomous functional components and tools that are interoperable through a variety of service delivery models.An important goal of services computing is to enable computing technology and information technology to be delivered and managed as services to advance and empower business services with ultimate elasticity and effi cacy.We are entering a new era of computing, with services computing as a cornerstone, interweaved by an exciting collection of technological revolutions from multimedia computing, pervasive computing, and social computing to cloud computing and big data computing.I am honored to have been elected as the new Editor-in-Chief (EIC) of the IEEE Transactions on Services Computing (TSC) at this exciting time and for the opportunity to serve and shepherd the community's premier publication venue in its next milestone.Since its inception in 2008, under the leadership of founding EIC Dr. Liang-Jie Zhang, TSC has grown to be a fl agship publication for research results in the core technology of services computing, including serviceoriented architecture, web services, mashups, cloud computing, software as a service, business consulting methodology and utilities, and business process modeling, transformation, and integration.I am also thankful to Dr. Liang-Jie Zhang and the steering committee of the IEEE Services Congress (co-located multiconferences including ICWS, SCC, and IEEE Cloud), which have nurtured and fueled TSC with cutting-edge research contributions in Services Computing from a diverse set of active researchers as well as practitioners worldwide.Moving forward, I would like to continue exploring opportunities of making TSC an exciting and defi nitive forum for attracting and publishing impactful research contributions that are transformative.For example, as information digitization becoming ubiquitous, the relevance of academic journals is being challenged by highly selective conferences as alternative venues for high quality research.It is worth examining the opportunities for and challenges to making TSC serve not only as an archival journal that will provide a historical perspective of research development but also as a forum for disseminating timely and exciting ongoing research that can stimulate innovation.To achieve such a promise for TSC, several directions are worth pursuing, including seeking ways to reach the broadest possible readership over time, broadening the scope of TSC from core technology and concepts in services computing.TSC will cover all areas of services computing, especially emerging technologies, such as data storage technology, green computing or energyeffi cient computing technology, and data analytics as a service to name a few.Furthermore, we will explore ways to shorten both the submission-review-decision cycle and the acceptance-to-publication production cycle.I would like to conclude by emphasizing the importance of all the contributors to TSC: the editorial board, the reviewers, the authors, the IEEE Computer Society staff, and the production crews.Their active contributions and dedication are critical for TSC to continue its success.Together with my predecessor, Dr. Liang-Jie Zhang,

The Editor-in-Chief introduces 13 new members of the Editorial Board. A brief professional biography is presented for each. They are: Rong N. Chang, Elena Ferrari, Joao Eduardo Ferreira, Bugra Gedik, Andrzej Goscinski, Keqing He, James Joshi, Wang-Chien Lee, Aameek Singh, Munindar Singh, Hans Weigand, I-Ling Yen, and Yanyong Zhang.

Server consolidation and application consolidation through virtualization are key performance optimizations in cloud-based service delivery industry. In this paper, we argue that it is important for both cloud consumers and cloud providers to understand the various factors that may have significant impact on the performance of applications running in a virtualized cloud. This paper presents an extensive performance study of network I/O workloads in a virtualized cloud environment. We first show that current implementation of virtual machine monitor (VMM) does not provide sufficient performance isolation to guarantee the effectiveness of resource sharing across multiple virtual machine instances (VMs) running on a single physical host machine, especially when applications running on neighboring VMs are competing for computing and communication resources. Then we study a set of representative workloads in cloud-based data centers, which compete for either CPU or network I/O resources, and present the detailed analysis on different factors that can impact the throughput performance and resource sharing effectiveness. For example, we analyze the cost and the benefit of running idle VM instances on a physical host where some applications are hosted concurrently. We also present an in-depth discussion on the performance impact of colocating applications that compete for either CPU or network I/O resources. Finally, we analyze the impact of different CPU resource scheduling strategies and different workload rates on the performance of applications running on different VMs hosted by the same physical machine.

As a promising standard of television, Internet protocol television (IPTV) gains increasing popularity with its on-demand services. However, supporting scalable on-demand IPTV services remains to be an important challenge. Existing IPTV architecture dedicates a regional station to serve subscribers in the respective region regardless of temporal and spatial dynamics in service demand. This may cause significant imbalance of resource utilization and service provisioning delay among stations, especially with increasing subscribers and service catalog. In this paper, we propose CiTV, a collaborative Internet TV service platform that achieves efficient and scalable IPTV service delivery through interstation collaboration. First, CiTV employs a novel request dispatching algorithm that enables stations to autonomously dispatch requests among themselves and achieves low per-view cost and load variance at the same time. Second, CiTV also leverages a distributed content placement algorithm to globally optimize content utilization at multistation level, which is important for deployment scalability. Last but not the least, CiTV also ensures service delivery quality and efficiency through fine-grained collaboration scope control. We perform both theoretical and empirical analysis of CiTV. Our experiment results suggest that CiTV significantly improves the scalability of on-demand IPTV services for the existing IPTV architecture.

User-perceived performance continues to be the most important QoS indicator in cloud-based data centers today. Effective allocation of virtual machines (VMs) to handle both CPU intensive and I/O intensive workloads is a crucial performance management capability in virtualized clouds. Although a fair amount of researches have dedicated to measuring and scheduling jobs among VMs, there still lacks of in-depth understanding of performance factors that impact the efficiency and effectiveness of resource multiplexing and scheduling among VMs. In this paper, we present the experimental research on performance interference in parallel processing of CPU-intensive and network-intensive workloads on Xen virtual machine monitor (VMM). Based on our study, we conclude with five key findings which are critical for effective performance management and tuning in virtualized clouds. First, colocating network-intensive workloads in isolated VMs incurs high overheads of switches and events in Dom0 and VMM. Second, colocating CPU-intensive workloads in isolated VMs incurs high CPU contention due to fast I/O processing in I/O channel. Third, running CPU-intensive and network-intensive workloads in conjunction incurs the least resource contention, delivering higher aggregate performance. Fourth, performance of network-intensive workload is insensitive to CPU assignment among VMs, whereas adaptive CPU assignment among VMs is critical to CPU-intensive workload. The more CPUs pinned on Dom0 the worse performance is achieved by CPU-intensive workload. Last, due to fast I/O processing in I/O channel, limitation on grant table is a potential bottleneck in Xen. We argue that identifying the factors that impact the total demand of exchanged memory pages is important to the in-depth understanding of interference costs in Dom0 and VMM.

State monitoring is widely used for detecting critical events and abnormalities of distributed systems. As the scale of such systems grows and the degree of workload consolidation increases in Cloud data centers, node failures and performance interferences, especially transient ones, become the norm rather than the exception. Hence, distributed state monitoring tasks are often exposed to impaired communication caused by such dynamics on different nodes. Unfortunately, existing distributed state monitoring approaches are often designed under the assumption of always-online distributed monitoring nodes and reliable inter-node communication. As a result, these approaches often produce misleading results which in turn introduce various problems to Cloud users who rely on state monitoring results to perform automatic management tasks such as auto-scaling. This paper introduces a new state monitoring approach that tackles this challenge by exposing and handling communication dynamics such as message delay and loss in Cloud monitoring environments. Our approach delivers two distinct features. First, it quantitatively estimates the accuracy of monitoring results to capture uncertainties introduced by messaging dynamics. This feature helps users to distinguish trustworthy monitoring results from ones heavily deviated from the truth, yet significantly improves monitoring utility compared with simple techniques that invalidate all monitoring results generated with the presence of messaging dynamics. Second, our approach also adapts to non-transient messaging issues by reconfiguring distributed monitoring algorithms to minimize monitoring errors. Our experimental results show that, even under severe message loss and delay, our approach consistently improves monitoring accuracy, and when applied to Cloud application auto-scaling, outperforms existing state monitoring techniques in terms of the ability to correctly trigger dynamic provisioning.

This paper argues that trust and reputation models should take into account not only direct experiences (local trust)and experiences from the circle of ”friends”, but also be attack resilient by design in the presence of dishonest feedbacks and sparse network connectivity. We first revisit EigenTrus

Network I/O workloads are dominating in most of the Cloud datacenters today. One way to improve inter-VM communication efficiency is to support co-resident VM communication using a faster communication protocol than the traditional TCP/IP commonly used regardless whether VMs are located on the same

Group communication is essential for multi-user applications. However, due to unpredictable node departures and non-deterministic network partitions, providing reliable and scalable group communication services is challenging when the applications are utilized by the users with heterogeneous capacities on a large scale. To address this challenge, we propose a novel replication scheme to achieve high reliability and low-cost scalability in group communication with following three features. First, it introduces a new concept of replication based on topological similarity, which empowers each node with an ability of measuring similarity between the nodes in topology. By eliminating the topological similarity between the replicas, it intelligently mitigates service interruptions caused by node failures and network partitions. Second, instead of specifying the number of replicas, it provides a technique for nodes to dynamically adapt the replication placement schemes by exploiting functionality importance of the nodes in the group- communication session. It eliminates the bottleneck problem and improves the network resource utilization. Third, the scheme is self-converging and it can stabilize within a few adaptations even facing a high churn rate. Extensive simulations show that it yields significant improvements in reduction of replication overhead and service interruption when comparing to existing approaches.

Mining trajectory data has been gaining significant interest in recent years. However, existing approaches to trajectory clustering are mainly based on density and Euclidean distance measures. We argue that when the utility of spatial clustering of mobile object trajectories is targeted at road network aware location based applications, density and Euclidean distance are no longer the effective measures. This is because traffic flows in a road network and the flow-based density characterization become important factors for finding interesting trajectory clusters of mobile objects travelling in road networks. In this paper, we propose NEAT-a road network aware approach for fast and effective clustering of spatial trajectories of mobile objects travelling in road networks. Our method takes into account the physical constraints of the road network, the network proximity and the traffic flows among consecutive road segments to organize trajectories into spatial clusters. The clusters discovered by NEAT are groups of sub-trajectories which describe both dense and highly continuous traffic flows of mobile objects. We perform extensive experiments with mobility traces generated using different scales of real road network maps. Our experimental results demonstrate that the NEAT approach is highly accurate and runs orders of magnitude faster than existing density-based trajectory clustering approaches.

Spatial alarm services are essential components of many location-based applications. One of the key technical challenges for supporting spatial alarms as a service is performance and scalability. This paper shows that the Euclidean distance-based spatial alarm processing techniques are inadequate for mobile users traveling on road networks due to the high overhead in terms of server load for alarm checks and the high energy consumption in terms of client wakeups. We design and develop RoadAlarm, a road network aware spatial alarm processing service, with three unique features. First, we introduce the concept of road network-based spatial alarms using road network distance measures and a set of metrics specialized for spatial alarm processing. Second, we develop the basic model for spatial alarm processing by exploiting two types of filters: subscription filter and Euclidean lower bound filter. Third and but not the least, we develop a suite of optimization techniques to further reduce the frequency of wakeups at mobile clients and the number of alarm checks at the alarm processing server, while ensuring high accuracy of spatial alarm processing. Our experimental results show that RoadAlarm outperforms existing Euclidean space-based approaches with high success rate (accuracy) and significantly increased hibernation time.

Mix-zones are recognized as an alternative and complementary approach to spatial cloaking based approach to location privacy protection. Mix-zones break the continuity of location exposure by ensuring that users' movements cannot be traced while they reside in a mix-zone. In this paper we provide an overview of various known attacks that make mix-zones on road networks vulnerable and illustrate a set of counter measures to make road network mix-zones attack resilient. Concretely, we categorize the vulnerabilities of road network mix-zones into two classes: one due to the road network characteristics and user mobility, and the other due to the temporal, spatial and semantic correlations of location queries. For instance, the timing information of users' entry and exit into a mix-zone provides information to launch a timing attack. The non-uniformity in the transitions taken at the road intersection may lead to transition attack. An example query correlation attack is the basic continual query (CQ) attacks, which attempt to break the anonymity of road network aware mix-zones by performing query correlation based inference. The CQ-timing attacks carry out inference attacks based on both query correlation and timing correlation, and the CQ-transition attacks execute inference attacks based on both query correlation and transition correlation. We study the factors that impact on the effectiveness of each of these attacks and evaluate the efficiency of the counter measures, such as non-rectangle mix-zones and delay tolerant mix-zones, through extensive experiments on traces produced by GTMobiSim at different scales of geographic maps.

The immensity and variety of personal information (e.g., profile, photo, and microblog) on social sites require access control policies tailored to individuals' privacy needs. Today such policies are still mainly specified manually by ordinary users, which is usually coarse-grained, tedious, and error-prone. This paper presents the design, implementation, and evaluation of an automated access control policy specification tool, XACCESS, that helps non-expert users effectively specify who should have access to which part of their data. A series of key features distinguish XACCESS from prior work: 1) it adopts a role-based access control model (instead of the conventional rule-based paradigm) to capture the implicit privacy/interest preference of social site users; 2) it employs a novel hybrid mining method to extract a set of semantically interpretable, functional "social roles", from both static network structures and dynamic historical activities; 3) based on the identified social roles, confidentiality setting of personal data, and (optional and possibly inconsistent) predefined user-permission assignments, it recommends a set of high-quality privacy settings; 4) it allows user feedback in every phase of the process to further improve the quality of the suggested privacy policies. A comprehensive experimental evaluation is conducted over real social network and user study data to validate the efficacy of XACCESS.

