Deep hash networks are widely used in tasks such as large-scale image retrieval due to high search efficiency and low storage costs through binary hash codes. With the growing demand for deploying deep hash networks on resource-constrained devices, it is crucial to perform network compression on them, in which automatic pruning constitutes a priority option owing to efficacy maintenance. However, existing pruning methods are mostly designed for image classification, while hashing networks must generate compact binary codes, making each channel more sensitive to retrieval objectives. As a result, their performance often degrades when applied to image retrieval tasks. In this paper, we propose a novel Automatic Channel Pruning framework by Searching with Structure Embedding (ACP-SSE). To the best of our knowledge, this is the first study to explore pruning techniques for deep hash networks and the first automatic pruning method by searching based on network topology structure. Specifically, we first design a structure encoding model by Graph Convolutional Networks (GCNs) whose graph is constructed by hash network and nodes' features are initialized by pruning strategies. The model is trained by contrastive learning loss efficiently without accuracy supervision by fine-tuning pruned models. In addition, we introduce a dynamic pruning search space in consideration of the resource constraints. By converting the automatic channel pruning task into searching the pruned structure with effect similar to the unpruned structure, it enables the method to adapt to various network architectures. Finally, the optimal networks are selected from the candidate set according to their performance in specific downstream tasks. Extensive experiments demonstrate that ACP-SSE indeed works in the automatic channel pruning area, outperforming state-of-the-art baselines in hashing-based image retrieval, while maintaining competitive accuracy in image classification.

Next Point-of-Interest (POI) recommendation task focuses on mining user behavioral preference patterns from historical check-ins to provide personalized suggestions for the next destination. Existing methods primarily rely on shallow contextual information and handcrafted feature interactions to predict the next POI. However, the inherent sparsity and complexity of user mobility patterns limit the computational capacity of non-reasoning models to capture deep intent, while large language models (LLMs) perform suboptimally because they lack a deep understanding of semantic IDs (SIDs) when SIDs are trained separately. To address these limitations, we propose Think2Go, a novel generative next POI recommendation framework, which enhances the model's comprehension of SID representations and explores diverse spatial-temporal patterns via test-time computational scaling. We unify supervised fine-tuning (SFT) and reinforcement learning (RL)-based reasoning within a single architecture, enabling joint optimization of memorization and adaptive reasoning to better retain user behavior patterns while exploring diverse user preferences. To further calibrate policy optimization in adaptive reasoning, we propose two advantage weighting mechanisms that integrate (1) prompt epistemic uncertainty, estimated via kernel density methods to assess the spatial-temporal periodic pattern alignment between queries and user history, promoting increased exploration under high epistemic uncertainty; and (2) reward-informed advantage scaling, captured by normalizing rewards against their maxima to adapt update magnitudes, thereby improving training stability and mitigating overfitting to noisy signals. This joint calibration forms an implicit curriculum learning strategy, delivering fine-grained, instance-aware policy updates that prevent entropy collapse and support robust exploration. Extensive experiments conducted on three real-world datasets demonstrate that Think2Go exhibits strong generalization capabilities and enhances the LLM's understanding of SIDs.

The escalating complexity of urban traffic congestion makes single-scale analysis inadequate for supporting effective top-down management strategies. However, existing researches face challenges in representing continuous, interpretable macro-scale patterns without excessively compromising the fidelity of micro-scale movement details. To address this, we proposed a bottom-up vector field-based framework powered by a data-driven flow model to extract coherent macro-scale congestion patterns from micro-scale mobility. We first constructed a network-constrained demand vector field using a path flow model to capture micro-scale mobility. Building on this, we developed congestion exposure fields to characterize the direction and intensity of individual congestion exposure. To inform meaningful spatiotemporal aggregation and improve interpretability, we incorporated irrotationality theory as a scale-aware criterion, identifying scales where micro-scale mobility exhibited stable and coherent structures suitable for macro-scale representation. At these scales, a congestion potential field was derived to reveal system-wide risk dynamics. A case study in Wuhan demonstrates that the proposed framework effectively reduces micro-scale noise and highlights coherent macro-scale patterns, which offer actionable insights into local traffic complexity by serving as real-time warnings for areas prone to congestion or exhibiting increased vulnerability. This work provides an interpretable approach for selecting meaningful aggregation scales, supporting integrative assessments of micro-macro traffic dynamics.

Federated Learning (FL) is a popular distributed machine learning method that enables the development of a robust global model through decentralized computation and periodic model aggregation, without requiring direct access to clients' data. However, data heterogeneity poses a significant challenge in FL, and the global long-tail distribution exacerbates this issue. While substantial research has focused on mitigating performance degradation caused by long-tailed distributions, existing methods typically concentrate on addressing discrepancies between local and global class distributions, often overlooking the fact that these discrepancies stem from variations in the data itself. To address this, we propose a novel approach, Federated Context Optimization and Feature Information Decoupling (FedDR), which generates partition strategies for each sample to extract and leverage long-tail, global, personalized, and label-text information within its features to enhance the representational distinction of tail classes. Specifically, we first design a Feature Information Decoupling module that separates global, personalized, and long-tail information within the features and incorporates this information into the loss function to strengthen the global model's focus on personalized information in tail samples. Furthermore, to exploit the textual label information embedded in the samples, we integrate a cross-modal model, CoOp, which utilizes open-vocabulary prior knowledge, and implement dynamic knowledge distillation between the client model and CoOp to enhance the client model's feature representation capability. Extensive experimental results on multiple benchmarks demonstrate that the proposed FedDR outperforms state-of-the-art methods in the federated long-tailed learning setting.

Distributed stream processing has recently gained much interest due to the need of extracting meaningful results from continuous data stream. To keep the extracted results fresh, the underlying network flows are often required to transmit packets continuously. Otherwise, these results will become stale, and their staleness is determined by the slowest flow. At this point,coflowscan be semantically comprised. Hence, efficient coflow transmission is critical for streaming applications. However, prior coflow-based solutions have significant limitations. They use a one-shot performance metric—CCT (coflow completion time), which cannot continuously reflect the staleness of the output results for a streaming application. To this end, we propose a new performance metric—coflow age(CA), for coflows generated by distributed streaming applications. The CA tracks thelongest time-since-last-serviceamong all flows in a coflow. In such a context, we consider a data center network with multiple coflows that continuously transmit packets between their source-destination pairs and address the problem of minimizing the average long-term CA while simultaneously satisfying the throughput constraints from the coflows. To solve this problem efficiently, we design a randomized algorithm and a drift-plus-age algorithm, and show that they can make the average long-term CA to achieve nearly two times and arbitrarily close to the optimal value, respectively. Through extensive simulations, we further demonstrate that both of the proposed algorithms can significantly reduce the CA of coflows, without violating the throughput requirement of any coflow, when compared to the state-of-the-art solution in both scenario with the packet arrival probability being known and unknown a prior.

Next Point-of-Interest (POI) prediction is important to various human mobility applications, such as route planning and location-based advertising. To address the spatial-temporal sparsity issues arising from users' irregular and inconsistent visit times to different POIs, multi-granular structures can be incorporated to enhance feature representation through hierarchical relationships. However, existing methods often fall short in capturing the comprehensive multi-granularity spatial-temporal correlations due to three primary limitations: (1) users' complex mobility patterns entangled in single trajectory data, (2) limited mobility patterns details due to independent modeling at each granularity, and (3) low inference efficiency in cascaded multi-granularity predictions. To tackle these challenges, we propose a novel approach that models transformations across different granularities in both spatial regions and temporal periods as a diffusion process, leading to the development of the Multi-Granularity Spatial-Temporal Diffusion Network (MGSTDN). In particular, this model adopts a multi-task architecture, where predictions at varying spatial-temporal granularities (i.e., different diffusion steps) are treated as distinct tasks. By employing a multi-granularity diffusion mechanism in both spatial and temporal dimensions, it captures more nuanced spatial-temporal correlations, enhancing the physical constraints and behavioral pattern dependencies across granularities. During the diffusion process's forward stage, coarser-grained regions and periods are derived based on fine-grained features. In the reverse stage, finer-grained regions and periods are recovered from coarse-grained features, guided by encoded historical trajectory information, until the next POI is determined. To improve computational efficiency, we introduce a multi-granularity mapping propagation matrix, enabling parallel computation and accelerating the prediction process across different granularities. We evaluated the effectiveness of MGSTDN through extensive experiments on three datasets, demonstrating significant improvements over existing methods.

Amid the accelerating pace of global digital transformation, the Industrial Internet of Things (IIoT) has progressively emerged as a vital force in promoting industrial upgrading and economic restructuring. The proliferation of IIoT devices has augmented the complexity of security management, making the deployment of intrusion traffic detection solutions imperative. Existing solutions for network traffic classification have certain limitations. This paper presents GRID, a Graph-based Robust Intrusion Detection solution for IIoT, encompassing two main modules: the Hierarchical Traffic Graph Constructor (HTGC) and the Cascaded Graph Attention Network (CGATN). The HTGC exploits the inherent packet-flow-conversation hierarchy of traffic data to construct the graph structure and fuse packet-level and behavioral features. The CGATN addresses the issues faced by conventional multi-layer Graph Neural Networks (GNNs) and employs contrastive representation learning during training to enhance the robustness of the solution. GRID demonstrates significant advantages compared to state-of-the-art solutions. The experimental results in both closed-world and open-world scenarios reveal an average increase of 3.09% in classification accuracy, 0.23% in balanced accuracy, and 10.03% in Matthews correlation coefficient.

Next point-of-interest (POI) recommendation is to explore the historical check-in sequence information in location-based social networks (LBSNs) to recommend the next location that he/she might be interested in. However, most previous methods used only limited information of unimodal data (i.e., check-in sequences), while some recent methods have attempted to explore multimodal data (e.g., textual content) but lacked sufficient interactions between geographic behavior patterns and content behavior patterns. In this work, we argue that users usually consider geographical trajectories and textual content interdependently to determine the next location to visit. To this end, we propose a novel cross-modal aggregation attention network (CMAAN), which interactively learns multiview representations from POI sequence and content sequence for predicting the next POI. Our approach models inter-modal interaction correlations, intra-modal sequence correlations, and intra-modal semantic correlations simultaneously to fully discover contextual potential relations along the trajectories. Specifically, the intra-modal semantic correlations are able to capture the variable location functionalities under different contextual relationships of cross-modal interaction information. Moreover, we apply the aggregation attention to adaptively aggregate multiview representations which represent the comprehensive hidden state of the next POI. Extensive experiments on two large-scale datasets clearly demonstrate that our CMAAN achieves state-of-the-art performance.