Social influences, the phenomena that one individual's actions can induce similar behaviors among his/her friends via their social ties, have been observed prevailingly in socially networked systems. While most existing work focuses on studying general, macro-level influence (e.g., diffusion); equally important is to understand social influence at microscopic scales (i.e., at the granularity of single individuals, actions, and time-stamps), which may benefit a range of applications. We propose μSI, a microscopic social-influence model wherein: individuals' actions are modeled as temporary interactions between social network (formed by individuals) and object network (formed by targets of actions); one individual's actions influence his/her friends in a dynamic, network-wise manner (i.e., dependent on both social and object networks). We develop for μSI a suite of novel inference tools that enable to answer questions of the form: How may an occurred interaction trigger another? More importantly, when and where may a new interaction be observed? We carefully address the computational challenges for inferencing over such semantically rich models by dynamically identifying sub-domains of interest and varying the precision of solutions over different sub-domains. We demonstrate the breadth and generality of μSI using two seemingly disparate applications. In the context of social tagging service, we show how it can help improve the accuracy and freshness of resource recommendation; in the context of mobile phone call service, we show how it can help improve the efficiency of paging operation.

Most proposed DHTs engage certain topology maintenance mechanisms specific to the static graphs on which they are based. The designs of these mechanisms are complicated and repeated with graph-relevant concerns. In this paper, we propose the “distributed line graphs” (DLG), a universal technique for designing DHTs based on arbitrary regular graphs. Using DLG, the main features of the initial graphs are preserved, and thus people can design a new DHT by simply choosing the graph with desirable features and applying DLG to it. We demonstrate the power of DLG by illustrating four DLG-enabled DHTs based on different graphs, namely, Kautz, de Bruijn, butterfly, and hypertree graphs. The effectiveness of our proposals is demonstrated through analysis, simulation, and implementation.

Social trust and recommendation services are the most popular social rating systems today for service providers to learn about the social opinion or popularity of a product, item, or service, such as a book on Amazon, a seller on eBay, a story on Digg or a movie on Netflix. Such social rating systems are very convenient and offer alternative learning environments for decision makers, but they open the door for attackers to manipulate the social rating systems by selfishly promoting or maliciously demoting certain items. Although a fair amount of effort has been made to understand various risks and possible defense mechanisms to counter such attacks, most of the existing work to date has been devoted to studying specific types of attacks and their countermeasures. In this article, we argue that vulnerabilities in social rating systems and their countermeasures should be examined and analyzed in a systematic manner. We first give an overview of the common vulnerabilities and attacks observed in some popular social rating services. Next, we describe three types of attack strategies in two types of social rating systems, including a comprehensive theoretical analysis of their attack effectiveness and attack costs. Three context-aware countermeasures are then presented: (i) hiding user-item relationships, (ii) using confidence weight to distinguish popular and unpopular items, and (iii) incorporating time windows in trust establishment. We also provide an in-depth discussion on how these countermeasures can be used effectively to improve the robustness and trustworthiness of the social rating services.

The increasing popularity of large-scale distributed applications in datacenters has led to the growing demand of distributed application state monitoring. These application state monitoring tasks often involve collecting values of various status attributes from a large number of nodes. One challenge in such large-scale application state monitoring is to organize nodes into a monitoring overlay that achieves monitoring scalability and cost effectiveness at the same time. In this paper, we present REMO, a REsource-aware application state MOnitoring system, to address the challenge of monitoring overlay construction. REMO distinguishes itself from existing works in several key aspects. First, it jointly considers intertask cost-sharing opportunities and node-level resource constraints. Furthermore, it explicitly models the per-message processing overhead which can be substantial but is often ignored by previous works. Second, REMO produces a forest of optimized monitoring trees through iterations of two phases. One phase explores cost-sharing opportunities between tasks, and the other refines the tree with resource-sensitive construction schemes. Finally, REMO also employs an adaptive algorithm that balances the benefits and costs of overlay adaptation. This is particularly useful for large systems with constantly changing monitoring tasks. Moreover, we enhance REMO in terms of both performance and applicability with a series of optimization and extension techniques. We perform extensive experiments including deploying REMO on a BlueGene/P rack running IBM's large-scale distributed streaming system - System S. Using REMO in the context of collecting over 200 monitoring tasks for an application deployed across 200 nodes results in a 35-45 percent decrease in the percentage error of collected attributes compared to existing schemes.

Reputation-based trust systems provide important capability in open and service-oriented computing environments. Most existing trust models fail to assess the variance of a reputation prediction. Moreover, the summation method, widely used for reputation feedback aggregation, is vulnerable to malicious feedbacks. This paper presents a general trust model, called RLM, for a more comprehensive and robust reputation evaluation. Concretely, we define a comprehensive reputation evaluation method based on two attributes: reputation value and reputation prediction variance. The reputation predication variance serves as a quality measure of the reputation value computed based on aggregation of feedbacks. For feedback aggregation, we propose the novel Kalman aggregation method, which can inherently support robust trust evaluation. To defend against malicious and coordinated feedbacks, we design the Expectation Maximization algorithm to autonomously mitigate the influence of a malicious feedback, and further apply the hypothesis test method to resist malicious feedbacks precisely. Through theoretical analysis, we demonstrate the robustness of the RLM design against adulating and defaming attacks, two popular types of feedback attacks. Our experiments show that the RLM model can effectively capture the reputation's evolution and outperform the popular summation-based trust models in terms of both accuracy and attack resilience. Concretely, under the attack of collusive malicious feedbacks, RLM offers higher robustness for the reputation prediction and a lower false positive rate for the malicious feedback detection.

The past few years have witnessed an increasing demand for the next generation health information networks (e.g., NHIN[1]), which hold the promise of supporting large-scale information sharing across a network formed by autonomous healthcare providers. One fundamental capability of such information network is to support efficient, privacy-preserving (for both users and providers) search over the distributed, access controlled healthcare documents. In this paper we focus on addressing the privacy concerns of content providers; that is, the search should not reveal the specific association between contents and providers (a.k.a. content privacy). We propose SS-PPI, a novel privacy-preserving index abstraction, which, in conjunction of distributed access control-enforced search protocols, provides theoretically guaranteed protection of content privacy. Compared with existing proposals (e.g., flipping privacy-preserving index[2]), our solution highlights with a series of distinct features: (a) it incorporates access control policies in the privacy-preserving index, which improves both search efficiency and attack resilience; (b) it employs a fast index construction protocol via a novel use of the secrete-sharing scheme in a fully distributed manner (without trusted third party), requiring only constant (typically two) round of communication; (c) it provides information-theoretic security against colluding adversaries during index construction as well as query answering. We conduct both formal analysis and experimental evaluation of SS-PPI and show that it outperforms the state-of-the-art solutions in terms of both privacy protection and execution efficiency.

Parallel processing is the key to speedup performance and to achieve high throughput in processing large scale data analytical workloads. However, failures of nodes involved in the analytical query can interrupt the whole process, resulting in the complete restart of the query if the system does not

As interests in sharing and mining social network data continue to grow, we see a growing demand for privacy preserving social network data publishing. In this paper, we discuss privacy risks in publishing social network data and the design principles for developing countermeasures. The main co

This paper presents MobiMix, a road network based mix-zone framework to protect location privacy of mobile users traveling on road networks. In contrast to spatial cloaking based location privacy protection, the approach in MobiMix is to break the continuity of location exposure by using mix-zones, where no applications can trace user movement. This paper makes two original contributions. First, we provide the formal analysis on the vulnerabilities of directly applying theoretical rectangle mix-zones to road networks in terms of anonymization effectiveness and attack resilience. We argue that effective mix-zones should be constructed and placed by carefully taking into consideration of multiple factors, such as the geometry of the zones, the statistical behavior of the user population, the spatial constraints on movement patterns of the users, and the temporal and spatial resolution of the location exposure. Second, we develop a suite of road network mix-zone construction methods that provide higher level of attack resilience and yield a specified lower-bound on the level of anonymity. We evaluate the MobiMix approach through extensive experiments conducted on traces produced by GTMobiSim on different scales of geographic maps. Our experiments show that MobiMix offers high level of anonymity and high level of resilience to attacks compared to existing mix-zone approaches.

Information leakage via the networks formed by subjects (e.g., Facebook, Twitter) and objects (e.g., blogosphere) - some of whom may be controlled by malicious insiders - often leads to unpredicted access control risks. While it may be impossible to precisely quantify information flows between two entities (e.g., two friends in a social network), this paper presents a first attempt towards leveraging recent advances in modeling socio-information networks to develop a statistical risk estimation paradigm for quantifying such insider threats. In the context of socio-information networks, our models estimate the following likelihoods: prior flow - has a subject $s$ acquired covert access to object o via the networks? posterior flow - if s is granted access to o, what is its impact on information flows between subject s' and object o'? network evolution - how will a newly created social relationship between s and s' influence current risk estimates? Our goal is not to prescribe a one-size-fits-all solution; instead we develop a set of composable network-centric risk estimation operators, with implementations configurable to concrete socio-information networks. The efficacy of our solutions is empirically evaluated using real-life datasets collected from the IBM SmallBlue project and Twitter.

We present Purlieus, a MapReduce resource allocation system aimed at enhancing the performance of MapReduce jobs in the cloud. Purlieus provisions virtual MapReduce clusters in a locality-aware manner enabling MapReduce virtual machines (VMs) access to input data and importantly, intermediate data from local or close-by physical machines. We demonstrate how this locality-awareness during both map and reduce phases of the job not only improves runtime performance of individual jobs but also has an additional advantage of reducing network traffic generated in the cloud data center. This is accomplished using a novel coupling of, otherwise independent, data and VM placement steps. We conduct a detailed evaluation of Purlieus and demonstrate significant savings in network traffic and almost 50% reduction in job execution times for a variety of workloads.

Abstract Distributed Hash Tables (DHTs) are scalable, self‐organizing, and adaptive to underlying topology changes, thus being a promising infrastructure for hosting large‐scale distributed applications. The ever‐wider use of DHT infrastructures has found more and more applications that require support for range queries. Recently, a number of DHT‐based range query schemes have been proposed. However, most of them suffer from high query delay or imbalanced load distribution. To address these problems, in this paper we first present an efficient indexing structure called Balanced Kautz (BK) tree that uniformly maps them‐dimensional data space onto DHT nodes, and then propose a BK tree‐based range query scheme called ERQ that processes range queries in a parallel fashion and guarantees to return the results in a bounded delay. In a DHT withNnodes, ERQ can answer any range of query in less thanrmlog N(2loglog N+ 1) hops in a load‐balanced manner, irrespective of the queried range, the whole space size, or the number of queried attributes. The effectiveness of our proposals is demonstrated through experiments. Copyright © 2010 John Wiley & Sons, Ltd.

In location-based, mobile continual query (CQ) systems, two key measures of quality-of-service (QoS) are: freshness and accuracy. To achieve freshness, the CQ server must perform frequent query reevaluations. To attain accuracy, the CQ server must receive and process frequent position updates from the mobile nodes. However, it is often difficult to obtain fresh and accurate CQ results simultaneously, due to 1) limited resources in computing and communication and 2) fast-changing load conditions caused by continuous mobile node movement. Hence, a key challenge for a mobile CQ system is: How do we achieve the highest possible quality of the CQ results, in both freshness and accuracy, with currently available resources? In this paper, we formulate this problem as a load shedding one, and develop MobiQual—a QoS-aware approach to performing both update load shedding and query load shedding. The design of MobiQual highlights three important features. 1) Differentiated load shedding: We apply different amounts of query load shedding and update load shedding to different groups of queries and mobile nodes, respectively. 2) Per-query QoS specification: Individualized QoS specifications are used to maximize the overall freshness and accuracy of the query results. 3) Low-cost adaptation: MobiQual dynamically adapts, with a minimal overhead, to changing load conditions and available resources. We conduct a set of comprehensive experiments to evaluate the effectiveness of MobiQual. The results show that, through a careful combination of update and query load shedding, the MobiQual approach leads to much higher freshness and accuracy in the query results in all cases, compared to existing approaches that lack the QoS-awareness properties of MobiQual, as well as the solutions that perform query-only or update-only load shedding.

Monitoring global states of a distributed cloud application is a critical functionality for cloud datacenter management. State monitoring requires meeting two demanding objectives: high level of correctness, which ensures zero or low error rate, and high communication efficiency, which demands minimal communication cost in detecting state updates. Most existing work follows an instantaneous model which triggers state alerts whenever a constraint is violated. This model may cause frequent and unnecessary alerts due to momentary value bursts and outliers. Countermeasures of such alerts may further cause problematic operations. In this paper, we present a WIndow-based StatE monitoring (WISE) framework for efficiently managing cloud applications. Window-based state monitoring reports alerts only when state violation is continuous within a time window. We show that it is not only more resilient to value bursts and outliers, but also able to save considerable communication when implemented in a distributed manner based on four technical contributions. First, we present the architectural design and deployment options for window-based state monitoring with centralized parameter tuning. Second, we develop a new distributed parameter tuning scheme enabling WISE to scale to much more monitoring nodes as each node tunes its monitoring parameters reactively without global information. Third, we introduce two optimization techniques, including their design rationale, correctness and usage model, to further reduce the communication cost. Finally, we provide an in-depth empirical study of the scalability of WISE, and evaluate the improvement brought by the distributed tuning scheme and the two performance optimizations. Our results show that WISE reduces communication by 50-90 percent compared with instantaneous monitoring approaches, and the improved WISE gains a clear scalability advantage over its centralized version.