High-definition (HD) mapping is essential for autonomous driving and localization services, providing detailed lane-level road graphs for various applications. Current methodologies primarily segment the geometric structure of lane lines from remote sensing images and extract vectorized road graphs using heuristic methods. However, these approaches fail to adequately account for lane instance information and topological structures. Furthermore, the semi-automated process imposes constraints on the spatial scalability of HD maps. To overcome these limitations, we propose AutoLRG, a two-stage method for lane-level road graph construction. In lane geometry prediction, we propose a lane segmentation network based on directional supervision and multimodal fusion, incorporating an angle-direction loss and a cross-attention-based fusion module to enhance lane perception and connectivity. In lane instance modeling, we develop a Transformer-based lane decoder, which leverages an object detection architecture to extract vectorized lane instances and road vertices in an end-to-end manner. In lane topology construction, we introduce a "road segment–intersection" decoupled model, which establishes the connectivity relationships of intersection nodes based on traffic regulations to form a lane-level topological directed road graph. The ablation studies conducted on the two benchmark datasets (UrbanLaneGraph and OpenSatMap) have validated the effectiveness of the method. Comparative experiments with other methods demonstrate that our approach exhibits superior performance in lane segmentation, instance modeling, and topology construction. Code is available at https://github.com/EchoQiHeng/AutoLRG.

Next Point-of-Interest (POI) recommendation has been proven effective at utilizing sparse, intricate spatial-temporal trajectory data to recommend subsequent POIs to users. While existing methods commonly alleviate the problem of data sparsity by integrating spatial-temporal context information, POI category features, and social relationships, they largely overlook the fact that the trajectory sequences collected in the datasets are often incomplete. This oversight limits the model’s potential to fully leverage historical context. In light of this background, we propose Trajectory Data Augmentation with Uncertainty (TAU) for Next POI Recommendation. TAU is a general graph-based trajectory data augmentation method designed to complete user mobility patterns by marrying uncertainty estimation into the next POI recommendation task. More precisely, TAU taps into the global transition pattern graph to identify sets of intermediate nodes located between every pair of locations, effectively leveraging edge weights as transition probabilities. During trajectory sequence construction, TAU selectively prompts intermediate nodes, chosen based on their likelihood of occurrence as pseudo-labels, to establish comprehensive trajectory sequences. Furthermore, to gauge the certainty and impact of pseudo-labels on the target location, we introduce a novel confidence-aware calibration strategy using evidence deep learning (EDL) for improved performance and reliability. The experimental results clearly indicate that our TAU method achieves consistent performance improvements over existing techniques across two real-world datasets, verifying its effectiveness as the state-of-the-art approach to the task.

Asynchronous federated learning (AFL) allows participants to immediately submit trained models without waiting for other participants, building upon the federated learning (FL). Due to privacy concerns, FL is more susceptible to financial fraud, compounded by the gradient delay issues introduced by asynchronous submissions, making defense against financial fraud more challenging. Motivated by the above finding, we propose a secure and privacy-preserving AFL defense method for image datasets with an implanted backdoor via filtering redundant neurons (BDAFL), enhancing its resilience against such attacks without compromising privacy. We utilize gradient compensation to mitigate the impact of delays introduced by asynchrony. To counter financial fraud, we employ an anomaly detection algorithm based on neurons’ weights and ensemble distillation to eliminate the affected neurons implanted with the backdoor, rendering the attack ineffective. Extensive experiments demonstrate the effectiveness and superiority of our approach.

The rapid proliferation of Internet of Things (IoT) devices has brought about unprecedented convenience to people’s daily lives. However, this growth has also created opportunities for hackers to launch large-scale botnet attacks using these devices. As a result, it is critical to deploy real-time traffic classifiers on edge gateways to detect network intrusions and improve near-source protection capabilities. To this end, we propose iDetector, a novel real-time intrusion detection solution for IoT networks that is simple in structure and easy to reproduce. iDetector samples network conversations in real-time using a sliding sampling window and generates traffic samples that integrate multiple features. This allows the samples to accurately capture the patterns of each type of traffic. We propose the nonlinear feature transformation (NFT) algorithm based on the prior distribution of traffic features to increase the information entropy of the samples and thereby improve the classification performance. To enable deployment on edge gateways, we propose EdgeNet, a lightweight deep neural network model that utilizes depthwise separable convolution and self-attention mechanism to enhance classification performance while reducing the number of model parameters. Experimental evaluations show that our solution outperforms state-of-the-art deep learning-based solutions in terms of classification performance and has faster classification speed on resource-constrained edge gateways.

Data privacy is becoming increasingly important as data becomes more valuable, as evidenced by the enactment of right-to-be-forgotten laws and regulations. However, in a federated learning (FL) system, simply deleting data from the database when a user requests data revocation is not sufficient, as the training data is already implicitly contained in the parameter distribution of the models trained with it. Furthermore, the global model in the FL system is vulnerable to data poisoning attacks by malicious nodes. Exploring a reliable data poisoning reversal method can effectively counter such attacks. In this article, we analyze the necessity of decoupling the processes of unlearning and training and propose a training-agnostic and efficient method that can effectively perform two types of unlearning tasks: 1) client revocation and 2) category removal. Specifically, we decompose the unlearning process into two steps: 1) knowledge erasure and 2) memory guidance. We first propose a novel knowledge erasure strategy called momentum degradation (MoDe) which realizes the erasure of implicit knowledge in the model and ensures that the model can move smoothly to the early state of the retrained model. To mitigate the performance degradation caused by the first step, the memory guidance strategy implements guided fine-tuning of the model on different data points, which can effectively restore the discriminability of the model on the remaining data points. Extensive experiments demonstrate that our method outperforms the existing task-specific algorithms and matches the performance of retraining, accelerating the execution time by 5–20 times compared to retraining on different data sets.

Federated Learning (FL), which has been employed to train machine learning models on the data with a distributed manner, could enhance the immersive user experience for the human-centric metaverse. However, it’s challenging to train machine learning models accurately and promptly with FL for the human-centric metaverse due to massive data communication and user unreliability. User experience could be negatively affected by using low-quality machine learning models for human-centric metaverse, e.g., it cannot scrutinize and arrive at decisions accurately and timely. To resolve this pressing issue, we propose MetaFul a federated unlearning solution which reduces the negative influences of low-quality data with no data transmission by removing low-quality training models at the server side. To be specific, MetaFul includes three main components. (i) Low-throughput federated learning (LT-FL) addresses the issue of large model transmission in FL by decreasing the dimension and the number of transmitted model parameters. (ii) Loss-based model quality assessment (LM-QA) utilizes the model loss generated in LT-FL to estimate user data quality. (iii) Non-communicative federated unlearning (NC-FUL) revokes the low-quality data impact on the FL model with careful designed federated unlearning at the server side. Both LM-QA and NC-FUL have no communications with clients. Finally, extensive evaluations are conducted to show MetaFul could improve the model accuracy by at least 2.5% and decrease the user perception time by at least 19.3% in human-centric metaverse compared to benchmarks.

Unmanned Aerial Vehicle (UAV)-enabled mobile edge computing has been proposed as an efficient task-offloading solution for user equipments (UEs). Nevertheless, the presence of heterogeneous UAVs makes centralized navigation policies impractical. Decentralized navigation policies also face significant challenges in knowledge sharing among heterogeneous UAVs. To address this, we present the soft hierarchical deep reinforcement learning network (SHDRLN) and dual-end federated reinforcement learning (DFRL) as a decentralized navigation policy solution. It enhances overall task-offloading energy efficiency for UAVs while facilitating knowledge sharing. Specifically, SHDRLN, a hierarchical DRL network based on maximum entropy learning, reduces policy differences among UAVs by abstracting atomic actions into generic skills. Simultaneously, it maximizes the average efficiency of all UAVs, optimizing coverage for UEs and minimizing task-offloading waiting time. DFRL, a federated learning (FL) algorithm, aggregates policy knowledge at the cloud server and filters it at the UAV end, enabling adaptive learning of navigation policy knowledge suitable for the UAV's performance parameters. Extensive simulations demonstrate that the proposed solution not only outperforms other baseline algorithms in overall energy efficiency but also achieves more stable navigation policy learning under different levels of heterogeneity of different UAV performance parameters.

Federated Learning (FL) has emerged as a privacy-preserving paradigm enabling collaborative model training among distributed clients. However, current FL methods operate under the closed-world assumption, i.e., all local training data originates from a global labeled dataset balanced across classes, which is often invalid for practical scenarios. In contrast, in many open-world settings, data have been observed to exhibit heavy-tailed distributions, particularly in the realm of mobile computing and Internet of Things (IoT). Heavy-tailed data can have a significant negative impact on the performance of learning algorithms due to amplifying the heterogeneity in the FL environment. To this end, we introduce a novel framework to counter biased training caused by diverse and imbalanced classes. This framework includes a balance-aware reward aggregation mechanism addressing local majority and global minority class disparities. Rewards are assigned based on client class prevalence for fair aggregation. A calibration module supplements global aggregation to manage conflicts from inconsistent data distribution among clients. Using reward aggregation and calibration, we effectively mitigate heavy-tailed distribution effects, enhancing FL model performance. This framework seamlessly integrates with leading FL methods, demonstrated through extensive experiments on benchmark and real-world datasets.

Federated unlearning (FUL) is an emerging distributed machine learning paradigm which enables the removal or unlearning of specific training data effects from trained Federated Learning (FL) models. While current studies mostly focus on client-side FUL to address the “right to be forgotten”, and ignore the server's right to remove local models from the global model, particularly when clients are trained with low-quality data. In this paper, we introduce the Server-Initiated Federated Unlearning (SIFU) algorithm, devised to eliminate low-quality data from the global model. SIFU consists of two main components: (i) Identifying low-quality data: we develop a category-based method for quantifying low-quality data for each client and filter out clients containing such data. Datasets are then divided accordingly. (ii) Unlearning low-quality data: we employ gradient ascent training to counteract the adverse effects of low-quality data on local models. To minimize any bias introduced, we concurrently perform several batches of boosting training with good-quality data. SIFU could identify and promptly eliminate the impact of low-quality data on the FL global model while still preserving the benefits of good-quality data. Finally, extensive evaluations are conducted to verify the performance of SIFU with four different kinds of datasets and models. Results show that, compared to retraining from scratch, SIFU accelerates the speed of unlearning by 15× for small datasets (i.e., MNIST and FMNIST) and 20× for large datasets (i.e., CIFAR-10 and CelebA) without any degradation in accuracy, which also outperforms the state of the arts.

Anomaly detection aims to identify data or behav-iors that are different from the usual patterns. In traditional anomaly detection settings, edge devices collect the data and send it to a centralized server for model training, which faces two critical issues: (1) it risks data exposure during transmission; (2) it demands a large amount of network bandwidth for data transfer. To tackle these problems, we propose a Visual Federated Learning algorithm (VFLA) for anomalous behavior identification in the multi-UAVs system. To the best of our knowledge, we are the first to merge federated learning with video-based anomaly detection. VFLA consists of two phases: The initial phase is training a pseudo-label generator. UAVs collect a dataset and manually annotate it. This labeled data is then used to train the pseudo-label generator on the server, which is subsequently distributed back to the UAVs. The second phase is the federated learning-based anomaly detection model training. UAVs leverage the pseudo-label generator to automatically annotate the collected video footage. These annotated videos are fed into an anomaly detection network for training. Once the local training is completed, UAVs upload their local models to a server for federated aggregation. The global model is then redistributed to the UAVs for additional training rounds, until reach the target accuracy. Finally, we simulate the federated learning anomaly detection algorithm on the Shanghai-tech dataset, it demonstrates an average accuracy boost of 5.6% compared to baselines.

Social Internet of Things (SIoT), a promising and emerging paradigm that injects the notion of social networking into smart objects (i.e., things), paving the way for the next generation of Internet of Things. However, due to the risks and uncertainty, a crucial and urgent problem to be settled is establishing reliable relationships within SIoT, that is, trust evaluation. Graph neural networks for trust evaluation typically adopt a straightforward way such as one-hot or node2vec to comprehend node characteristics, which ignores the valuable semantic knowledge attached to nodes. Moreover, the underlying structure of SIoT is usually complex, including both the heterogeneous graph structure and pairwise trust relationships, which renders hard to preserve the properties of SIoT trust during information propagation. To address these aforementioned problems, we propose a novel knowledge-enhanced graph neural network (KGTrust) for better trust evaluation in SIoT. Specifically, we first extract useful knowledge from users’ comment behaviors and external structured triples related to object descriptions, in order to gain a deeper insight into the semantics of users and objects. Furthermore, we introduce a discriminative convolutional layer that utilizes heterogeneous graph structure, node semantics, and augmented trust relationships to learn node embeddings from the perspective of a user as a trustor or a trustee, effectively capturing multi-aspect properties of SIoT trust during information propagation. Finally, a trust prediction layer is developed to estimate the trust relationships between pairwise nodes. Extensive experiments on three public datasets illustrate the superior performance of KGTrust over state-of-the-art methods.

Urban tracking plays a vital role for people’s urban life in intelligent transportation systems, e.g., public safety, case investigation, finding missing items, etc. However, the current tracking methods consume a large amount of communication and computing resources since they mainly offload all related sensing data, i.e., videos, generated by widely deployed cameras to the cloud where data are stored, processed, and analyzed. In this paper, we propose a graph optimized data offloading algorithm leveraging a crowd-AI hybrid method to minimize the data offloading cost and ensure the reliable urban tracking result. To be specific, we first formulate a crowd-AI hybrid urban tracking scenario, and prove the proposed data offloading problem in this scenario is NP-hard. Then, we solve it by decomposing the problem into two parts, i.e., trajectory prediction and task allocation. The trajectory prediction algorithm, leveraging the state graph, computes possible tracking areas of the target object, and the task allocation algorithm, using the dependency graph, chooses the optimal set of crowds and cameras to cover the tracking area while minimizing the data offloading cost separately. Finally, the extensive simulations with large real world data set are conducted showing that the proposed algorithm outperforms benchmarks in reducing data offloading cost while ensuring the tracking success rate in intelligent transportation systems.

Recently, the Weisfeiler-Lehman (WL) graph isomorphism test was used to measure the expressiveness of graph neural networks (GNNs), showing that the neighborhood aggregation GNNs were at most as powerful as 1-WL test in distinguishing graph structures. There were also improvements proposed in analogy to k-WL test ($k>1$). However, the aggregations in these GNNs are far from injective as required by the WL test, and suffer from weak distinguishing strength, making it become the expression bottleneck. In this paper, we improve the expressiveness by exploring powerful aggregations. We reformulate an aggregation with the corresponding aggregation coefficient matrix, and then systematically analyze the requirements on this matrix for building more powerful and even injective aggregations. We also show the necessity of applying nonlinear units ahead of aggregations, which is different from most existing GNNs. Based on our theoretical analysis, we develop ExpandingConv. Experimental results show that our model significantly boosts performance, especially for large and densely connected graphs.

Major manufacturers and retailers are increasingly using RFID systems in supply-chain scenarios, where theft of goods during transport typically causes significant economic losses for the consumer. This paper studies how to achieve time-efficient and secure integrity authentication problems in RFID systems. We start with a straightforward solution called SecAuth, which uses a secure identity stored on reserved memory to authenticate tags in a secure way. We then propose a time efficient KTAuth protocol, which design a verification chain mechanism to efficiently verify a small set of key tags using limited on-tag memory. We point out that the limitation of KTAuth is that it takes too much overhead to write a large block of data to tag memory, which leads to the proposed group selection mechanism. The KTAuth with group selection (KTAuth-GS) enables you to select key tags with a single select command, which helps to quickly check the existence of key tags and reduces the data writes on the tags. Experiments and simulation results demonstrate that the proposed KTAuth-GS can defend against counterfeiting attacks by providing more reliable results and reducing the execution time by as much as a factor of 5 when compared with a baseline tag identification protocol.

Many improvements on GNNs can be deemed as operations on the spectrum of the underlying graph matrix, which motivates us to directly study the characteristics of the spectrum and their effects on GNN performance. By generalizing most existing GNN architectures, we show that the correlation issue caused by the unsmooth spectrum becomes the obstacle to leveraging more powerful graph filters as well as developing deep architectures, which therefore restricts GNNs’ performance. Inspired by this, we propose the correlation-free architecture which naturally removes the correlation issue among different channels, making it possible to utilize more sophisticated filters within each channel. The final correlation-free architecture with more powerful filters consistently boosts the performance of learning graph representations. Code is available at https://github.com/qslim/gnn-spectrum.

Spatial-temporal forecasting (e.g., traffic flow forecasting) plays a vital role in various applications and has important research significance. Most existing spatial-temporal forecasting methods improve the forecasting accuracy by stacking discrete modules. However, the discontinuous hidden state trajectories between discrete modules may lead to high numerical errors, parameter redundancy, and model complexity. Neural Controlled Differential Equations (NCDEs) can solve the above issues effectively, which have a mechanism that can continuously adjust the hidden state trajectories according to controlled signals. In this paper, we propose a novel Spatial-Temporal Continuous Encoder-Decoder (STCED) method for spatial-temporal forecasting. Specifically, we first propose an overall spatial-temporal continuous encoder-decoder architecture based on NCDEs, which can not only promote the spatial-temporal message passing, but also capture the periodicity without introducing multiple modules or extra parameters. Then, we refactor the core of NCDEs based on Fast Weight Programmers (FWPs) and Graph Neural Networks (GNNs) to overcome the limitations that previous NCDEs cannot capture long-range spatial-temporal dependencies. We conduct experiments on two representative spatial-temporal datasets, demonstrating the effectiveness and superiority of our proposed algorithm.

Load balancing is essential for datacenter networks. However, prior solutions have significant limitations: they either are oblivious to congestion or involve a daunting and time-consuming parameter-tunning task over their heuristics for achieving good performance. Thus, we ask: is it possible to learn to balance datacenter traffic? While deep reinforcement learning (DRL) sounds like a good answer, we observe that it is too heavyweight due to the long decision-making latency. Therefore, we introduce BULB, a lightweight and automated datacenter load balancer. BULB learns link weights to guide the end-hosts to spread traffic, so as to free the central agent from quick flow-level decision-making. BULB offline trains a DRL agent for optimizing link weights but employs an imitation learning based approach to faithfully translate this agent’s DNN to a decision tree for online deployment. We implement a BULB prototype with a popular machine learning framework and evaluate it extensively in ns-3. The results show that BULB achieves up to 36.6%/56.4%, 19.9%/42.5%, 35.9%/54.8%, and 45.1%/67.7% better average/tail flow completion time than ECMP, CONGA, LetFlow, and Hermes, respectively. Moreover, BULB reduces the decision latency by 175 times while incurring only 2% performance loss after converting the DNN into a decision tree.

Federated Learning (FL) is susceptible to gradient leakage attacks, as recent studies show the feasibility of obtaining private training data on clients from publicly shared gradients. Existing work solves this problem by incorporating a series of privacy protection mechanisms, such as homomorphic encryption and local differential privacy to prevent data leakage. However, these solutions either incur significant communication and computation costs, or significant training accuracy loss. In this paper, we show that the sensitivity of gradient changes w.r.t. training data is an essential measure of information leakage risk. Based on this observation, we present a novel defense, whose intuition is perturbing gradients to match information leakage risk such that the defense overhead is lightweight while privacy protection is adequate. Our another key observation is that global correlations of gradients could compensate for this perturbation. Based on such compensation, training can achieve guaranteed accuracy. We conduct experiments on MNIST, Fashion-MNIST and CIFAR-10 for defending against two gradient leakage attacks. Without sacrificing accuracy, the results demonstrate that our lightweight defense can decrease the PSNR and SSIM between the reconstructed images and raw images by up to more than 60% for both two attacks, compared with baseline defensive methods.