Publish-subscribe (pub-sub) is an emerging paradigm for building a large number of distributed systems. A wide area pub-sub system is usually implemented on an overlay network infrastructure to enable information dissemination from publishers to subscribers. Using an open overlay network raises several security concerns such as: confidentiality and integrity, authentication, authorization and Denial-of-Service (DoS) attacks. In this article we present EventGuard, a framework for building secure wide-area pub-sub systems. The EventGuard architecture is comprised of three key components: (1) a suite of security guards that can be seamlessly plugged-into a content-based pub-sub system, (2) a scalable key management algorithm to enforce access control on subscribers, and (3) a resilient pub-sub network design that is capable of scalable routing, handling message dropping-based DoS attacks, and node failures. The design of EventGuard mechanisms aims at providing security guarantees while maintaining the system’s overall simplicity, scalability, and performance metrics. We describe an implementation of the EventGuard pub-sub system to show that EventGuard is easily stackable on any content-based pub-sub core. We present detailed experimental results that quantify the overhead of the EventGuard pub-sub system and demonstrate its resilience against various attacks.

Privacy has been identified as a vital requirement in designing and implementing data mining systems. In general, privacy preservation demands protecting both input and output privacy: the former refers to sanitizing the raw data itself before performing mining; while the latter refers to preventing the mining output (models or patterns) from malicious inference attacks. This article presents a systematic study on the problem of protecting output privacy in data mining, and particularly, stream mining: (i) we highlight the importance of this problem by showing that even sufficient protection of input privacy does not guarantee that of output privacy; (ii) we present a general inferencing and disclosure model that exploits the intrawindow and interwindow privacy breaches in stream mining output; (iii) we propose a light-weighted countermeasure that effectively eliminates these breaches without explicitly detecting them, while minimizing the loss of output accuracy; (iv) we further optimize the basic scheme by taking account of two types of semantic constraints, aiming at maximally preserving utility-related semantics while maintaining hard privacy guarantee; (v) finally, we conduct extensive experimental evaluation over both synthetic and real data to validate the efficacy of our approach.

(A short version of this paper appears in IEEE INFOCOM 2009: http://www.research.ibm.com/people/i/iyengar/INFOCOM2009-kanon.pdf.) Peer-to-peer VoIP (voice over IP) networks, exemplified by Skype, are becoming increasingly popular due to their significant cost advantage and richer call forwarding features than traditional public switched telephone networks. One of the most important features of a VoIP network is privacy (for VoIP clients). Unfortunately, most peer-to-peer VoIP networks neither provide personalization nor guarantee a quantifiable privacy level. In this paper, we propose novel flow analysis attacks that demonstrate the vulnerabilities of peer-to-peer VoIP networks to privacy attacks. We then address two important challenges in designing privacy-aware VoIP networks: Can we provide personalized privacy guarantees for VoIP clients that allow them to select privacy requirements on a per-call basis? How to design VoIP protocols to support customizable privacy guarantee? This paper proposes practical solutions to address these challenges using a quantifiable k-anonymity metric and a privacy-aware VoIP route setup and route maintenance protocols. We present detailed experimental evaluation that demonstrates the performance and scalability of our protocol, while meeting customizable privacy guarantees.

Virtualization is a key technology for cloud based data centers to implement the vision of infrastructure as a service (IaaS) and to promote effective server consolidation and application consolidation. However, current implementation of virtual machine monitor does not provide sufficient performance isolation to guarantee the effectiveness of resource sharing, especially when the applications running on multiple virtual machines of the same physical machine are competing for computing and communication sources. In this paper, we present our performance measurement study of network I/O applications in virtualized cloud. We focus our measurement based analysis on performance impact of co-locating applications in a virtualized cloud in terms of throughput and resource sharing effectiveness, including the impact of idle instances on applications that are running concurrently on the same physical host. Our results show that by strategically co-locating network I/O applications, performance improvement for cloud consumers can be as high as 34%, and the cloud providers can achieve over 40% performance gain.

Server virtualization offers the ability to slice large, underutilized physical servers into smaller, parallel virtual machines (VMs), enabling diverse applications to run in isolated environments on a shared hardware platform. Effective management of virtualized cloud environments introduces new and unique challenges, such as efficient CPU scheduling for virtual machines, effective allocation of virtual machines to handle both CPU intensive and I/O intensive workloads. Although a fair number of research projects have dedicated to measuring, scheduling, and resource management of virtual machines, there still lacks of in-depth understanding of the performance factors that can impact the efficiency and effectiveness of resource multiplexing and resource scheduling among virtual machines. In this paper, we present our experimental study on the performance interference in parallel processing of CPU and network intensive workloads in the Xen Virtual Machine Monitors (VMMs). We conduct extensive experiments to measure the performance interference among VMs running network I/O workloads that are either CPU bound or network bound. Based on our experiments and observations, we conclude with four key findings that are critical to effective management of virtualized cloud environments for both cloud service providers and cloud consumers. First, running network-intensive workloads in isolated environments on a shared hardware platform can lead to high overheads due to extensive context switches and events in driver domain and VMM. Second, co-locating CPU-intensive workloads in isolated environments on a shared hardware platform can incur high CPU contention due to the demand for fast memory pages exchanges in I/O channel. Third, running CPU-intensive workloads and network-intensive workloads in conjunction incurs the least resource contention, delivering higher aggregate performance. Last but not the least, identifying factors that impact the total demand of the exchanged memory pages is critical to the in-depth understanding of the interference overheads in I/O channel in the driver domain and VMM.

With Cloud Computing gaining tremendous importance in the recent past, understanding low-level implications of the cloud infrastructure becomes necessary. One of the key technologies deployed in large Cloud infrastructures namely the Amazon EC2 for providing isolation and separate protection domains for multiple clients is virtualization. Therefore, identifying the performance bottlenecks in a virtualized setup and understanding the implications of workload combinations and resource configurations on the overall I/O performance helps both the cloud providers in managing their infrastructure efficiently and also their customers by means of better performance. In this paper we present the measurement results of detailed experiments conducted on a virtualized setup focusing on the storage I/O performance. We categorize our experimental evaluation into four components, each of which presenting some significant factors that affect storage I/O performance. Our experimental results can be useful for cloud application developers to tune their applications for better I/O performance and for the cloud service providers to make more effective decisions on resource provisioning and workload scheduling.

Multi-tiered storage systems today are integrating Solid State Disks (SSD) on top of traditional rotational hard disks for performance enhancement due to the significant IO improvements in SSD technology. It is widely recognized that automated data migration between SSD and HDD plays a critical role in effective integration of SSD into multi-tiered storage systems. Furthermore, effective data migration has to take into account of application specific workload characteristics, deadlines, and IO profiles. An important and interesting challenge for automated data migration in multi-tiered storage systems is how to fully release the power of data migration while guaranteeing the migration deadline is critical to maximizing the performance of SSD-enabled multi-tiered storage system. In this paper, we present an adaptive look ahead data migration model that can incorporate application specific characteristics and I/O profiles as well as workload deadlines. Our adaptive data migration model has three unique features. First, it incorporates a set of key factors that may impact on the performance of look ahead migration efficiency in our formal model develop. Second, our data migration model can adaptively determine the optimal look ahead window size, based on several parameters, to optimize the effectiveness of look ahead migration. Third, we formally and experimentally show that the adaptive data migration model can improve overall system performance and resource utilization while meeting workload deadlines. Through our trace driven experimental study, we compare the adaptive look ahead migration approach with the basic migration model and show that the adaptive migration model is effective and efficient for continuously improving and tuning of the performance and scalability of multi-tier storage systems.

With the widespread use of electronic health record (EHR), building a secure EHR sharing environment has attracted a lot of attention in both healthcare industry and academic community. Cloud computing paradigm is one of the popular healthIT infrastructure for facilitating EHR sharing and EHR integration. In this paper we discuss important concepts related to EHR sharing and integration in healthcare clouds and analyze the arising security and privacy issues in access and management of EHRs. We describe an EHR security reference model for managing security issues in healthcare clouds, which highlights three important core components in securing an EHR cloud. We illustrate the development of the EHR security reference model through a use-case scenario and describe the corresponding security countermeasures and state of art security techniques that can be applied as basic security guards.

This paper presents MAPS - a personalized Multi-Attribute Probabilistic Selection framework - to estimate the probability of an item being a user's best choice and rank the items accordingly. The MAPS framework makes three original contributions in this paper. First, we capture the inter-attribute t

Multicast is a common platform for supporting group communication applications, such as IPTV, multimedia content delivery, and location-based advertisements. Distributed hash table (DHT) based overlay networks such as Chord and CAN presents a popular distributed computing architecture for multicast

Operational networks typically generate massive monitoring data that consist of local (in both space and time) observations of the status of the networks. It is often hypothesized that such data exhibit both spatial and temporal correlation based on the underlying network topology and time of occurrence; identifying such correlation patterns offers valuable insights into global network phenomena (e.g., fault cascading in communication networks). In this paper we introduce a new class of models suitable for learning, indexing, and identifying spatio-temporal patterns in network monitoring data. We exemplify our techniques with the application of fault diagnosis in enterprise networks. We show how it can help network management systems (NMSes) to effciently detect and localize potential faults (e.g., failure of routing protocols or network equipments) by analyzing massive operational event streams (e.g., alerts, alarms, and metrics). We provide results from extensive experimental studies over real network event and topology datasets to explore the effcacy of our solution.

We present the design and implementation of a new indexing technique, Mondrian tree. The Mondrian tree indexing method partitions the entire universe of discourse into spatial alarm monitoring regions and alarm-free regions. This enables us to reduce the number of on-demand alarm-free region computations, significant savings of both server load and client-to-server communication cost. We evaluate the efficiency of the Mondrian tree indexing approach and show that the Mondrian tree offers significant performance enhancements on spatial alarm processing at both the server side and the client side.

Wireless Positioning Systems (WPS) are popular alternative localization methods, especially in dense urban areas where GPS has known limitations. Map-matching (MM) has been used as an approach to improve the accuracy of the estimated locations of WiFi Access Points (APs), and thus the accuracy of a wireless positioning system. Large-scale wireless positioning differs from satellite based positioning in at least two aspects: First, wireless positioning systems typically derive the location estimates based on war-driving access point (AP) data. Second, the locations of the AP beacons are not generally known at the same precision as that of the satellite locations. This results in lower accuracy and a lower confidence factor in the use of wireless positioning. This paper presents a fast selective look-ahead map-matching technique, called SLAMM. Existing MM algorithms developed for real-time location tracking of a moving vehicle are ill-suited for matching large collections of war-driving data due to the time complexity. Another unique feature of SLAMM is the map-matching of critical location samples in an AP trace to the road network before matching non-critical samples. Our experiments over a real dataset of 70 million AP samples show that SLAMM is accurate and significantly faster than the traditional MM approaches.

Monitoring global states of an application deployed over distributed nodes becomes prevalent in today's datacenters. State monitoring requires not only correct monitoring results but also minimum communication cost for efficiency and scalability. Most existing work adopts an instantaneous state monitoring approach, which triggers state alerts whenever a constraint is violated. Such an approach, however, may cause frequent and unnecessary state alerts due to unpredictable monitored value bursts and momentary outliers that are common in large-scale Internet applications. These false alerts may further lead to expensive and problematic counter-measures. To address this issue, we introduce window-based state monitoring in this paper. Window-based state monitoring evaluates whether state violation is continuous within a time window, and thus, gains immunity to short-term value bursts and outliers. Furthermore, we find that exploring the monitoring time window at distributed nodes achieves significant communication savings over instantaneous monitoring. Based on this finding, we develop WISE, a system that efficiently performs WIndow-based StatE monitoring at datacenter-scale. WISE is highlighted with three sets of techniques. First, WISE uses distributed filtering time windows and intelligently avoids global information collecting to achieve communication efficiency, while guaranteeing monitoring correctness at the same time. Second, WISE provides a suite of performance tuning techniques to minimize communication cost based on a sophisticated cost model. Third, WISE also employs a set of novel performance optimization techniques. Extensive experiments over both real world and synthetic traces show that WISE achieves a 50% - 90% reduction in communication cost compared with existing instantaneous monitoring approaches and simple alternative schemes.

As a severe threat in anonymized data publication, proximity breach is gaining increasing attention. Such breach occurs when an attacker learns with high confidence that the sensitive information of a victim associates with a set of semantically proximate values, even though not sure about the exact one. Recently (¿, ¿)-dissimilarity [14] has been proposed as an effective countermeasure against general proximity attack. In this paper, we present a detailed analytical study on the fulfillment of this principle, derive criteria to efficiently test its satisfiability for given microdata, and point to a novel anonymization model, XCOLOR, with theoretical guarantees on both operation efficiency and utility preservation.

IPTV has emerged as the future standard of television and drawn enormous attention from both industry and research communities. Among different IPTV services, on-demand services are the most popular ones due to their convenience and rich content. However, supporting scalable and reliable on-demand IPTV services remains to be an important challenge. Existing IPTV architecture dedicates a centralized regional station to serve subscribers in the respective region regardless of temporal and spatial dynamics in service demand. As a result, it may cause significant imbalance of resource utilization and service provisioning delay at different stations, especially with increasing subscribers and video content. In this paper, we propose to allow IPTV stations of different regions to collaboratively serve user requests for delivering scalable and reliable IPTV services. One key challenge in achieving this station-wise collaboration is to route service requests to appropriate stations according to cost-effectiveness and load distribution in a fully distributed manner. We devise a novel request dispatching protocol which runs on each IPTV station, and yet forms a collaborative dispatching strategy that avoids hot spots and reduces service delivery cost at the same time. Our experiment results suggest that our service request dispatching algorithm significantly improves the scalability of on-demand IPTV services for the existing IPTV architecture.

The significant IO improvements of Solid State Disks (SSD) over traditional rotational hard disks makes it an attractive approach to integrate SSDs in tiered storage systems for performance enhancement. However, to integrate SSD into multi-tiered storage system effectively, automated data migration between SSD and HDD plays a critical role. In many real world application scenarios like banking and supermarket environments, workload and IO profile present interesting characteristics and also bear the constraint of workload deadline. How to fully release the power of data migration while guaranteeing the migration deadline is critical to maximizing the performance of SSD-enabled multi-tiered storage system. In this paper, we present an automated, deadline-aware, lookahead migration scheme to address the data migration challenge. We analyze the factors that may impact on the performance of lookahead migration efficiency and develop a greedy algorithm to adaptively determine the optimal lookahead window size to optimize the effectiveness of lookahead migration, aiming at improving overall system performance and resource utilization while meeting workload deadlines. We compare our lookahead migration approach with the basic migration model and validate the effectiveness and efficiency of our adaptive lookahead migration approach through a trace driven experimental study.

Mobile commerce and location based services (LBS) are some of the fastest growing IT industries in the last five years. Location update of mobile clients is a fundamental capability in mobile commerce and all types of LBS. Higher update frequency leads to higher accuracy, but incurs unacceptably high cost of location management at the location servers. We propose RoadTrack -- a road-network based, query-aware location update framework with two unique features. First, we introduce the concept of precincts to control the granularity of location update resolution for mobile clients that are not of interest to any active location query services. Second, we define query encounter points for mobile objects that are targets of active location query services, and utilize these encounter points to define the adequate location update schedule for each mobile. The RoadTrack framework offers three unique advantages. First, encounter points as a fundamental query awareness mechanism enable us to control and differentiate location update strategies for mobile clients in the vicinity of active location queries, while meeting the needs of location query evaluation. Second, we employ system-defined precincts to manage the desired spatial resolution of location updates for different mobile clients and to control the scope of query awareness to be capitalized by a location update strategy. Third, our road-network based check-free interval optimization further enhances the effectiveness of the Road-Track query-aware location update scheduling algorithm. This optimization provides significant cost reduction for location update management at both mobile clients and location servers. We evaluate the RoadTrack location update approach using a real world road-network based mobility simulator. Our experimental results demonstrate that the RoadTrack query aware location update approach outperforms existing representative location update strategies in terms of both client energy efficiency and server processing load.

With the popularity of voting systems in cyberspace, there is growing evidence that current voting systems can be manipulated by fake votes. This problem has attracted many researchers working on guarding voting systems in two areas: relieving the effect of dishonest votes by evaluating the trust of voters, and limiting the resources that can be used by attackers, such as the number of voters and the number of votes. In this paper, we argue that powering voting systems with trust and limiting attack resources are not enough. We present a novel attack named as Reputation Trap (RepTrap). Our case study and experiments show that this new attack needs much less resources to manipulate the voting systems and has a much higher success rate compared with existing attacks. We further identify the reasons behind this attack and propose two defense schemes accordingly. In the first scheme, we hide correlation knowledge from attackers to reduce their chance to affect the honest voters. In the second scheme, we introduce robustness-of-evidence, a new metric, in trust calculation to reduce their effect on honest voters. We conduct extensive experiments to validate our approach. The results show that our defense schemes not only can reduce the success rate of attacks but also significantly increase the amount of resources an adversary needs to launch a successful attack.

International audience

Within collaborative computing, computer mediated communications are evolving rapidly thanks to the development of new technologies. The facilitation of awareness and discovery of users in the communications networks is a key requirement for the success of these collaborative systems. Besides the ne

Spatial alarms are considered as one of the basic capabilities in future mobile computing systems for enabling personalization of location-based services. In this paper, we propose a distributed architecture and a suite of safe region techniques for scalable processing of spatial alarms. We show that safe region-based processing enables resource optimal distribution of partial alarm processing tasks from the server to the mobile clients. We propose three different safe region computation algorithms to explore the impact of size and shape of the safe region on network bandwidth, server load and client energy consumption. Concretely, we show that the maximum weighted perimeter rectangular safe region approach outperforms previous techniques in terms of performance and accuracy. We further explore finer granularity safe regions by introducing grid-based and pyramid-based representation of rectilinear polygonal shapes using bitmap encoding. Our experimental evaluation shows that the distributed safe region-based architecture outperforms the two most popular server-centric approaches, periodic and safe period-based, for spatial alarm processing.

To observe, analyze and control large scale distributed systems and the applications hosted on them, there is an increasing need to continuously monitor performance attributes of distributed system and application states. This results in application state monitoring tasks that require fine-grained attribute information to be collected from relevant nodes efficiently. Existing approaches either treat multiple application state monitoring tasks independently and build ad-hoc monitoring trees for each task, or construct a single static monitoring tree for multiple tasks. We argue that a careful planning of multiple application state monitoring tasks by jointly considering multi-task optimization and node level resource constraints can provide significant gains in performance and scalability. In this paper, we present REMO, a REsource-aware application state MOnitoring system. REMO produces a forest of optimized monitoring trees through iterations of two phases, one phase exploring cost sharing opportunities via estimation and the other refining the monitoring plan through resource-sensitive tree construction. Our experimental results include those gathered by deploying REMO on a BlueGene/P rack running IBM's large-scale distributed streaming system - System S. Using REMO running over 200 monitoring tasks for an application deployed across 200 nodes results in a 35%-45% decrease in the percentage error of collected attributes compared to existing schemes.

This work presents a systematic study of the problem of protecting general proximity privacy, with findings applicable to most existing data models. Our contributions are multi-folded: we highlighted and formulated proximity privacy breaches in a data-model-neutral manner; we proposed a new privacy principle (epsiv,delta)k-dissimilarity, with theoretically guaranteed protection against linking attacks in terms of both exact and proximate QI-SA associations; we provided a theoretical analysis regarding the satisfiability of (epsiv,delta)k-dissimilarity, and pointed to promising solutions to fulfilling this principle.

DHTs are scalable, self-organizing, and adaptive to underlying topology changes, thus being a promising infrastructure for realizing efficient Web service discovery. Range queries play an important role in service discovery, and in recent years a number of DHT-based range query schemes have been proposed. However, most of them suffer from high query delay and high processing cost. This paper presents ERQ, an Efficient scheme for delay bounded Range Query processing over DHTs. We first emulate the PHT structure and design a balanced Kautz (BK) tree to uniformly map the m-dimensional data space onto DHT nodes, and then present a novel algorithm that processes range queries in a parallel fashion, where an on-the-fly space pruning mechanism is adopted to reduce the processing cost. In a DHT with N nodes, ERQ can answer any range query in less than logN (2loglogN+1) hops with low processing cost, irrespective of the queried range, the whole space size, or the number of queried attributes. The effectiveness of ERQ is demonstrated through extensive experiments.

One of the critical challenges for service oriented computing systems is the capability to guarantee scalable and reliable service provision. This paper presents Reliable GeoGrid, a decentralized service computing architecture based on geographical location aware overlay network for supporting reliable and scalable mobile information delivery services. The reliable GeoGrid approach offers two distinct features. First, we develop a distributed replication scheme, aiming at providing scalable and reliable processing of location service requests in decentralized pervasive computing environments. Our replica management operates on a network of heterogeneous nodes and utilizes a shortcut-based optimization to increase the resilience of the system against node failures and network failures. Second, we devise a dynamic load balancing technique that exploits the service processing capabilities of replicas to scale the system in anticipation of unexpected workload changes and node failures by taking into account of node heterogeneity, network proximity, and changing workload at each node. Our experimental evaluation shows that the reliable GeoGrid architecture is highly scalable under changing service workloads with moving hotspots and highly reliable in the presence of both individual node failures and massive node failures.

Peer-to-peer VoIP (voice over IP) networks, exemplified by Skype, are becoming increasingly popular due to their significant cost advantage and richer call forwarding features than traditional public switched telephone networks. One of the most important features of a VoIP network is privacy (for VoIP clients). Unfortunately, most peer-to-peer VoIP networks neither provide personalization nor guarantee a quantifiable privacy level. In this paper we propose novel flow analysis attacks that demonstrate the vulnerabilities of peer-to-peer VoIP networks to privacy attacks. We present detailed experimental evaluation that demonstrates these attacks quantifying performance and scalability degradation.

Modern communication networks generate massive volume of operational event data, e.g., alarm, alert, and metrics, which can be used by a network management system (NMS) to diagnose potential faults. In this work, we introduce a new class of indexable fault signatures that encode temporal evolution of events generated by a network fault as well as topological relationships among the nodes where these events occur. We present an efficient learning algorithm to extract such fault signatures from noisy historical event data, and with the help of novel space-time indexing structures, we show how to perform efficient, online signature matching. We provide results from extensive experimental studies to explore the efficacy of our approach and point out potential applications of such signatures for many different types of networks including social and information networks.

Consider a mobile client who travels over roads and wishes to receive location-based services (LBS) from untrusted service providers. How might the user obtain such services without exposing her private position information? Meanwhile, how could the privacy protection mechanism incur no disincentive, e.g., excessive computation or communication cost, for any service provider or mobile user to participate in such a scheme? We detail this problem and present a general model for privacy-aware mobile services. A series of key features distinguish our solution from existing ones: a) it adopts the network-constrained mobility model (instead of the conventional random-waypoint model) to capture the privacy vulnerability of mobile users; b) it regards the attack resilience (for mobile users) and the query-processing cost (for service providers) as two critical measures for designing location privatization solutions, and provides corresponding analytical models; c) it proposes a robust and scalable location anonymization model, XStar, which best leverages the two measures; d) it introduces multi-folded optimizations in implementing XStar, which lead to further performance improvement. A comprehensive experimental evaluation is conducted to validate the analytical models and the efficacy of XStar.

Link-based analysis of the Web provides the basis for many important applications-like Web search, Web-based data mining, and Web page categorization-that bring order to the massive amount of distributed Web content. Due to the overwhelming reliance on these important applications, there is a rise in efforts to manipulate (or spam) the link structure of the Web. In this manuscript, we present a parameterized framework for link analysis of the Web that promotes spam resilience through a source-centric view of the Web. We provide a rigorous study of the set of critical parameters that can impact source-centric link analysis and propose the novel notion of influence throttling for countering the influence of link-based manipulation. Through formal analysis and a large-scale experimental study, we show how different parameter settings may impact the time complexity, stability, and spam resilience of Web link analysis. Concretely, we find that the source-centric model supports more effective and robust rankings in comparison with existing Web algorithms such as PageRank.

In multiparty collaborative data mining, participants contribute their own data sets and hope to collaboratively mine a comprehensive model based on the pooled data set. How to efficiently mine a quality model without breaching each party's privacy is the major challenge. In this paper, we propose an approach based on geometric data perturbation and data mining service-oriented framework. The key problem of applying geometric data perturbation in multiparty collaborative mining is to securely unify multiple geometric perturbations that are preferred by different parties, respectively. We have developed three protocols for perturbation unification. Our approach has three unique features compared to the existing approaches: with geometric data perturbation, these protocols can work for many existing popular data mining algorithms, while most of other approaches are only designed for a particular mining algorithm; both the two major factors: data utility and privacy guarantee are well preserved, compared to other perturbation-based approaches; and two of the three proposed protocols also have great scalability in terms of the number of participants, while many existing cryptographic approaches consider only two or a few more participants. We also study different features of the three protocols and show the advantages of different protocols in experiments.