We explore the problem of selectively forgetting categories from trained CNN classification models in federated learning (FL). Given that the data used for training cannot be accessed globally in FL, our insights probe deep into the internal influence of each channel. Through the visualization of feature maps activated by different channels, we observe that different channels have a varying contribution to different categories in image classification.

The machine learning performance usually could be improved by training with massive data. However, requesters can only select a subset of devices with limited training data to execute federated learning (FL) tasks as a result of their limited budgets in today’s IoT scenario. To resolve this pressing issue, we devise a blockchain-enhanced FL market (BFL) to$(i)$make data in computationally bounded devices available for training with social Internet of things,$(ii)$maximize the amount of training data with given budgets for an FL task, and$(iii)$decentralize the FL market with blockchain. To achieve these goals, we firstly propose a trust-enhanced collaborative learning strategy (TCL) and a quality-oriented task allocation algorithm (QTA), where TCL enables training data sharing among trusted devices with social Internet of things, and QTA allocates suitable devices to execute FL tasks while maximizing the training quality with fixed budgets. Then, we devise an encrypted model training scheme (EMT) based on a simple but countervailable differential privacy methodology to prevent attacks from malicious devices. In addition, we also propose a contribution-driven delegated proof of stake (DPoS) consensus mechanism to guarantee the fairness of reward distribution in the block generation process. Finally, extensive evaluations are conducted to verify the proposed BFL could improve the total utility of requesters and average accuracy of FL models significantly.

Summary The cyber‐physical system (CPS) has become a promising direction that can enrich the interaction between people and people, people and objects, objects and objects in the physical world and the virtual world. Because of the development of CPS, the number and types of smart device connected to the Internet has rapidly increased, bringing with it the current issues of flexibility, efficiency, availability, security, and scalability of the Internet‐of‐Things (IoT) network. These problems are caused by the key mechanism of large‐scale distribution to the IoT network. In this article, we propose CPSS inspired by software‐defined networking (SDN) and service‐oriented architecture, which is a powerful software‐defined cyber‐physical system to solve the challenges of broader CPS adoption. CPSS has three complementary modules: application‐as‐a‐service (AAAS), network‐as‐a‐service (NAAS), and infrastructure‐as‐a‐service (IAAS). IAAS collects network information from a different set of controllers and devices and generates a domain‐wide network view. NAAS collects information from different domains and generates a global view, which helps data transmission between different domains and enables seamless switching between applications and infrastructure devices. AAAS provides personalized service for each user. To test CPSS, we develop a prototype system. Our experimental results show that CPSS can effectively control cyber‐physical system, and CPSS has greatly improved network latency, network throughput, and network reliability compared to traditional controller‐controlled CPS.

Distributed machine learning (DML) is an increasingly important workload. In a DML job, each communication phase can comprise acoflow, and there are dependencies among its coflows. Thus, efficient coflow scheduling becomes critical for DML jobs. However, the majority of existing solutions focus on scheduling single-stage coflows with no dependencies. While there are a few studies schedule dependent coflows of multi-stage jobs, they suffer from either practical or theoretical issues. Motivated by this situation, we study how to schedule dependent coflows of multiple DML jobs to minimize the total JCT in a shared cluster. We present a formal mathematical formulation for this problem and prove its NP-hardness. To solve this problem without job size information, we present an online coflow-aware optimization framework calledParrot. The core idea inParrotis to infer the job with the shortest remaining processing time (SRPT) each time and dynamically control the inferred job's bandwidth based on how confident it is an SRPT job while being mindful of not starving any other job. Specifically, in the design ofParrot, we present a least per-coflow attained service (LPCAS) policy to infer the SRPT job. We further propose a dynamic job weight assignment mechanism and a linear program (LP) based weighted bandwidth scaling strategy for sharing bandwidth among DML jobs. We have proved thatParrotalgorithm has a non-trivial competitive ratio. The results from large-scale trace-driven simulations further demonstrate that ourParrotcan reduce the total JCT by up to 58.4 percent, compared to the state-of-the-art Aalo solution.

The Inter-datacenter transfer is a fundamental service for global cloud applications. Geo-distributed data centers become an essential resource for their application performance which may be destroyed by network congestion. Recent years, most inter-datacenter transfer methods focus on allocating transfers by bandwidth allocation to achieve low cost or high utilization. However, the congestion condition is rarely considered in these works. In this article, we introduce a congestion-aware traffic allocation method named CONA (CONgestion-Aware), whose target is to maximize the profit of allocation transfer among multiple data centers. On this purpose, a maximizing optimization model is proposed, and an efficient link grading strategy is presented. A matrix transformation method is also introduced to simplify the optimization problem. Furthermore, the link congestion condition is considered by the controller, as well as the prediction on link congestion. To verify our proposed method, simulation model is established and comprehensive experiments are conducted. The experimental results show that our method brings higher profit than fair share and greedy traffic allocation method.

In the era of global-scale services, analytical queries are performed on datasets that span multiple data centers (DCs). Such geo-distributed queries generate a large amount of inter-DC data transfers at run time. Due to the expensive inter-DC bandwidth, various methods have been proposed to reduce the traffic cost in geo-distributed data analytics. However, current methods do not attempt to address the throughput issue in geo-distributed analytics. In this article, we target at characterizing and optimizing a cost-throughput tradeoff problem in geo-distributed data analytics. Our objectives are two-fold: (1) we minimize the inter-DC traffic cost when serving geo-distributed analytics with uncertain query demand, and (2) we maximize the system throughput, in terms of the number of query requests that can be successfully served with guaranteed queuing delay. Specifically, we formulate a stochastic optimization problem that seamlessly combines these two objectives. To solve this problem, we take advantage of Lyapunov optimization techniques to design and analyze a two-timescale online control framework. Without prior knowledge of future query requests, this framework makes online decisions on input data placement and admission control of query requests. Rigorous theoretical analyses show that our framework can achieve a near-optimal solution and maintain system stability and robustness as well. Extensive trace-driven simulation results further demonstrate that our framework is capable of reducing inter-DC traffic cost, improving system throughput, and guaranteeing a maximum delay for each query request.

Approximate Nearest Neighbor(ANN) search is the core problem in many large-scale machine learning and computer vision applications such as multimodal retrieval. Hashing is becoming increasingly popular, since it can provide efficient similarity search and compact data representations suitable for handling such large-scale ANN search problems. Most hashing algorithms concentrate on learning more effective projection functions. However, the accuracy loss in the quantization step has been ignored and barely studied. In this paper, we analyse the importance of various projected dimensions, distribute them into several groups and quantize them with two types of values which can both better preserve the neighborhood structure among data. One is Variable Integer-based Quantization (VIQ) that quantizes each projected dimension with integer values. The other is Variable Codebook-based Quantization (VCQ) that quantizes each projected dimension with corresponding codebook values. We conduct experiments on five common public data sets containing up to one million vectors. The results show that the proposed VCQ and VIQ algorithms can both achieve much higher accuracy than state-of-the-art quantization methods. Furthermore, although VCQ performs better than VIQ, ANN search with VIQ provides much higher search efficiency.

Hashing has been widely applied to the large-scale approximate nearest neighbor search problem owing to its high efficiency and low storage requirement. Most investigations concentrate on learning hashing methods in a centralized setting. However, in existing big data systems, data is often stored across different nodes. In some situations, data is even collected in a distributed manner. A straightforward way to solve this problem is to aggregate all the data into the fusion center to obtain the search result (aggregating method). However, this strategy is not feasible because of the prohibitive communication cost. Although a few distributed hashing methods have been proposed to reduce this cost, they only focus on designing a distributed algorithm for a specific global optimization objective without considering scalability. Moreover, existing distributed hashing methods aim at finding a distributed solution to hashing, meanwhile avoiding accuracy loss, rather than improving accuracy. To address these challenges, we propose a Scalable Distributed Hashing (SDisH) model in which most existing hashing methods can be extended to process distributed data with no changes. Furthermore, to improve accuracy, we utilize the search radius as a global variable across different nodes to achieve a global optimum search result for every iteration. In addition, a voting algorithm is presented based on the results produced by multiple iterations to further reduce search errors. Theoretical analyses of communication, computation, and accuracy demonstrate the superiority of the proposed model. Numerical simulations on three large-scale and two relatively small benchmark datasets also show that the SDisH model achieves up to 44.75% and 10.23% accuracy gains compared to the aggregating method and state-of-the-art distributed hashing methods, respectively.

Next location prediction aims to find the location that the user will visit next. It plays a fundamental role for location-based applications. However, the heterogeneity and sparsity of the trajectory data pose great challenges to the task. Recently, RNN-based methods have shown promising performance in learining the spatio-temporal characteristics of the trajectory. While the effectiveness of location prediction has been improved, the computational efficiency and the long-term preferences still leave space for further research. The self-attention mechanism is viewed as a promising solution for parallel computation and exploiting sequential regularities from sparse data. But the huge memory cost and the neglect of temporal information make it infeasible to directly modeling human mobility regularities. In this paper, we propose a temporal-context-based self-attention network named TCSA-Net, which can simultaneously exploit long- and short-term mvoement preferences from sparse and long trajectories. In particular, we design a novel two-stage self-attention architecture that can learn long-term dependency under constrained memory budget. Further, we propose a multi-modal embedding layer to model two complementary temporal contexts and provide more abundant temporal and sequential information. Extensive experiments on two real-life datasets show that the TCSA-Net significantly outperforms the state-of-the-art methods in terms of standard evaluation metrics.

Traffic prediction plays an essential role in intelligent transportation system. Accurate traffic prediction can assist route planing, guide vehicle dispatching, and mitigate traffic congestion. This problem is challenging due to the complicated and dynamic spatio-temporal dependencies between different regions in the road network. Recently, a significant amount of research efforts have been devoted to this area, especially deep learning method, greatly advancing traffic prediction abilities. The purpose of this paper is to provide a comprehensive survey on deep learning-based approaches in traffic prediction from multiple perspectives. Specifically, we first summarize the existing traffic prediction methods, and give a taxonomy. Second, we list the state-of-the-art approaches in different traffic prediction applications. Third, we comprehensively collect and organize widely used public datasets in the existing literature to facilitate other researchers. Furthermore, we give an evaluation and analysis by conducting extensive experiments to compare the performance of different methods on a real-world public dataset. Finally, we discuss open challenges in this field.

Tag identification is a critical operation in large-scale RFID applications. Typically, in the RFID-enabled warehouse, the reader needs to execute tag identifications to obtain the inventory information of numerous tagged items. The existing schemes usually divide the time frame into multiple slots and map each tag to one of them for replying its identity. This imposes serious tag collisions because two or more tags may be mapped to the same slot and their responses corrupt with each other. When tag collision happens, all the collided tags cannot be identified by the reader, which significantly increases the identification delay. To overcome the collision problem in the identification process, this paper proposes a Group Tag Identification (GTI) framework to identify grouped tags in both singleton and collision slots. The key novelty of GTI is in leveraging tag-correlation to identify grouped tags in the collision slots without any extra transmission overhead. The main challenge of this work is to overcome the communication and architectural limitations of RFID systems in the context of building ID and slot-correlation between tags. Extensive simulations show that GTI significantly reduces the identification delay by up to 40 percent when compared with the state-of-the-art dynamical frame slotted aloha schemes.

In this paper, we propose a novel super-resolution and infection edge detection co-guided learning network for COVID-19 CT segmentation (CogSeg). Our CogSeg is a coherent framework consisting of two branches. Specifically, we use image super-resolution (SR) as an auxiliary task, which assist segmentation to recover high-resolution representations. Moreover, we propose an infection edge detection guided region mutual information (RMI) loss, which uses the edge detection results of segmentation to explicitly maintain the high order consistency between segmentation prediction and ground truth around infection edge pixels. Our CogSeg network can effectively maintain high-resolution representation and leverages edge details to improve the segmentation performance. When evaluated on two publicly available COVID-19 CT datasets, our CogSeg improves 10.63 and 13.02 points than the established baseline method (i.e. U-Net) w.t.r mIoU. Moreover, our CogSeg achieves more appealing results both quantitatively and qualitatively than the state-of-the-art methods.

Accurate traffic prediction is of great importance in Intelligent Transportation System. This problem is very challenging due to the complex spatial and long-range temporal dependencies. Existing models generally suffer two limitations: (1) GCN-based methods usually use a fixed Laplacian matrix to model spatial dependencies, without considering their dynamics; (2) RNN and its variants are only capable of modeling a limited-range temporal dependencies, resulting in significant information loss. In this paper, we propose a novel spatial-temporal graph neural network (STNN), an end-to-end solution for traffic prediction that simultaneously captures dynamic spatial and long-range temporal dependencies. Specifically, STNN first uses a spatial attention network to model complex and dynamic spatial correlations, without any expensive matrix operations or relying on predefined road network topologies. Second, a temporal transformer network is utilized to model long-range temporal dependencies across multiple time steps, which considers not only the recent segment, but also the periodic dependencies of historical data. Making full use of historical data can alleviate the difficulty of obtaining real-time data and improve the prediction accuracy. Experiments are conducted on two real-world traffic datasets, and the results verify the effectiveness of the proposed model, especially in long-term traffic prediction.

Major manufacturers and retailers are increasingly using RFID systems in supply-chain scenarios, where theft of goods during transport typically causes significant economic losses for the consumer. Recent sample-based authentication methods attempt to use a small set of random sample tags to authenticate the integrity of the entire tag population, which significantly reduces the authentication time at the expense of slightly reduced reliability. The problem is that it still incurs extensive initialization overhead when writing the authentication information to all of the tags. This paper presents KTAuth, a lightweight integrity authentication approach to efficiently and reliably detect missing tags and counterfeit tags caused by stolen attacks. The competitive advantage of KTAuth is that it only requires writing the authentication information to a small set of deterministic key tags, offering a significant reduction in initialization costs. In addition, KTAuth strictly follows the C1G2 specifications and thus can be deployed on Commercial-Off-The-Shelf RFID systems. Furthermore, KTAuth proposes a novel authentication chain mechanism to verify the integrity of tags exclusively based on data stored on them. To evaluate the feasibility and deployability of KTAuth, we implemented a small-scale prototype system using mainstream RFID devices. Using the parameters achieved from the real experiments, we also conducted extensive simulations to evaluate the performance of KTAuth in large-scale RFID systems.

Over the last decade, traffic engineering (TE) has always been a research hotspot in data center networks. For routing flows efficiently and practically, existing TE schemes explore experience-driven heuristics or machine learning (ML) techniques to predict/identify network flows’ size information. However, these TE schemes have significant limitations: they either identify the flow size information too late or are unaware of the ML models’ prediction errors. In this paper, we present DarkTE, a novel TE solution that can learn to predict flow size information timely for achieving better routing performance while being robust to the prediction errors. At its heart, DarkTE employs an ensemble learning technique (i.e., random forest) to classify flows into mice and elephant flows with high accuracy. It then leverages a confidence-based rate allocation and path selection scheme to mitigate the occasional classification errors. Large-scale simulations demonstrate that DarkTE classifies flows within hundreds of microseconds, and the classification accuracy is at least 86.4% over three different realistic workloads. Further, DarkTE completes flows 2.94 times faster on average and makes more links to experience over 90% bandwidth utilization than the Hedera solution.

For learning graph representations, not all detailed structures within a graph are relevant to the given graph tasks. Task-relevant structures can be localized or sparse which are only involved in subgraphs or characterized by the interactions of subgraphs (a hierarchical perspective). A graph neural network should be able to efficiently extract task-relevant structures and be invariant to irrelevant parts, which is challenging for general message passing GNNs. In this work, we propose to learn graph representations from a sequence of subgraphs of the original graph to better capture task-relevant substructures or hierarchical structures and skip noisy parts. To this end, we design soft-mask GNN layer to extract desired subgraphs through the mask mechanism. The soft-mask is defined in a continuous space to maintain the differentiability and characterize the weights of different parts. Compared with existing subgraph or hierarchical representation learning methods and graph pooling operations, the soft-mask GNN layer is not limited by the fixed sample or drop ratio, and therefore is more flexible to extract subgraphs with arbitrary sizes. Extensive experiments on public graph benchmarks show that soft-mask mechanism brings performance improvements. And it also provides interpretability where visualizing the values of masks in each layer allows us to have an insight into the structures learned by the model.

Nowadays, more and more sensors, devices and applications are connected in Industrial Internet of Things (IIoT), producing massive real-time flows which need to be scheduled for Quality-of-Service provision. To realize application-aware and adaptive flow scheduling, the problem of traffic classification must be addressed at first. When edge computing paradigm is introduced into IIoT, the traffic classification service can be deployed on edge node in the near-end. Recently, deep-learning-based IIoT traffic classification methods show better performance, but the computational cost of deep learning model is too high to be deployed on edge node. Moreover, increasingly unknown flows generated by new devices and emerging industrial APPs lead to frequent training of traffic classifiers. It is difficult to migrate the complex process of classifier training from cloud server to edge nodes with limited resources. To address these issues, we take the benefits of hash mechanism and consensus mechanism in blockchain to design a lightweight IIoT traffic classification service, which is more applicable for edge computing paradigm. First, inspired by the hash mechanism in blockchain and the learning to hash for big data, we propose a new learning-to-hash method named extension hashing. By this method, we can build the set of binary coding tress (BCT set), then generating hash table for more efficient k-nearest neighbor-based classification without complex classifier training. Then, we design a new voting-based consensus algorithm to synchronize the BCT sets and the hash tables across edge nodes, thereby providing the traffic classification service. Finally, we conduct data-driven simulations to evaluate the proposed service. By comparing traffic classification results on public data set, we can see that the proposed service achieves the highest classification accuracy with the minimal time cost and memory usage.

Hashing methods have been widely used in Approximate Nearest Neighbor (ANN) search for big data due to low storage requirements and high search efficiency. These methods usually map the ANN search for big data into the k -Nearest Neighbor ( k NN) search problem in Hamming space. However, Hamming distance calculation ignores the bit-level distinction, leading to confusing ranking. In order to further increase search accuracy, various bit-level weights have been proposed to rank hash codes in weighted Hamming space. Nevertheless, existing ranking methods in weighted Hamming space are almost based on exhaustive linear scan, which is time consuming and not suitable for large datasets. Although Multi-Index hashing that is a sub-linear search method has been proposed, it relies on Hamming distance rather than weighted Hamming distance. To address this issue, we propose an exact k NN search approach with Multiple Tables in Weighted Hamming space named WHMT, in which the distribution of bit-level weights is incorporated into the multi-index building. By WHMT, we can get the optimal candidate set for exact k NN search in weighted Hamming space without exhaustive linear scan. Experimental results show that WHMT can achieve dramatic speedup up to 69.8 times over linear scan baseline without losing accuracy in weighted Hamming space.

Accurate traffic forecasting is important to enable intelligent transportation systems in a smart city. This problem is challenging due to the complicated spatial, short-term temporal and long-term periodical dependencies. Existing approaches have considered these factors in modeling. Most solutions apply CNN, or its extension Graph Convolution Networks (GCN) to model the spatial correlation. However, the convolution operator may not adequately model the non-Euclidean pair-wise correlations. In this paper, we propose a novel Attention-based Periodic-Temporal neural Network (APTN), an end-to-end solution for traffic foresting that captures spatial, short-term, and long-term periodical dependencies. APTN first uses an encoder attention mechanism to model both the spatial and periodical dependencies. Our model can capture these dependencies more easily because every node attends to all other nodes in the network, which brings regularization effect to the model and avoids overfitting between nodes. Then, a temporal attention is applied to select relevant encoder hidden states across all time steps. We evaluate our proposed model using real world traffic datasets and observe consistent improvements over state-of-the-art baselines.