This paper addresses the problem of optimizing multiple distributed stream queries that are executing simultaneously in distributed data stream systems. We argue that the static query optimization approach of "plan, then deployment" is inadequate for handling distributed queries involving multiple streams and node dynamics faced in distributed data stream systems and applications. Thus, the selection of an optimal execution plan in such dynamic and networked computing systems must consider operator ordering, reuse, network placement, and search space reduction. We propose to use hierarchical network partitions to exploit various opportunities for operator-level reuse while utilizing network characteristics to maintain a manageable search space during query planning and deployment. We develop top-down, bottom-up, and hybrid algorithms for exploiting operator-level reuse through hierarchical network partitions. Formal analysis is presented to establish the bounds on the search space and suboptimality of our algorithms. We have implemented our algorithms in the IFLOW system, an adaptive distributed stream management system. Through simulations and experiments using a prototype deployed on Emulab, we demonstrate the effectiveness of our framework and our algorithms.

Serverless distributed computing has received significant attention from both the industry and the research community. Among the most popular applications are the wide area network file systems, exemplified by CFS, Farsite and OceanStore. These file systems store files on a large collection of untrusted nodes that form an overlay network. They use cryptographic techniques to maintain file confidentiality and integrity from malicious nodes. Unfortunately, cryptographic techniques cannot protect a file holder from a Denial-of-Service (DoS) or a host compromise attack. Hence, most of these distributed file systems are vulnerable to targeted file attacks, wherein an adversary attempts to attack a small (chosen) set of files by attacking the nodes that host them. This paper presents LocationGuard - a location hiding technique for securing overlay file storage systems from targeted file attacks. LocationGuard has three essential components: (i) location key, (ii) routing guard, a secure algorithm that protects accesses to a file in the overlay network given its location key, and (iii) a set of location inference guards. Our experimental results quantify the overhead of employing LocationGuard and demonstrate its effectiveness against DoS attacks, host compromise attacks and various location inference attacks.

With fast-paced growth of digital data and exploding storage management costs, enterprises are looking for new ways to effectively manage their data. One such cost-effective paradigm is the cloud storage model also referred to as Storage-as-a-Service, in which enterprises outsource their storage to a storage service provider (SSP) by storing data (usually encrypted) at a remote SSP-managed site and accessing it over a high speed network. Along with storage capacity used, the SSP often charges clients on the amount of data that is accessed from the SSP site. Thus, it is in the interest of the client enterprise to download only relevant content. This makes search over outsourced storage an important capability. Searching over encrypted outsourced storage, however, is a complex challenge. Each enterprise has different access privileges for different users and this access control needs to be preserved during search (for example, ensuring that a user cannot search through data that is inaccessible from the filesystem due to its permissions). Secondly, the search mechanism has to preserve confidentiality from the SSP and indices can not be stored in plain text. In this article, we present a new filesystem search technique that integrates access control and indexing/search mechanisms into a unified framework to support access control aware search. Our approach performs indexing within the trusted enterprise domain and uses a novel access control barrel (ACB) primitive to encapsulate access control within these indices. The indices are then systematically encrypted and shipped to the SSP for hosting. Unlike existing enterprise search techniques, our approach is resilient to various common attacks that leak private information. Additionally, to the best of our knowledge, our approach is a first such technique that allows search indices to be hosted at the SSP site, thus effectively providing search-as-a-service . This does not require the client enterprise to fully trust the SSP for data confidentiality. We describe the architecture and implementation of our approach and a detailed experimental analysis comparing with other approaches.

Most proposed DHTs have their unique maintenance mechanisms specific to the static graphs on which they are based. In this paper we propose distributed line graphs (DLG), a universal framework for building DHTs based on arbitrary constant-degree graphs. We prove that in a DLG-enabled, N-node DHT, the out-degree is d, the in-degree is between 1 and 2d, and the diameter is less than 2(logdN-logdN0+D0+1), where d, D0and N0represent the degree, diameter and number of nodes of the initial graph, respectively. The maintenance cost of DLG-enabled DHTs is O(logdN). We show the power of DLG technique by applying it to Kautz graphs to propose a new DHT scheme.

Freshness and accuracy are two key measures of quality of service (QoS) in location-based, mobile continual queries (CQs). However, it is often difficult to provide both fresh and accurate CQ results due to (a) limited resources in computing and communication and (b) fast-changing load conditions caused by continuous mobile node movement. Thus a key challenge for a mobile CQ system is: How do we achieve the highest possible quality of the query results, in both freshness and accuracy, with currently available resources under changing load conditions? In this paper, we formulate this problem as a load shedding one, and develop MobiQual - a QoS-aware framework for performing both update load shedding and query load shedding. The design of MobiQual highlights three important features. (1)Differentiatedloadshedding: Different amounts of query and update load shedding are applied to different groups of queries and mobile nodes, respectively. (2)Per-queryQoSspecifications: The overall freshness and accuracy of the query results are maximized with individualized QoS specifications. (3)Low-costadaptation: MobiQual dynamically adapts, with a minimal overhead, to changing load conditions and available resources. We show that, through a careful combination of update and query load shedding, the MobiQual approach leads to much higher freshness and accuracy in the query results in all cases, compared to existing approaches.

With fast paced growth of digital data and exploding storage management costs, enterprises are looking for new ways to effectively manage their data. One such cost-effective paradigm is the storage-as-a-service model, in which enterprises outsource their storage to a storage service provider (SSP) by storing data at a remote SSP-managed site and accessing it over a high speed network. Often for a variety of reasons, enterprises find it unacceptable to fully trust the SSP and prefer to store data in an encrypted form. This typically limits collaboration and data sharing among enterprise users due to complex key management and access control challenges. In this paper, we propose a platform called SHAROES that provides data sharing capability over such outsourced storage environments. SHAROES provide rich *nix-like data sharing semantics over SSP stored data, without trusting the SSP for data confidentiality or access control. SHAROES is unique in its ability in reducing user involvement during setup and operation through the use of in-band key management and allows a near-seamless transition of existing storage environments to the new model. It is also superior in performance by minimizing the use of expensive public-key cryptography in metadata management. We present the architecture and implementation of various SHAROES components and our experiments demonstrate performance superior to other proposals by over 40% on a number of benchmarks.

Privacy preservation in data mining demands protecting both input and output privacy. The former refers to sanitizing the raw data itself before performing mining. The latter refers to preventing the mining output (model/pattern) from malicious pattern-based inference attacks. The preservation of input privacy does not necessarily lead to that of output privacy. This work studies the problem of protecting output privacy in the context of frequent pattern mining over data streams. After exposing the privacy breaches existing in current stream mining systems, we propose Butterfly, a light-weighted countermeasure that can effectively eliminate these breaches without explicitly detecting them, meanwhile minimizing the loss of the output accuracy. We further optimize the basic scheme by taking account of two types of semantic constraints, aiming at maximally preserving utility-related semantics while maintaining the hard privacy and accuracy guarantee. We conduct extensive experiments over real- life datasets to show the effectiveness and efficiency of our approach.

Effective management of Web Services systems relies on accurate understanding of end-to-end transaction flows, which may change over time as the service composition evolves. This work takes a data mining approach to automatically recovering end-to-end transaction flows from (potentially obscure) monitoring events produced by monitoring tools. We classify the caller-callee relationships among monitoring events into three categories(identity, direct-invoke, and cascaded-invoke), and propose unsupervised learning algorithms to generate rules for each type of relationship. The key idea is to leverage the temporal information available in the monitoring data and extract patterns that have statistical significance. By piecing together the caller-callee relationships a teach step along the invocation path, we can recover the end-to-end flow for every executed transaction. Experiments demonstrate that our algorithms outperform human experts in terms of solution quality, scale well with the data size, and are robust against noises in monitoring data.

One important problem for public broadcast Location-Based Services (LBS) is to enforce access control on a large number of subscribers. In such a system a user typically subscribes to a LBS for a time interval (a, b) and a spatial region (xbl, ybl, xir, ytr) according to a 3-dimensional spatial-temporal authorization model. In this paper, we argue that current approaches to access control using group key management protocols are not scalable. Our proposal STauth minimizes the number of keys which needs to be distributed and is thus scalable to a much higher number of subscribers and the dimensionality of the authorization model. We analytically and experimentally demonstrate the performance and scalability benefits of our approach against other group key management protocols.

Applications such as VoIP need to provide anonymity to clients while maintaining low latency to satisfy quality of service (QoS) requirements. Existing solutions for providing anonymity such as mix networks are not well suited to applications like VoIP, SSH, and gaming which require low communication latency. This paper investigates the problem of on-demand construction of QoS sensitive routes on anonymizing networks using the VoIP application. We first describe triangulation based timing analysis attacks on shortest path route set up protocols. We show that even when a small fraction (~1%) of the network is malicious, the adversary can infer the source (caller) with reasonably high probability. Second, we describe random walk based route set up protocols that significantly improve anonymity while satisfying latency- based QoS guarantees. We describe a prototype implementation of our proposal and show that our protocols can significantly reduce the probability of inferring the caller. We present a detailed experimental evaluation to demonstrate our attacks and quantify the performance and scalability of our guards.

This paper presents PrivacyGrid - a framework for supporting anonymous location-based queries in mobile information delivery systems. The PrivacyGrid framework offers three unique capabilities. First, it provides a location privacy protection preference profile model, called location P3P, which allows mobile users to explicitly define their preferred location privacy requirements in terms of both location hiding measures (e.g., location k-anonymity and location l-diversity) and location service quality measures (e.g., maximum spatial resolution and maximum temporal resolution). Second, it provides fast and effective location cloaking algorithms for location k-anonymity and location l-diversity in a mobile environment. We develop dynamic bottom-up and top-down grid cloaking algorithms with the goal of achieving high anonymization success rate and efficiency in terms of both time complexity and maintenance cost. A hybrid approach that carefully combines the strengths of both bottom-up and top-down cloaking approaches to further reduce the average anonymization time is also developed. Last but not the least, PrivacyGrid incorporates temporal cloaking into the location cloaking process to further increase the success rate of location anonymization. We also discuss PrivacyGrid mechanisms for supporting anonymous location queries. Experimental evaluation shows that the PrivacyGrid approach can provide close to optimal location k-anonymity as defined by per user location P3P without introducing significant performance penalties.

We propose the SocialTrust framework for tamper-resilient trust establishment in online social networks. Two of the salient features of SocialTrust are its dynamic revision of trust by (i) distinguishing relationship quality from trust; and (ii) incorporating a personalized feedback mechanism for adapting as the social network evolves.

This paper investigates the strategic decisions of online vendors for offering different mechanism, such as sampling and online reviews of information products, to increase their online sales. Focusing on measuring the effectiveness of electronic market design (offering reviews, sampling, or both), our study shows that online markets behavior as communication markets, and consumers learn product quality information both passively (reading online reviews) and actively but subjectively (listening to music sampling). Using data from Amazon, first we show that sampling along is a strong product quality signal that reduces the product uncertainty after controlling for halo effect. In general, products with sampling option enjoy a higher conversion rate (which leads to better sales) than those without sampling because sampling decreases the uncertainty of consuming experience goods. Second, the impact of online reviews on sales conversion rate is lower for experience goods with a sampling option than those without. Third, when the uncertainty of the societal reviews is higher, sampling plays a more important role because it mitigates such uncertainty introduced by online reviews.

In this paper, we analyse the data sharing mechanisms of *nix systems and identify an immediate need for better privacy support. For example, using a simple insider attack we were able to access over 84 GB of private data at one organisation of 825 users, including 300 000 e-mails and 579 passwords to financial and other private services websites, without exploiting any technical vulnerability. We present two solutions to address this problem: 1. an administrative auditing tool which can alert administrators and users when their private data is at risk; 2. a new View Based Access Control (VBAC) mechanism which provides stronger and yet convenient privacy support. We also describe a proof-of-concept filesystem-based implementation and performance analysis of VBAC. Our evaluations with three well-known filesystem benchmarks show little overhead of using VBAC.

Continued advances in mobile networks and positioning technologies have created a strong market push for location-based applications. Examples include location-aware emergency response, location-based advertisement, and location-based entertainment. An important challenge in wide deployment of location-based services (LBSs) is the privacy-aware management of location information, providing safeguards for location privacy of mobile clients against vulnerabilities for abuse. This paper describes a scalable architecture for protecting location privacy from various privacy threats resulting from uncontrolled usage of LBSs. This architecture includes the development of a personalized location anonymization model and a suite of location perturbation algorithms. A unique characteristic of our location privacy architecture is the use of a flexible privacy personalization framework to support location k-anonymity for a wide range of mobile clients with context-sensitive privacy requirements. This framework enables each mobile client to specify the minimum level of anonymity it desires and the maximum temporal and spatial tolerances it is willing to accept when requesting for k-anonymity preserving LBSs. We devise an efficient message perturbation engine to implement the proposed location privacy framework. The prototype we develop is designed to be run by the anonymity server on a trusted platform and performs location anonymization on LBS request messages of mobile clients, such as identity removal and spatio-temporal cloaking of location information. We study the effectiveness of our location cloaking algorithms under various conditions using realistic location data that is synthetically generated from real road maps and traffic volume data. Our experiments show that the personalized location k-anonymity model together with our location perturbation engine can achieve high resilience to location privacy threats without introducing any significant performance penalty.