The last few years have witnessed the rise of the big data era in which approximate nearest neighbor search is a fundamental problem in many applications, such as large-scale image retrieval. Recently, many research results have demonstrated that hashing can achieve promising performance due to its appealing storage and search efficiency. Since complex optimization problems for loss functions are difficult to solve, most hashing methods decompose the hash code learning problem into two steps: projection and quantization. In the quantization step, binary codes are widely used because ranking them by the Hamming distance is very efficient. However, the massive information loss produced by the quantization step should be reduced in applications where high search accuracy is required, such as in image retrieval. Since many two-step hashing methods produce uneven projected dimensions in the projection step, in this paper, we propose a novel dimension analysis-based quantization (DAQ) on two-step hashing methods for image retrieval. We first perform an importance analysis of the projected dimensions and select a subset of them that are more informative than others, and then we divide the selected projected dimensions into several regions with our quantizer. Every region is quantized with its corresponding codebook. Finally, the similarity between two hash codes is estimated by the Manhattan distance between their corresponding codebooks, which is also efficient. We conduct experiments on three public benchmarks containing up to one million descriptors and show that the proposed DAQ method consistently leads to significant accuracy improvements over state-of-the-art quantization methods.

Data-parallel applications generate a mix of coflows with and without deadlines. Deadline coflows are mission-critical and must be completed within deadlines, while the non-deadline coflows desire to be completed as soon as possible. Scheduling such mix-coflows is an important problem in modern datacenters. However, existing solutions only focus on one of the two types of coflows: they either solely concentrate on meeting the deadlines of deadline-aware coflows or reducing the coflow completion times (CCTs) of non-deadline coflows. In this article, we study the problem of optimizing deadline and non-deadline coflows simultaneously. To this end, we present a new optimization framework,mixCoflow, to schedule deadline coflows to minimize and balance their bandwidth footprint, such that non-deadline coflows can be scheduled as early as possible. Specifically, we develop the mathematical model and formulate the scheduling problem for deadline coflows as a lexicographical min-max integer linear programming (ILP) problem. Through rigorous theoretical analysis, this ILP problem has been proved to be equivalent to a linear programming (LP) problem that can be solved with standard LP solvers. By solving this LP,mixCoflowis able to balance the bandwidth footprint of deadline coflows while guaranteeing their deadlines. As a result, non-deadline coflows can be scheduled as soon as possible whenever they arrive. To demonstrate the effectiveness of our work, we have conducted extensive simulations based on a widely used Facebook data trace. The simulation results verify thatmixCoflowcan achieve significant improvement on the average CCT of non-deadline coflows, at no expense of increasing the deadline miss rates of deadline coflows, when compared to the state-of-art solutions.

Low latency stream processing on large clusters consisting of hundreds to thousands of servers is an increasingly important challenge. A crucial barrier to tackling this challenge is stragglers, i.e., tasks that are significantly straggling behind others in processing the stream data. However, prior straggler mitigation solutions have significant limitations. They balance streaming workloads among tasks but may incur imbalanced backlogs when the workloads exhibit variance, causing stragglers as well. Fortunately, we observe that carefully scheduling the outgoing tuples of different tasks can yield benefits for balancing backlogs, and thus avoids stragglers. To this end, we present Hone, a tuple scheduler that aims to minimize the maximum queue backlog of all tasks over time. Hone leverages an online Largest-Backlog-First (LBF) algorithm with a provable good competitive ratio to perform efficient tuple scheduling. We have implemented Hone based on Apache Storm and evaluated it extensively via both simulations and testbed experiments. Our results show that under the same workload balancing strategy-shuffle grouping, Hone outperforms the original Storm significantly, with the end-to-end tuple processing latency reduced by 78.7 percent on average.

High-resolution (HR) medical magnetic resonance images or computer tomography (CT) images can provide clearer anatomical details of human body, which facilitates early diagnosis of the diseases. However, due to the limitations of imaging environments, imaging systems and human factors, it is not easy to obtain clear HR medical images. In this paper, we propose a novel medical image super-resolution (SR) network, namely granular multi-scale network (GMSN), in the non-subsampled contourlet transform (NSCT) domain. GMSN mainly consists of a series of cascaded Res2Net blocks to exploit the multi-scale potential features at a granular level from medical images and increase the range of receptive fields for each network layer. The architecture possesses stronger multi-scale feature extraction ability, while maintaining a low computational load. In addition, most previous methods predict the HR images in the spatial domain, producing over-smoothed outputs while losing texture details. Thus, we formulate the medical image SR problem as the prediction of NSCT coefficients, which is able to further GMSN preserve richer structure details than that in spatial domain. The experimental results on our constructed medical image database show that our proposed method is capable of obtaining higher PSNR/SSIM values and preserving global topological structure and local texture detail better than other state-of-the-art methods.

Geographically distributed cloud is a promising technique to achieve high performance for service providers. For inter-datacenter transfers, deadline guarantee and fairness are the two most important requirements. On the one hand, to ensure more transfers finish before their deadlines, preemptive scheduling policies are widely used, leading to the transfer starvation problem and is hence unfair. On the other hand, to ensure fairness, inter-datacenter bandwidth is fairly shared among transfers with per-flow bandwidth allocation, which leads to deadline missing problem. A mechanism that achieves these two seemingly conflicting objectives simultaneously is still missing. In this paper, we propose TINA to schedule network transfers fairly while providing deadline guarantees. TINA allows each transfer to compete freely with each other for bandwidth. More specifically, each transfer is assigned a probability to indicate whether to transmit or not. We formulate the competition among the transfers as an El Farol game while keeping the traffic load under a threshold to avoid congestion. We then prove that the Nash Equilibrium is the optimal strategy and propose a light-weight algorithm to derive it. Finally, both simulations and testbed experiments results show that TINA achieves superior performance than state-of-art methods in terms of fairness and deadline guarantee rate.

Distributed streaming applications require the underlying network flows to transmit packets continuously to keep their output results fresh. These results will become stale if no updates come, and their staleness is determined by the slowest flow. At this point, coflows can be semantically comprised. Hence, efficient coflow transmission is critical for streaming applications. However, prior coflow-based solutions have significant limitations. They use a one-shot performance metric-CCT (coflow completion time), which cannot continuously reflect the staleness of the output results for a streaming application.To this end, we propose a new performance metric-coflow age (CA), for coflows generated by distributed streaming applications. The CA tracks the longest time-since-last-service among all flows in a coflow. In such a context, we consider a data center network with multiple coflows that continuously transmit packets between their source-destination pairs and address the problem of minimizing the average long-term CA while simultaneously satisfying the throughput constraints from the coflows. To solve this problem efficiently, we design a randomized algorithm and a drift-plus-age algorithm, and show that they can make the average long-term CA to achieve nearly two times and arbitrarily close to the optimal value, respectively. Through extensive simulations, we further demonstrate that both of the proposed algorithms can significantly reduce the CA of coflows, without violating the throughput requirement of any coflow, when compared to the state-of-the-art solution.

This paper studies the practically important problem of data collection for sensor-augmented RFID systems. However, existing RFID data collection protocols suffer from two common limitations: execution time is naturally in proportion to the number of tags, thus they cannot satisfy time-stringent application scenarios; none of them is complaint with the C1G2 standard, thus they cannot be implemented using Commercial-Off-The-Shelf (COTS) RFID tags. To overcome these two limitations, this paper proposes the Geographical correlation-based RF-data Collection (GRC) protocol. GRC is fast because it is able to approximately capture the sensing data of all tags by only actually gathering data from a small set of sampled tags. This is based on the observation from the real-world data set that sensing data has a strong geographical correlation, i.e., data gathered from nearby RFID tags has similar values. In GRC, we use a greedy approach to find the minimum sampling tag set to cover the whole monitoring region such that each un-sampled tag has at least one sampled tag nearby. Then, RFID reader runs the Framed Slotted Aloha (FSA) protocol specified in C1G2 standard to collect sensing data from the sampled tags. For each un-sampled tag, we approximate its sensing data by calculating weight-average of the data collected from its nearby sampled tags, where a faraway sampled tag should be given a small weight, and vice versa. Compared with existing RFID data collection schemes, the advantages of GRC are two-fold: (1) Extensive simulation results demonstrate that the time cost of our GRC scheme is only 1/28~1/3 of the state-of-the-art data collection scheme; (2) GRC is totally complaint with C1G2 standard, thus it can be easily deployed on the COTS RFID tags.

Tag inventory is one of the most fundamental tasks for RFID systems. However, the Framed Slotted Aloha (FSA) protocol specified in the C1G2 standard is of low time-efficiency, because it needs to collect all tags in the system. To improve time-efficiency, research communities proposed a batch of sampling-based approaches, in which the reader only needs to collect a small set of sampled tags instead of all. Although time-efficiency has been improved, existing sampling-based approaches still have two common limitations. First, all tags in the system are assumed to have the same sampling probability. It is unfair that tags attached to differential items (e.g., different values) have the same chance to be sampled and collected. Second, all existing sampling-based approaches stay in theory level and cannot be deployed on Commercial Off-The-Shelf (COTS) RFID devices, because the C1G2 standard does not support the sampling function at all. To deal with the above two limitations, this paper studies the new problem of differential tag sampling-letting each RFID tag be identified with a given sampling probability. In this paper, we use the COTS RFID devices including Impinj Speedway R420 reader and Monza 4QT tags to implement the Differential Tag Sampling (DTS) operation. Then, we apply probabilistic analytics on the collected tag data to address some practically important problems such as Multi-category Tag Cardinality Estimation (MTCE), and Value-based Missing Tag Detection (VMTD). Although the analytics results are not 100 percent accurate, the deviation in the results can be controlled below a small threshold and DTS can significantly improve the time-efficiency. DTS can be easily deployed on the COTS RFID systems, because it is totally compliant with the C1G2 standard. Extensive experiments demonstrate that DTS is able to let each tag take the given sampling probability to be sampled and identified. Moreover, the proposed DTS protocol can significantly reduce the execution time of MTCE and VMTD by nearly 70 percent than the FSA protocol.