Recently, we have seen increasing numbers of denial of service (DoS) attacks against online services and Web applications either for extortion reasons or for impairing and even disabling the competition. These DoS attacks have increasingly targeted the application level. Application-level DoS attacks emulate the same request syntax and network-level traffic characteristics as those of legitimate clients, thereby making the attacks much harder to detect and counter. Moreover, such attacks often target bottleneck resources such as disk bandwidth, database bandwidth, and CPU resources. In this article, we propose handling DoS attacks by using a twofold mechanism. First, we perform admission control to limit the number of concurrent clients served by the online service. Admission control is based on port hiding that renders the online service invisible to unauthorized clients by hiding the port number on which the service accepts incoming requests. Second, we perform congestion control on admitted clients to allocate more resources to good clients. Congestion control is achieved by adaptively setting a client's priority level in response to the client's requests in a way that can incorporate application-level semantics. We present a detailed evaluation of the proposed solution using two sample applications: Apache HTTPD and the TPCW benchmark (running on Apache Tomcat and IBM DB2). Our experiments show that the proposed solution incurs low performance overhead and is resilient to DoS attacks.

Reputation systems provide a promising way for building trust through social control in collaborative communities by harnessing the community knowledge in the form of feedback. However, reputation systems also introduce vulnerabilities due to potential manipulations by dishonest or malicious players. In this paper, we focus on two closely related problems - feedback sparsity and potential feedback manipulations - and propose a feedback similarity based inference framework. We perform extensive evaluations of various algorithmic components of the framework and evaluate their effectiveness on countering feedback sparsity in the presence of feedback manipulations.

Secure event dissemination in a pub-sub network refers to secure distribution of events to clients subscribing to those events without revealing the secret attributes in the event to the unauthorized subscribers and the routing nodes in a pub-sub network. A common solution to provide confidentiality guarantees for the secret attributes in an event is to encrypt so that only authorized subscribers can read them. The key challenge here is to build a secure and scalable content-based event dissemination infrastructure that can handle complex and flexible subscription models while preserving the efficiency and scalability of key management algorithms. In this paper, we describe the design and implementation of PSGuard, for secure event dissemination in pub-sub networks. PSGuard exploit hierarchical key derivation algorithms to encode publication-subscription matching semantics for scalable key management. An experimental evaluation of our prototype system shows that PSGuard meets the security requirements while maintaining the performance and scalability of a pub-sub network.

This paper presents GeoGrid, a geographical location service overlay network system, for providing scalable location-based services to a large and growing number of mobile users. GeoGrid is designed as a decentralized and geographical location aware overlay network and provides system-level facilities and optimizations for balancing load in the presence of node heterogeneity, dynamically moving hot-spots (location queries), and unpredictable rate of node join, departure and failure. GeoGrid uses geographical mapping of nodes to regions and geographical proximity based routing to take advantage of the similarity between physical and network proximity. Furthermore, GeoGrid exploits multiple opportunities for dynamic workload adaptation in the presence of static hotspot queries and moving hotspot queries. Its dynamic load balancing algorithms can efficiently utilize the heterogeneous capacities of end systems and balance both the location query workload and the routing workload.Our initial prototype development and experimental study demonstrate that GeoGrid can effectively reduce the workload imbalance by an order of magnitude.

We describe DSphere - a decentralized system for crawling, indexing, searching and ranking of documents in the World Wide Web. Unlike most of the existing search technologies that depend heavily on a page-centric view of the Web, we advocate a source-centric view of the Web and propose a decentralized architecture for crawling, indexing and searching the Web in a distributed source-specific fashion. A fully decentralized crawler is developed to crawl the World Wide Web where each peer is assigned the responsibility of crawling a specific set of documents referred to as a source collection. Link analysis techniques are used for ranking documents. Traditional link analysis techniques suffer from problems like slow refresh rate and vulnerabilities to Web Spam. We propose a source-based link analysis approach, which computes fast and accurate ranking scores for all crawled documents.

To provide high-quality results for location-based, continual queries (CQs) in a mobile system, the query processor usually demands receiving frequent position updates from the mobile nodes. However, processing frequent updates often causes the query processor to become overloaded, under which updates must be dropped randomly, bringing down the query-result accuracy and negating the benefits of frequent updates. In this paper, we develop LIRA - a lightweight, region-aware load-shedding technique for preventively reducing the position-update load of a query processor, while maintaining high-quality query results. Instead of receiving too many updates and then randomly dropping some of them, LIRA uses a region-aware partitioning mechanism to identify the most beneficial shedding regions to cut down the position updates sent by the mobile nodes within those regions. Based on the densities of mobile nodes and queries in a region, LIRA judiciously applies different amounts of update reduction for different regions, aiming to minimize the negative impacts of load shedding on query-result accuracy. Experimental results show that LIRA is vastly superior to random update dropping and clearly outperforms other alternatives that do not possess region-aware load-shedding capabilities. Moreover, due to its lightweight nature, LIRA introduces very little overhead.

Tuple dropping, though commonly used for load shedding in most stream operations, is inadequate for m-way, windowed stream joins. The join output rate can be overly reduced because it fails to exploit the time correlations likely to exist among interrelated streams. In this paper, we introduce GrubJoin; an adaptive, m-way, windowed stream join that effectively performs time correlation-aware CPU load shedding. GrubJoin maximizes the output rate by achieving near-optimal window harvesting, which picks only the most profitable window segments for the join. Due to combinatorial explosion of possible m-way join sequences involving window segments, m-way, windowed stream joins pose several unique challenges. We focus on addressing two of them: (1) How can we quickly determine the optimal window harvesting configuration for any m-way, windowed stream join? (2) How can we monitor and learn the time correlations among the streams with high accuracy and minimal overhead? To tackle these challenges, we formalize window harvesting as an optimization problem, develop greedy heuristics to determine near-optimal window harvesting configurations and use approximation techniques to capture the time correlations. Our experimental results show that GrubJoin is vastly superior to tuple dropping when time correlations exist and is equally effective when time correlations are nonexistent.

Service computing has emerged as a primary platform for packaging and delivering computing and information technology to various enterprises ranging from business to governments and organizations. In recent years we have witnessed many emerging standards and industrial developments in the area of service computing.

With fast paced growth of digital data, keyword based search has become a critical enterprise application. Research has shown that nearly 85% of enterprise data lies in flat filesystems [10] that allow multiple users with different access privileges. Any search tool for such systems needs to be efficient and yet cognizant of access control semantics imposed by the underlying filesystem. Current enterprise search techniques use two disjoint search and access- control components by creating a single system-wide index and filtering search results for access control. This approach is ineffective as index and query statistics subtly leak private information. The other approach of using separate indices for each user is undesirable as it not only increases disk consumption due to shared files, but also increases overheads of updating indices whenever a file changes. We propose a distributed approach that couples search and access-control into a unified framework and pro vides secure multiuser search. Our scheme (logically) divides data into independent access-privileges based chunks, called access-control barrels (ACB). ACBs not only manage security but also improve overall efficiency as they can be indexed and searched in parallel by distributing them to multiple enterprise machines. We describe the architecture of ACBs based search and propose an optimization that ensures the scalability of our approach. We validate our design with a detailed evaluation using industry benchmarks and datasets. Our initial experiments show secure search with 38% improved indexing efficiency and low overheads for ACB processing.

Web search is one of the most critical applications for managing the massive amount of distributed Web content. Due to the overwhelming reliance on Web search, there is a rise in efforts to manipulate (or spam) Web search engines. In this paper, we develop a spam-resilient ranking model that promotes a source-based view of the Web. One of the most salient features of our spam-resilient ranking algorithm is the concept of influence throttling. We show how to utilize influence throttling to counter Web spam that aims at manipulating link-based ranking systems, especially PageRank-like systems. Through formal analysis and experimental evaluation, we show the effectiveness and robustness of our spam-resilient ranking model in comparison with existing Web algorithms such as PageRank.

We consider the problem of query optimization in distributed data stream systems where multiple continuous queries may be executing simultaneously. In order to achieve the best performance, query planning (such as join ordering) must be considered in conjunction with deployment planning (e.g., assigning operators to physical nodes with optimal ordering). However, such a combination involves not only a large number of network nodes but also many query operators, resulting in an extremely large search space for optimal solutions. Our paper aims at addressing this problem by utilizing hierarchical network partitions. We propose two algorithms - top-down and bottom-up which utilize hierarchical network partitions to provide scalable query optimization. Formal analysis is presented to establish the bounds on the search-space and to show the sub-optimality of our algorithms. Through simulations and experiments using a prototype deployed on Emulab we demonstrate the effectiveness of our algorithms.

With continued growth of music content available on the Internet, music information retrieval has attracted increasing attention. An important challenge for music searching is its ability to support both keyword and content based queries efficiently and with high precision. In this paper, we present a music query system - QueST (Query by acouStic and Textual features) to support both keyword and content based retrieval in large music databases. QueST has two distinct features. First, it provides new index schemes that can efficiently handle various queries within a uniform architecture. Concretely, we propose a hybrid structure consisting of Inverted file and Signature file to support keyword search. For content based query, we introduce the notion of similarity to capture various music semantics like melody and genre. We extract acoustic features from a music object, and map it to multiple high-dimension spaces with respect to the similarity notion using PCA and RBF neural network. Second, we design a result fusion scheme, called the Quick Threshold Algorithm, to speed up the processing of complex queries involving both textual and multiple acoustic features. Our experimental results show that QueST offers higher accuracy and efficiency compared to existing algorithms.

Determining the location of mobile devices is a necessary system function for a growing number of location based services (LBS). The most popular method for location determination is GPS, however GPS has known accuracy and environmental limitations, many of which are exacerbated in dense urban areas. Wireless positioning systems (WPS) on the other hand, demonstrate location accuracy largely inverted to that of GPS - the denser the urban setting, the more accurate the location estimation is in general. Large-scale WPS differs from satellite based positioning in at least two aspects; first, wireless positioning systems typically derive their location estimates based on observed beacon locations such as through wardriving, and second WPS lacks a mechanism to maintain highly synchronized clocks. This results in lower accuracy and a lower confidence factor in the use of wireless positioning. This paper presents a location estimation method that improves location accuracy for WPS through the use of digital map-matching of wardriving data. We have conducted initial experiments to evaluate our map-matching algorithm along with the enhanced location estimation approach and demonstrate its effectiveness for measuring and improving the accuracy of large-scale wireless positioning systems. We demonstrate extensions to a look-ahead map-matching algorithm to improve the accuracy and motivate further research in the area of large-scale map matching.

This paper proposes an adaptive spatio-temporal placement scheme for caching location dependent queries. The cache placement decision is made according to the potential cache benefit of the query results based on the spatio-temporal properties of query results and the movement patterns of the mobile client, aiming at increasing the cache hit ratio. We introduce the concept of 'overlapping cache benefit' as a measure of the hit rate of a cached item, and present three spatio-temporal cache placement schemes, which provide a step-by-step in-depth analysis of various factors that may affect the performance of a client cache in mobile environments. We implemented the spatio-temporal placement model in the first prototype of the MOBICACHE system. Our experimental evaluation shows that the spatial locality and the movement patterns of mobile clients are critical factors that impact the effectiveness of cache placement and the performance of client cache, and the proposed adaptive spatio-temporal cache placement approach yields higher hit ratio and better response time compared to existing mobile cache solutions.

We introduce the concept of link credibility, identify the conflation of page quality and link credibility in popular Web link analysis algorithms, and discuss how to decouple link credibility from page quality. Our credibility-based link analysis exhibits three distinct features. First, we develop several techniques for semi-automatically assessing link credibility for all Web pages. Second, our link credibility assignment algorithms allow users to assess credibility in a personalized manner. Third, we develop a novel credibility-based Web ranking algorithm - CredibleRank - which incorporates credibility information directly into the quality assessment of each page on the Web. Our experimental study shows that our approach is significantly and consistently more spam-resilient than both PageRank and TrustRank.

No abstract available.

Data perturbation is a popular technique for privacy-preserving data mining. The major challenge of data perturbation is balancing privacy protection and data quality, which are normally considered as a pair of contradictive factors. We propose that selectively preserving only the task/model specific information in perturbation would improve the balance. Geometric data perturbation, consisting of random rotation perturbation, random translation perturbation, and noise addition, aims at preserving the important geometric properties of a multidimensional dataset, while providing better privacy guarantee for data classification modeling. The preliminary study has shown that random geometric perturbation can well preserve model accuracy for several popular classification models, including kernel methods, linear classifiers, and SVM classifiers, while it also revealed some security concerns to random geometric perturbation. In this paper, we address some potential attacks to random geometric perturbation and design several methods to reduce the threat of these attacks. Experimental study shows that the enhanced geometric perturbation can provide satisfactory privacy guarantee while still well preserving model accuracy for the discussed data classification models.

Tuple dropping, though commonly used for load shedding in most data stream operations, is generally inadequate for multiway windowed stream joins. The join output rate can be unnecessarily reduced because tuple dropping fails to exploit the time correlations that are likely to exist among interrelated streams. In this paper, we introduce GrubJoin-an adaptive multiway windowed stream join that effectively performs time correlation-aware CPU load shedding. GrubJoin maximizes the output rate by achieving near-optimal window harvesting, which picks only the most profitable segments of individual windows for the join. Due mainly to the combinatorial explosion of possible multiway join sequences involving different window segments, GrubJoin faces unique challenges that do not exist for binary joins, such as determining the optimal window harvesting configuration in a time-efficient manner and learning the time correlations among the streams without introducing overhead. To tackle these challenges, we formalize window harvesting as an optimization problem, develop greedy heuristics to determine near-optimal window harvesting configurations, and use approximation techniques to capture the time correlations. Our experimental results show that GrubJoin is vastly superior to tuple dropping when time correlations exist and is equally effective when time correlations are nonexistent.

In recent years, edge computing has emerged as a popular mechanism to deliver dynamic Web content to clients. However, many existing edge cache networks have not been able to harness the full potential of edge computing technology. In this paper, we argue and experimentally demonstrate that cooperation among the individual edge caches coupled with scalable server-driven document consistency mechanisms can significantly enhance the capabilities and performance of edge cache networks in delivering fresh dynamic content. However, designing large-scale cooperative edge cache networks presents many research challenges. Toward addressing these challenges, this paper presents cooperative edge cache grid (cooperative EC grid, for short)-a large-scale cooperative edge cache network for efficiently delivering highly dynamic Web content with varying server update frequencies. The design of the cooperative EC grid focuses on the scalability and reliability of dynamic content delivery in addition to cache hit rates, and it incorporates several novel features. We introduce the concept of cache clouds as a generic framework of cooperation in large-scale edge cache networks. The architectural design of the cache clouds includes dynamic hashing-based document lookup and update protocols, which dynamically balance lookup and update loads among the caches in the cloud. We also present cooperative techniques for making the document lookup and update protocols resilient to the failures of individual caches. This paper reports a series of simulation-based experiments which show that the overheads of cooperation in the cooperative EC grid are very low, and our architecture and techniques enhance the performance of the cooperative edge networks.

Advances in distributed service-oriented computing and Internet technology have formed a strong technology push for outsourcing and information sharing. There is an increasing need for organizations to share their data across organization boundaries both within the country and with countries that may have lesser privacy and security standards. Ideally, we wish to share certain statistical data and extract the knowledge from the private databases without revealing any additional information of each individual database apart from the aggregate result that is permitted. In this article, we describe two scenarios for outsourcing data aggregation services and present a set of decentralized peer-to-peer protocols for supporting data sharing across multiple private databases while minimizing the data disclosure among individual parties. Our basic protocols include a set of novel probabilistic computation mechanisms for important primitive data aggregation operations across multiple private databases such as max, min, and top k selection. We provide an analytical study of our basic protocols in terms of precision, efficiency, and privacy characteristics. Our advanced protocols implement an efficient algorithm for performing k NN classification across multiple private databases. We provide a set of experiments to evaluate the proposed protocols in terms of their correctness, efficiency, and privacy characteristics.

One of the most prominent and comprehensive ways of data collection in sensor networks is to periodically extract raw sensor readings. This way of data collection enables complex analysis of data, which may not be possible with in-network aggregation or query processing. However, this flexibility in data analysis comes at the cost of power consumption. In this paper we develop ASAP − an adaptive sampling approach to energyefficient periodic data collection in sensor networks. The main idea behind ASAP is to use a dynamically changing subset of the nodes as samplers such that the sensor readings of the sampler nodes are directly collected, whereas the values of the non-sampler nodes are predicted through the use of probabilistic models that are locally and periodically constructed. ASAP can be effectively used to increase the network lifetime while keeping the quality of the collected data high, in scenarios where either the spatial density of the network deployment is superfluous relative to the required spatial resolution for data analysis or certain amount of data quality can be traded off in order to decrease the power consumption of the network. ASAP approach consists of three main mechanisms. First, sensing-driven cluster construction is used to create clusters within the network such that nodes with close sensor readings are assigned to the same clusters. Second, correlation-based sampler selection and model derivation are used to determine the sampler nodes and to calculate the parameters of the probabilistic models that capture the spatial and temporal correlations among the sensor readings. Last, adaptive data collection and model-based prediction are used to minimize the number of messages used to extract data from the network. A unique feature of ASAP is the use of in-network schemes, as opposed to the protocols requiring centralized control, to select and dynamically refine the subset of the sensor nodes serving as samplers and to adjust th

Distributed privacy preserving data mining tools are critical for mining multiple databases with a minimum information disclosure. We present a framework including a general model as well as multi-round algorithms for mining horizontally partitioned databases using a privacy preserving k Nearest Neighbor (kNN) classifier.

It is widely recognized that developing efficient and fully automated algorithms for clustering large transactional datasets is a challenging problem. In this paper, we propose a fast, memory-efficient, and scalable clustering algorithm for analyzing transactional data. Our approach has three unique features. First, we use the concept of Weighted Coverage Density as a categorical similarity measure for efficient clustering of transactional datasets. The concept of weighted coverage density is intuitive and allows the weight of each item in a cluster to be changed dynamically according to the occurrences of items. Second, we develop two transactional data clustering specific evaluation metrics based on the concept of large transactional items and the coverage density respectively. Third, we implement the weighted coverage density clustering algorithm and the two clustering validation metrics using a fully automated transactional clustering framework, called SCALE (Sampling, Clustering structure Assessment, cLustering and domain-specific Evaluation). The SCALE framework is designed to combine the weighted coverage density measure for clustering over a sample dataset with self-configuring methods that can automatically tune the two important parameters of the clustering algorithms: (1) the candidates of the best number K of clusters; and (2) the application of two domain-specific cluster validity measures to find the best result from the set of clustering results. We have conducted experimental evaluation using both synthetic and real datasets and our results show that the weighted coverage density approach powered by the SCALE framework can efficiently generate high quality clustering results in a fully automated manner.

Web ranking is one of the most successful and widely used collaborative computing applications, in which Web pages collaborate in the form of varying degree of relationships to assess their relative quality. Though many observe that links display strong source-centric locality, for example, in terms of administrative domains and hosts, most Web ranking analysis to date has focused on the flat page-level Web linkage structure. In this paper we develop a framework for link-based collaborative ranking of the Web by utilizing the strong Web link structure. We argue that this source-centric link analysis is promising since it captures the natural link-locality structure of the Web, can provide more appealing and efficient Web applications, and reflects many natural types of structured human collaborations. Concretely, we propose a generic framework for source-centric collaborative ranking of the Web. This paper makes two unique contributions. First, we provide a rigorous study of the set of critical parameters that can impact source-centric link analysis, such as source size, the presence of self-links, and different source-citation link weighting schemes (e.g., uniform, link count, source consensus). Second, we conduct a large-scale experimental study to understand how different parameter settings may impact the time complexity, stability, and spam-resilience of Web ranking. We find that careful tuning of these parameters is vital to ensure success over each objective and to balance the performance across all objectives

Cost-effective cooperation among a network of edge caches is widely accepted as an effective mechanism for enhancing the scalability, performance, and reliability of edge cache networks. However, the problem of how to form cache groups for achieving effective and efficient cooperation in edge cache networks has largely been unexplored. In this paper, we identify two important factors that need to be considered while forming cooperative groups, namely, network proximities of edge caches and network distances of the caches to the origin server. We propose two novel cache clustering schemes for accurately partitioning the caches of a given edge cache network into specified number of cache groups. The first scheme, called the Selective Landmarks scheme (SL scheme), accurately partitions the edge cache network into cooperative groups based on the network proximities of the caches. The second cache group formation scheme, called Server Distance sensitive Selective Landmarks scheme (SDSL scheme), provides a careful combination network proximities and server distances. Our experiments indicate that the proposed techniques can yield significant performance benefits.

Business processes continue to play an important role in today's service-oriented enterprise computing systems. Mining, discovering, and integrating process-oriented services has attracted growing attention in the recent year. In this paper we present a quantitative approach to modeling and capturing the similarity and dissimilarity between different process designs. We derive the similarity measures by analyzing the process dependency graphs of the participating workflow processes. We first convert each process dependency graph into a normalized process matrix. Then we calculate the metric space distance between the normalized matrices. This distance measure can be used as a quantitative and qualitative tool in process mining, process merging, and process clustering, and ultimately it can reduce or minimize the costs involved in design, analysis, and evolution of workflow systems

Analyzing clustering structures in data streams can provide critical information for making decision in realtime. In this paper, we present a framework for detecting the change of critical clustering structure in categorical data streams. The framework consists of the Hierarchical Entropy Tree structure (HE-Tree) and the extended ACE clustering algorithm. HE-Tree can efficiently capture the entropy property of the categorical data streams and allow us to draw precise clustering information from the data stream for high-quality BkPLots with the extended ACE algorithm.

We present an adaptive distributed query-sampling framework that is quality-conscious for extracting high-quality text database samples. The framework divides the query-based sampling process into an initial seed sampling phase and a quality-aware iterative sampling phase. In the second phase the sampling process is dynamically scheduled based on estimated database size and quality parameters derived during the previous sampling process. The unique characteristic of our adaptive query-based sampling framework is its self-learning and self-configuring ability based on the overall quality of all text databases under consideration. We introduce three quality-conscious sampling schemes for estimating database quality, and our initial results show that the proposed framework supports higher-quality document sampling than existing approaches.

We consider the problem of distributed continuous data delivery services in an overlay network of heterogeneous nodes. Each node in the system can be a source for any number of data streams and at the same time be a consumer node that is receiving streams sourced at other nodes. A consumer node may define a filter on a source stream such that only the desired portion of the stream is delivered, minimizing the amount of unnecessary bandwidth consumption. By heterogeneous, we mean that nodes not only may have varying network bandwidths and computing resources but also different interests in terms of the filters and the rates of the data streams they are interested in. Our objective is to construct an efficient stream delivery network in which nodes cooperate in forwarding data streams in the presence of constrained resources. We formalize this distributed stream delivery problem as an optimization one by starting with a simple setup where the network topology is fixed and node bandwidth characteristics are known. The goal of the optimization is to find valid delivery graphs with minimum bandwidth consumption. We extend this problem formulation to QoS-aware stream delivery, in order to handle the bandwidth constrained cases in which unwanted drops and delays are inevitable. We provide a classification of delivery graph construction schemes, and in light of this classification we develop pragmatic quality-aware stream delivery (QASD) algorithms. These algorithms aim at constructing efficient stream delivery graphs in a distributed setting, where global knowledge is not available and network characteristics are not known in advance. We introduce a set of evaluation metrics and provide experimental results to illustrate the effectiveness of our proposed algorithms under these metrics.

Denial of service (DoS) attacks attempt to consume a server's resources (network bandwidth, computing power, main memory, disk bandwidth etc.) to near exhaustion so that there are no resources left to handle requests from legitimate clients. An effective solution to defend against DoS attacks is to filter DoS attack requests at the earliest point (say, the Web site's firewall), before they consume much of the server's resources. Most defenses against DoS attacks attempt to filter requests from inauthentic clients before they consume much of the server's resources. Client authentication using techniques like IPSec or SSL may often require changes to the client-side software and may additionally require superuser privileges at the client for deployment. Further, using digital signatures (as in SSL) makes verification very expensive, thereby making the verification process itself a viable DoS target for the adversary. In this paper, we propose a light-weight client transparent technique to defend against DoS attacks with two unique features: (i) Our technique can be implemented entirely using JavaScript support provided by a standard client-side browser like Mozilla FireFox or Microsoft Internet Explorer. Client transparency follows from the fact that: (i) no changes to client-side software are required, (ii) no client-side superuser privileges are required, and (iii) clients (human beings or automated clients) can browse a DoS protected Web site in the same manner that they browse other Web sites, (ii) Although we operate using the client-side browser (HTTP layer), our technique enables fast IP level packet filtering at the server's firewall and requires no changes to the application(s) hosted by the Web server. In this paper we present a detailed design of our technique along with a detailed security analysis. We also describe a concrete implementation of our proposal on the Linux kernel and present an evaluation using two applications: bandwidth intensive Apache HTTPD and database intensive TPCW. Our experiments show that our approach incurs a low performance overhead and is resilient to DoS attacks

This paper describes a motion-adaptive indexing scheme for efficient evaluation of moving continual queries (MCQs) over moving objects. It uses the concept of motion-sensitive bounding boxes (MSBs) to model moving objects and moving queries. These bounding boxes automatically adapt their sizes to the dynamic motion behaviors of individual objects. Instead of indexing frequently changing object positions, we index less frequently changing object and query MSBs, where updates to the bounding boxes are needed only when objects and queries move across the boundaries of their boxes. This helps decrease the number of updates to the indexes. More importantly, we use predictive query results to optimistically precalculate query results, decreasing the number of searches on the indexes. Motion-sensitive bounding boxes are used to incrementally update the predictive query results. Furthermore, we introduce the concepts of guaranteed safe radius and optimistic safe radius to extend our motion-adaptive indexing scheme to evaluating moving continual k-nearest neighbor (kNN) queries. Our experiments show that the proposed motion-adaptive indexing scheme is efficient for the evaluation of both moving continual range queries and moving continual kNN queries.