Over the last decade, we have witnessed growing data volumes generated and stored across geographically distributed datacenters. Processing such geo-distributed datasets may suffer from significant slowdown as the underlying network flows have to go through the inter-datacenter networks with relatively low and highly heterogeneous available link bandwidth. Thus, optimizing the transmissions of inter-datacenter flows, especially coflows that capture application-level semantics, is important for improving the communication performance of such geo-distributed applications. However, prior solutions on coflow scheduling have significant limitations: they schedule coflows with already-fixed endpoints of flows, making them insufficient to optimize the coflow completion time (CCT). In this article, we focus on the problem of jointly considering endpoint placement and coflow scheduling to minimize the average CCT of coflows across geo-distributed datacenters. To solve this problem without any prior knowledge of coflow arrivals, we present a coflow-aware optimization framework called SmartCoflow. In SmartCoflow, we first apply an approximate algorithm to obtain the endpoint placement and scheduling decisions for a single coflow. Based on the single-coflow solution, we then develop an efficient online algorithm to handle the dynamically arrived coflows. Through rigorous theoretical analysis, we prove that SmartCoflow has a non-trivial competitive ratio. We also extend SmartCoflow to incorporate various design choices or requirements of applications and operators, such as enforcing an inter-datacenter bandwidth usage budget and considering coflow deadline. Through experimental results from testbed implementation and trace-driven simulations, we demonstrate that SmartCoflow can reduce the average CCT, lower bandwidth usage, and improve coflow deadline meet rate, when compared to the state-of-the-art scheduling-only method.

With the development of IoT techniques, urban computing has become an emerging topic in academia and industry. The goal of urban computing is to address some issues of urban planning by using the big data generated in urban facilities. The massive trajectory data processing is viewed as an important issue in urban computing. To satisfy the storage and processing requirements of massive trajectory data, a distributed system is usually adopted. However, existing distributed systems face challenges of data locality aware partitioning and various trajectory queries. In this paper, we propose a distributed framework of massive trajectory data analysis based on HBase, to realize spatio-temporal query more effectively. We first design a temporal-based pre-partitioning strategy to improve the performance of data written. Then we develop a Multi-Level Index to speed up the process of spatio-temporal query. Extensive experiments on real trajectory datasets demonstrate that the proposed framework significantly improves efficiency and usability.

In mobile edge computing (MEC) systems with energy harvesting, the mobile devices are empowered with the energy that harvested from renewable energy sources. On the other hand, mobile devices can offload their computation-intensive tasks to the MEC server to further save energy and reduce the task execution latency. However, the energy harvested is unstable and the mobile devices have to make sure that the energy should not be run out. Moreover, the wireless channel condition between the mobile device and the MEC server is dynamically changing, leading to unstable communication delay. Considering the energy constraints and unstable communication delay, the benefit of computation offloading is limited. In this paper, we investigate D2D-assisted computation offloading for mobile edge computing systems with energy harvesting. In our method, the mobile device is allowed to offload its tasks to the MEC server with the help of its neighbor node. More Specifically, the neighbor node acts as a relay to help the mobile device to communicate with the MEC server. Our goal is to minimize the average task execution time by selecting an optimal execution strategy for each task, i.e., whether to execute the task locally, or offload it to the MEC server directly, or offload it to the MEC server with the help of the most suitable neighbor node, or just to drop it. We propose a low-complexity online algorithm, which stem from Lyapunov Optimization-based Dynamic Computation Offloading (LODCO) algorithm, to solve this problem. Extensive simulations verified the effectiveness of the proposed algorithm, where the average task execution time is reduced around 50% as compared to that of the original LODCO algorithm.

In software-defined network, the use of distributed controllers to control forwarding devices has been proposed to solve the issues of scalability and load balance. However, the forwarding devices are statically assigned to the controllers in these distributed systems, which can overload some controllers while others are underutilized. In this paper, we propose an architecture named ASLB (active controller selection load balance), which proactively selects appropriate controllers for load balancing and minimize packet processing delays. We also present a novel active controller selection algorithm (ACS) for ASLB that efficiently schedules traffic from the switch to the controller and designs an intermediate coordinator for actively selecting a controller to serve a request. We built a system and evaluated it on a physical platform. The results show that ASLB is much better than the static allocation scheme in terms of minimizing latency, bandwidth utilization, and throughput.

Hashing methods have attracted much attention due to their superior time and storage properties for image retrieval. To learn similarity-preserving hash function, most existing methods are designed for the centralized setting. However, the current data storage systems are distributed to increase scalability. Obviously, it is infeasible to aggregate all the data into a fusion center because of the prohibitively expensive communication and computation overhead. Motivated by this, some methods are proposed to achieve hashing for distributed data. However, these methods mostly focus on extending one specific hashing to a distributed model without considering the generality. In this letter, we propose a novel general distributed hash learning model, which can be viewed as an effective distributed model of most hashing methods. The proposed model can achieve up to 15.2% accuracy gains over state-of-the-art distributed hashing methods, while the communication cost is independent on the data size.

The emerging deployment of large-scale cloud applications incurs significant inter-datacenter traffic, which makes the scarce wide-area bandwidth across data centers become the performance bottleneck. To achieve the desirable network performance, bandwidth guarantee should be provided for the resulting inter-datacenter traffic. However, the existing bandwidth allocation methods mainly focus on intra-datacenter traffic, and cannot achieve the cost-minimizing bandwidth guarantee for inter-datacenter traffic. In this paper, we focus on the bandwidth guarantee problem for inter-datacenter traffic and present a novel bandwidth allocation model. Our model can ensure the bandwidth guarantee, minimize the resulting network cost, and efficiently avoid the potential traffic overload on low cost links. To solve the large-scale optimization problem in our model, we are motivated to develop a distributed algorithm by blending the advantages of alternating direction method of multipliers (ADMM) and the auxiliary variable method. Specifically, we efficiently decompose the optimization problem into many small sub-problems, which are allowed to be processed in a large-scale computing environment, where each server solves a few small sub-problems. We further present a theoretically proved globally, asymptotically stable algorithm to solve these sub-problems. Extensive evaluation results demonstrate that our bandwidth allocation method can effectively realize the bandwidth guarantee for inter-datacenter traffic with reduced network cost and outperforms the prior method PS-L. In particular, the total network cost is reduced by 59.57 percent on average.

Currently, parallel data transmissions in large-scale datacenter networks are becoming increasingly crucial to application performance. Despite fine-grained control by SDN-enabled networks, some transmission errors, such as misconfigurations, will inevitably occur, resulting in high-level forwarding policies that cannot be conformed to at the data plane. Therefore, flow trajectory detection is very important for allowing datacenter network operators to troubleshoot problems and ensure that all traffic flows are running on the correct paths. However, existing solutions detect flow trajectories by recording the entire path of each packet. These methods are prone to imposing significant overheads in terms of both the number of switch entries and the amount of packet header space required. To considerably reduce this overhead, we present FlowTracer, an efficient flow trajectory detection solution, which can sample a path one link at a time instead of recording the entire path. FlowTracer consists of a method of probabilistic packet tagging and a method of trajectory reconstruction. In this paper, we first introduce the method of probabilistic packet tagging, which is performed in OpenFlow-enabled switches with very few switch entries and limited packet header space by means of double VLAN tags. Then, we explore the topological structure of datacenter networks and propose our method of trajectory reconstruction, which is performed at end hosts and achieves rapid convergence. Finally, we evaluate FlowTracer on a 48-ary fat-tree topology. The results show that FlowTracer can detect trajectories quickly while placing far smaller demands on both switch entries and packet header space than state-of-the-art techniques.

In data centers, a lot of cluster computing applications follow the coflow working pattern. That is, a collection of flows between two groups of machines is semantically related. On the other hand, network function virtualization sufficiently improves the performance of data center networks. It however complicates the network environment by introducing many multi-function middleboxes each with multiple resources. Coflows encounter extremely different processing delays under diverse network functions. Prior coflow scheduling schemes are insufficient to guarantee the coflow completion time in the multi-resource environment. In this paper, we propose, model, and analyze the coflow scheduling problem in the multi-resource environment. We present a dedicated method, data rate guarantee for coflow (DRGC), to guarantee the data rate requirements of coflows in this situation. DRGC prioritizes the coflow scheduling sequence, assigns precise data rates for coflows, and deploys a packet scheduling algorithm at middleboxes to guarantee their transmissions. In our experiments, DRGC efficiently guarantees the completion times of coflows and supports 15% more workload, compared with other scheduling schemes.

Coflow scheduling is crucial to improve the communication performance of data-parallel jobs, especially when these jobs running in the inter-datacenter networks with limited and heterogeneous link bandwidth. However, prior solutions on coflow scheduling assume the endpoints of flows in a coflow to be fixed, making them insufficient to optimize the coflow completion time (CCT). In this paper, we focus on the problem of jointly considering endpoint placement and coflow scheduling to minimize the average CCT of coflows across geo-distributed datacenters. We first develop the mathematical model and formulate a mixed integer linear programming (MILP) problem to characterize the intertwined relationship between endpoint placement and coflow scheduling, and reveal their impact on the average CCT. Then, we present SmartCoflow, a coflow-aware optimization framework, to solve the MILP problem without any prior knowledge of coflow arrivals. In SmartCoflow, we first apply an approximate algorithm to obtain the endpoint placement and scheduling decisions for a single coflow. Based on the single-coflow solution, we then develop an efficient online algorithm to handle the dynamically arrived coflows. To validate the efficiency and practical feasibility of SmartCoflow, we implement it as a real-world coflow scheduler based on the Varys open-source framework. Through experimental results from both a small-scale testbed implementation and large-scale simulations, we demonstrate that SmartCoflow can achieve significant improvement on the average CCT, when compared to the state-of-the-art scheduling-only method.

Summary Enterprise IT departments must decide whether to build a private cloud themselves or lease space in a cloud instead. Actually, most enterprise cloud systems are heterogeneous simultaneously consisting partly of private cloud and partly of public cloud. An enterprise needs to know the optimal mixing ratio to achieve the lowest cost. However, the optimal ratio is a dynamic value because the popularity of data changes constantly and the data increases rapidly. In this paper, we formulate maximizing the net present value (NPV) of the investment revenue as a dynamic decision‐making problem and propose a model to simplify the decision‐making problem. We use the K‐means algorithm to cluster the devices into groups. In addition, we also provide a solution with the kNN algorithm for locating data. Furthermore, our system provides a cost‐saving solution that requires no additional devices or investment. Subsequently, as the system grows larger, the decision‐making solution can be employed again to determine how to best extend the cloud system at the lowest cost.