With the growing popularity and availability of mobile communications, our ability to stay connected while on the move is becoming a reality instead of science fiction as it was just a decade ago. An important research challenge for modern location-based services is the scalable processing of location monitoring requests on a large collection of mobile objects. The centralized architecture, though studied extensively in literature, would create intolerable performance problems as the number of mobile objects grows significantly. This paper presents a distributed architecture and a suite of optimization techniques for scalable processing of continuously moving location queries. Moving location queries can be viewed as standing location tracking requests that continuously monitor the locations of mobile objects of interest and return a subset of mobile objects when certain conditions are met. We describe the design of MobiEyes, a distributed real time location monitoring system in a mobile environment. The main idea behind the MobiEyes' distributed architecture is to promote a careful partition of a real time location monitoring task into an optimal coordination of server-side processing and client-side processing. Such a partition allows evaluating moving location queries with a high degree of precision using a small number of location updates, thus providing highly scalable location monitoring services. A set of optimization techniques are used to limit the amount of computation to be handled by the mobile objects and enhance the overall performance and system utilization of MobiEyes. Important metrics to validate the proposed architecture and optimizations include messaging cost, server load, and amount of computation at individual mobile objects. We evaluate the scalability of the MobiEyes location monitoring approach using a simulation model based on a mobile setup. Our experimental results show that MobiEyes can lead to significant savings in terms of server load and messaging cost when compared to solutions relying on central processing of location information.

With continued advances in communication network technology and sensing technology, there is astounding growth in the amount of data produced and made available through cyberspace. Efficient and high-quality clustering of large datasets continues to be one of the most important problems in large-scale data analysis. A commonly used methodology for cluster analysis on large datasets is the three-phase framework of sampling/summarization, iterative cluster analysis, and disk-labeling. There are three known problems with this framework which demand effective solutions. The first problem is how to effectively define and validate irregularly shaped clusters, especially in large datasets. Automated algorithms and statistical methods are typically not effective in handling these particular clusters. The second problem is how to effectively label the entire data on disk (disk-labeling) without introducing additional errors, including the solutions for dealing with outliers, irregular clusters, and cluster boundary extension. The third obstacle is the lack of research about issues related to effectively integrating the three phases. In this article, we describe iVIBRATE---an interactive visualization-based three-phase framework for clustering large datasets. The two main components of iVIBRATE are its VISTA visual cluster-rendering subsystem which invites human interplay into the large-scale iterative clustering process through interactive visualization, and its adaptive ClusterMap labeling subsystem which offers visualization-guided disk-labeling solutions that are effective in dealing with outliers, irregular clusters, and cluster boundary extension. Another important contribution of iVIBRATE development is the identification of the special issues presented in integrating the two components and the sampling approach into a coherent framework, as well as the solutions for improving the reliability of the framework and for minimizing the amount of errors generated within the cluster analysis process. We study the effectiveness of the iVIBRATE framework through a walkthrough example dataset of a million records and we experimentally evaluate the iVIBRATE approach using both real-life and synthetic datasets. Our results show that iVIBRATE can efficiently involve the user in the clustering process and generate high-quality clustering results for large datasets.

Peer-to-peer (P2P) file sharing systems such as Gnutella have been widely acknowledged as the fastest-growing Internet applications ever. The P2P model has many potential advantages, including high flexibility and serverless management. However, these systems suffer from the well-known performance mismatch between the randomly constructed overlay network topology and the underlying IP-layer topology. This paper proposes to structure the P2P overlay topology using a heterogeneity-aware multitier topology to better balance the load at peers with heterogeneous capacities and to prevent low-capability nodes from throttling the performance of the system. An analytical model is developed to enable the construction and maintenance of heterogeneity-aware overlay topologies with good node connectivity and better load balance. We also develop an efficient routing scheme, called probabilistic selective routing, that further utilizes heterogeneity-awareness to enhance the routing performance. We evaluate our design through simulations. The results show that our multitier topologies alone can provide eight to 10 times improvement in the messaging cost, two to three orders of magnitude improvement in terms of load balancing, and seven to eight times lower topology construction and maintenance costs when compared to Gnutella's random power-law topology. Moreover, our heterogeneity-aware routing scheme provides further improvements on all evaluation metrics, when used with our heterogeneity-aware overlay topologies.

A publish-subscribe overlay service is a wide-area communication infrastructure that enables information dissemination across geographically scattered and potentially unlimited number of publishers and subscribers. A wide-area publish-subscribe (pub-sub) system is often implemented as a collection of spatially disparate nodes communicating on top of a peer to peer overlay network. Such a model presents many inherent benefits such as scalability and performance, as well as potential challenges such as: (i) confidentiality & integrity, (ii) authentication, and (iii) denial-of-service (DoS) attacks. In this paper we present EventGuard for securing pub-sub overlay services. EventGuard comprises of a suite of security guards that can be seamlessly plugged-into a content-based pub-sub system. EventGuard mechanisms aim at providing security guarantees while maintaining the system's overall simplicity, scalability and performance metrics. We present an implementation which shows that EventGuard is easily stackable on any content-based pub-sub core. Finally, our experimental results show that EventGuard can secure a pub-sub system with minimal performance penalty.

We present an adaptive load shedding approach for windowed stream joins. In contrast to the conventional approach of dropping tuples from the input streams, we explore the concept of selective processing for load shedding. We allow stream tuples to be stored in the windows and shed excessive CPU load by performing the join operations, not on the entire set of tuples within the windows, but on a dynamically changing subset of tuples that are learned to be highly beneficial. We support such dynamic selective processing through three forms of runtime adaptations: adaptation to input stream rates, adaptation to time correlation between the streams and adaptation to join directions. Indexes are used to further speed up the execution of stream joins. Experiments are conducted to evaluate our adaptive load shedding in terms of output rate. The results show that our selective processing approach to load shedding is very effective and significantly outperforms the approach that drops tuples from the input streams.

Peer-to-peer (P2P) networks offer a message exchanging overlay for distributed applications such as file sharing, application layer multicast, and publisher/subscriber system. The communication efficiency of the underlying overlay network is thus one of the primary factors that determine the performance of those applications. In this paper, we propose a P2P overlay network aiming at offering the low maintenance overhead of unstructured P2P networks and the scalability and communication efficiency of structured P2P networks. We design a distributed algorithm to construct low-diameter overlay networks with power law topologies. Peers consider both network proximity information and capacity of existing peers when choosing their P2P network neighbors. Using an application layer multicast system as our example, we demonstrate that our system can provide generic, scalable, and low diameter overlay networks for distributed applications that demand efficient P2P communication supports

This paper describes a personalized k-anonymity model for protecting location privacy against various privacy threats through location information sharing. Our model has two unique features. First, we provide a unified privacy personalization framework to support location k-anonymity for a wide range of users with context-sensitive personalized privacy requirements. This framework enables each mobile node to specify the minimum level of anonymity it desires as well as the maximum temporal and spatial resolutions it is willing to tolerate when requesting for k-anonymity preserving location-based services (LBSs). Second, we devise an efficient message perturbation engine which runs by the location protection broker on a trusted server and performs location anonymization on mobile users’ LBS request messages, such as identity removal and spatio-temporal cloaking of location information. We develop a suite of scalable and yet efficient spatio-temporal cloaking algorithms, called CliqueCloak algorithms, to provide high quality personalized location k-anonymity, aiming at avoiding or reducing known location privacy threats before forwarding requests to LBS provider(s). The effectiveness of our CliqueCloak algorithms is studied under various conditions using realistic location data synthetically generated using real road maps and traffic volume data.

Caching on the edge of the Internet is becoming a popular technique to improve the scalability and efficiency of delivering dynamic web content. In this paper we study the challenges in designing a large scale cooperative edge cache network, focusing on mechanisms and methodologies for efficient cooperation among caches to improve the overall performance of the edge cache network. This paper makes three original contributions. First, we introduce the concept of cache clouds, which forms the fundamental framework for cooperation among caches in the edge network. Second, we present dynamic hashing-based protocols for document lookups and updates within each cache cloud, which are not only efficient, but also effective in dynamically balancing lookup and update loads among the caches in the cloud. Third, we outline a utility-based mechanism for placing dynamic documents within a cache cloud. Our experiments indicate that these techniques can significantly improve the performance of the edge cache networks.

Advances in distributed service-oriented computing and global communications have formed a strong technology push for large scale data integration among organizations and enterprises. However, concerns about data privacy become increasingly important for large scale mission-critical data integration applications. Ideally, given a database query spanning multiple private databases, we wish to compute the answer to the query without revealing any additional information of each individual database apart from the query result. In practice, we may relax this constraint to allow efficient information integration while minimizing the information disclosure. In this paper, we propose an efficient decentralized peer-to-peer protocol for supporting aggregate queries over multiple private databases while respecting the privacy constraints of participants. The paper has three main contributions. First, it formalizes the notion of loss of privacy in terms of information revealed at individual participating databases. Second, it presents a novel probabilistic decentralized protocol for top k selection across multiple private databases that minimizes the loss of privacy. Third, it experimentally evaluates the protocol in terms of its correctness, efficiency and privacy characteristics.

Data perturbation techniques are one of the most popular models for privacy preserving data mining (Agrawal and Srikant, 2000; Aggarwal and Yu, 2004). It is especially convenient for applications where the data owners need to export/publish the privacy-sensitive data. A data perturbation procedure can be simply described as follows. Before the data owner publishes the data, they randomly change the data in certain way to disguise the sensitive information while preserving the particular data property that is critical for building the data models. Several perturbation techniques have been proposed recently, among which the most typical ones are randomization approach (Agrawal and Srikant, 2000) and condensation approach (Aggarwal and Yu, 2004).

In a distributed Web service integration environment, the selection of Web services should be based on their reputation and quality-of-service (QoS). Various trust models for web services have been proposed to evaluate the reputation of Web services/service providers. Current mechanisms are based on tracing the feedbacks to the past behaviors of Web services. However, very few of them consider the robustness and attack-resiliency of the trust models. In this paper, we present an attack resilient distributed trust management system in a Web service management environment. The proposed attack resilient trust model uses two vectors to capture the behavior and the trustworthiness of a Web service/service provider based on our analysis on the possible attacks against the trust models. We also present a set of experiments that show the effectiveness of our trust model in detecting malicious behavior of service providers.

This paper presents DynaBot, a domain-specific Web service discovery system. The core idea of the DynaBot service discovery system is to use domain-specific service class descriptions powered by an intelligent Deep Web crawler. In contrast to current registry-based service discovery systems -like the several available UDDI registries - DynaBot promotes focused crawling of the Deep Web of services and discovers candidate services that are relevant to the domain of interest. It uses intelligent filtering algorithms to match services found by focused crawling with the domain-specific service class descriptions. We demonstrate the capability of DynaBot through the BLAST scenario and describe our initial experience with DynaBot.

We present an architecture of a trust framework that can be used to intelligently tradeoff between security and performance in a SAN file system. The primary idea is to differentiate between various clients in the system based on their trustworthiness and provide them with differing levels of security and performance. Client trustworthiness reflects its expected behavior and is evaluated in an online fashion using a customizable trust model. We also describe the interface of the trust framework with an example block level security solution for an out-of-band virtualization based SAN file system (SAN FS). The proposed framework can be easily extended to provide differential treatment based on data sensitivity, using a configurable parameter of the trust model. This allows associating stringent security requirements for more sensitive data, while trading off security for better performance for less critical data, a situation regularly desired in an enterprise.

We present HSAN & a hybrid storage area network, which uses both in-band (like NFS (R. Sandberg et al., 1985)) and out-of-band visualization (like SAN FS (J. Menon et al., 2003)) access models. HSAN uses hybrid servers that can serve as both metadata and NAS servers to intelligently decide the access model per each request, based on the characteristics of requested data. This is in contrast to existing efforts that merely provide concurrent support for both models and do not exploit model appropriateness for requested data. The HSAN hybrid model is implemented using low overhead cache-admission and cache-replacement schemes and aims to improve overall response times for a wide variety of workloads. Preliminary analysis of the hybrid model indicates performance improvements over both models.

Reputation systems have been popular in estimating the trustworthiness and predicting the future behavior of nodes in a large-scale distributed system where nodes may transact with one another without prior knowledge or experience. One of the fundamental challenges in distributed reputation management is to understand vulnerabilities and develop mechanisms that can minimize the potential damages to a system by malicious nodes. In this paper, we identify three vulnerabilities that are detrimental to decentralized reputation management and propose TrustGuard - a safeguard framework for providing a highly dependable and yet efficient reputation system. First, we provide a dependable trust model and a set of formal methods to handle strategic malicious nodes that continuously change their behavior to gain unfair advantages in the system. Second, a transaction based reputation system must cope with the vulnerability that malicious nodes may misuse the system by flooding feedbacks with fake transactions. Third, but not least, we identify the importance of filtering out dishonest feedbacks when computing reputation-based trust of a node, including the feedbacks filed by malicious nodes through collusion. Our experiments show that, comparing with existing reputation systems, our framework is highly dependable and effective in countering malicious nodes regarding strategic oscillating behavior, flooding malevolent feedbacks with fake transactions, and dishonest feedbacks.