Summary Virtual cluster has recently emerged as a common abstraction for cloud applications or tenants to specify and reserve resources. Such virtual cluster brings valuable insights into the cloud elasticity when scaling up or down the number of resources on demand. Unfortunately, for global‐scale applications running on geo‐distributed datacenters, it is always a challenge to scale the virtual cluster. Due to the fact that the inter‐datacenter bandwidth is an expensive and scarce resource, it is increasingly important yet typically hard to achieve cost‐minimizing bandwidth guarantees when scaling. However, existing approaches mainly focus on the scaling within intra‐datacenter networks and cannot be simply extended to the inter‐datacenter scenario. In this paper, we study the problem of scaling up a virtual cluster with consideration of both bandwidth cost minimization and bandwidth guarantees fulfillment targeting inter‐datacenter networks. Specifically, we first propose an efficient algorithm to scale up the virtual cluster without changing its original VM placement. With the observation that such VM placement can hinder the cluster scalability, we further present an optimized algorithm, which exploits VM migration when scaling. Finally, we conduct extensive simulations to demonstrate the effectiveness of our algorithms, in terms of both bandwidth cost and the acceptance rate of scaling requests with bandwidth guarantees.

Nowadays, service function chaining (SFC) becomes more and more widespread and profound to implement flexible and economical virtual network infrastructures for the Internet of Things (IoT). With the benefits of SFC, the IoT service providers can steer massive traffic through a sequence of heterogeneous virtual network function instances based on their business logic. SFC is viewed as an attractive solution for building virtualized IoT-dedicated network. However, the SFC orchestration in IoT is still a challenge problem. Existing work usually focuses on the performance guarantee while ignoring the issue of resource idleness. To meet the sharp increase in IoT traffic amounts and the diversification of IoT traffic requirements, it is necessary to implement the performance and resource aware SFC orchestration system. Motivated by this, we propose a novel linear programming model and an effective approximation optimization algorithm for SFC orchestration, in order to achieve performance guarantee while avoiding resource idleness. Based on the proposed model and algorithm, a new prototype system named performance and resource aware orchestration system of SFC for IoT (PRSFC-IoT) is built upon OpenStack for online SFC orchestration. A large number of simulation experiments show that the PRSFC-IoT outperforms existing solutions for SFC orchestration in IoT.

Inefficient bandwidth sharing in a datacenter network, between different application frameworks, e.g., MapReduce and Spark, can lead to inelastic and skewed usage of link bandwidth and increased completion times for the applications. Existing work, however, either solely focuses on managing computation and storage resources or controlling only sending/receiving rate at hosts. In this paper, we present CoMan, a solution that provides global in-network bandwidth management in multiplexed data centers, with two goals: improving bandwidth utilization and reducing application completion time. CoMan first designs a novel abstraction of virtual link groups (VLGs) to establish a shared bandwidth resource pool. Based on this pool, CoMan implements a three-level bandwidth allocation model, which enables elastic bandwidth sharing among computing frameworks as well as guarantees network performance for the applications. CoMan further improves the bandwidth utilization by devising a VLG dependency graph and solves an optimization problem to guide the path selection using a 32-approximation algorithm. We conduct comprehensive trace-driven simulations as well as small-scale testbed experiments to evaluate the performance of CoMan. Extensive simulation results show that CoMan improves the bandwidth utilization and speeds up the application completion time by up to 2.83× and 6.68×, respectively, compared to the ECMP + ElasticSwitch solution. Our implementation also verifies that CoMan can realistically speed up the application completion times by 2.32× on average.

Increasing number of Internet-scale applications, such as video streaming, incur huge amount of wide area traffic. Such traffic over the unreliable Internet without bandwidth guarantee suffers unpredictable network performance. This result, however, is unappealing to the application providers. Fortunately, Internet giants like Google and Microsoft are increasingly deploying their private wide area networks (WANs) to connect their global datacenters. Such high-speed private WANs are reliable, and can provide predictable network performance. In this paper, we propose a new type of service-inter-datacenter network as a service (iDaaS), where traditional application providers can reserve bandwidth from those Internet giants to guarantee their wide area traffic. Specifically, we design a bandwidth trading market among multiple iDaaS providers and application providers, and concentrate on the essentialbandwidth pricingproblem. The involved challenging issue is that the bandwidth price of each iDaaS provider is not only influenced by other iDaaS providers, but also affected by the application providers. To address this issue, we characterize the interaction between iDaaS providers and application providers using a Stackelberg game model, and analyze the existence and uniqueness of the equilibrium. We further present an efficient bandwidth pricing algorithm by blending the advantage of a geometrical Nash bargaining solution and the demand segmentation method. For comparison, we present two bandwidth reservation algorithms, where each iDaaS provider's bandwidth is reserved in a weighted fair manner and a max-min fair manner, respectively. Finally, we conduct comprehensive trace-driven experiments. The evaluation results show that our proposed algorithms not only ensure the revenue of iDaaS providers, but also provide bandwidth guarantee for application providers with lower bandwidth price per unit.

The emerging deployment of geographically distributed data centers (DCs) incurs a significant amount of data transfers over the Internet. Such transfers are typically charged by Internet Service Providers (ISPs) with the widely adopted q-th percentile charging model. In such charging model, the time slots with top 100-q percent of data transmission do not affect the total transmission cost, and can be viewed as free. This brings the opportunity to optimize the scheduling of inter-DC transfers to minimize the entire transmission cost. However, very little work has been done to exploit those free time slots for scheduling inter-DC transfers. The crux is that existing work either lacks a mechanism to accumulate traffic to free time slots, or inevitably relies on prior knowledge of traffic arrival patterns. In this paper, we attempt to exploit those free time slots by leveraging diverse time-sensitivities among inter-DC transfers, so as to reduce or even minimize the transmission cost. Specifically, we advocate that a simple principle should be followed: more traffic peaks should be scheduled in free time slots, while less traffic differentiation should be maintained among the remaining time slots. To this end, we take advantage of the Lyapunov optimization techniques to design a pricing-aware control framework. This framework efficiently makes online decisions for inter-DC transfers without requiring a prior knowledge of traffic arrivals. To verify our proposed framework, we conduct small-scale testbed implementation. The results show that our framework can realistically reduce the transmission cost by up to 19.38%.

In the era of global-scale services, analytical queries are performed on datasets that span multiple data centers (DCs). Due to the scarce and expensive inter-DC bandwidth, various methods have been proposed to reduce either the traffic cost or the completion time for those analytics queries. However, current methods make no attempt to maximize the number of successfully served query requests. Moreover, most of them rely on unrealistic assumptions - such as analytical queries are repeated or known in advance. In this paper, we target at characterizing and optimizing the cost-performance tradeoff for geo-distributed data analytics. Our objectives are two-fold: (1) we minimize the inter-DC traffic cost when serving geo-distributed analytics with uncertain query demand, and (2) we maximize the system throughput, in terms of the number of query requests that can be successfully served with guaranteed queuing delay. To achieve these objectives, we take advantage of Lyapunov optimization techniques to design a two-timescale online control framework. Without prior knowledge of future query requests, this framework makes online decisions on input data placement and admission control of query requests. Extensive trace-driven simulation results demonstrate that our framework is capable of reducing inter-DC traffic cost, improving system throughput and guaranteeing a maximum delay for each query request.

Summary In wireless sensor networks, a lot of applications need the sensed information be transmitted to the sink node within a predefined time threshold. So end‐to‐end delay is an important performance metric in wireless sensor networks. Opportunistic routing protocols have been proposed to reduce the waiting delay. In the duty cycle networks, increasing the duty cycle ratio can also reduce the end‐to‐end delay. However, this method will consume more energy. It is obvious that there exists a trade‐off between delay and energy consumption. So adjusting the duty cycle ratio of each node can investigate this trade‐off. To the best of our knowledge, no existing work takes both of end‐to‐end delay and energy efficiency into consideration in the opportunistic routing networks. In this paper, we want to minimize the whole energy consumption while guaranteeing the expected end‐to‐end delay between the source nodes and the sink node is below the given threshold. To deal with this problem, we propose a dynamic duty cycle scheme which can significantly reduce the energy consumption and guarantee the expected end‐to‐end delay demand in the opportunistic routing network. To be specific, firstly, we formulate a new metric with the wake‐up time slots as the variable to measure the end‐to‐end delay. Secondly, for simplifying the complex problem, we decompose it into a set of single‐hop delay guarantee problems. Feedback controller has been used to solve the problem. We also analyze the influence of the multiple receivers in the same forwarding set. Finally, we conduct extensive simulations to evaluate the performance of the proposed algorithm. The experimental results reveal that our scheme can guarantee the delay requirement, meanwhile, significantly reduce the energy consumption compared with prior schemes.

Nowadays, Internet has became an indispensable part of our daily life. The network keeps expanding in scales, and the network technologies develop rapidly. From the network architecture to the network application, new technologies emerge constantly. For example, Software Defined Networking (SDN) introduces the programmability of network to achieve more flexible network management. Device-to-device (D2D) communication provides the promising solutions for social applications and content sharing applications in mobile network. Cloud computing and fog computing make it possible to maximize the utilization of network resources. This special issue is intended to introduce some new advances in the above future network technologies, in which eight high-quality papers are presented. There are three papers about SDN including the synchronization of multidomain controllers, the placement of SDN controller, and the traffic scheduling in SDN. Three papers focus on cloud computing, whose topics include the IaaS cloud broker, the fog computing, and the architecture of virtual machine storage. Other two papers are about D2D network and multicast network, respectively. The general objective of this special issue is to show the state-of-the-art in future network technologies and to point out the directions of the research on future network. As an emerging network technology, software-defined networking (SDN) attracts more and more attentions from both academia and industry. The SDN provides solutions to decouple the network control and the data transmission, which enables the network to become programmable. According to the northbound API provided by SDN controller, people can manage the network more flexibly and implement the network innovation applications more easily. However, the research in the context of SDN is still in its infancy. There are many important issues that should be addressed. Zou et al try to address the synchronization problem of multidomain SDN controllers in their paper “Active Synchronization of Multi-domain Controllers in Software Defined Networks.”1 In massive data center networks, the single controller is the bottleneck of the SDN deployment. To improve the scalability and reliability, multidomain controllers are usually adopted. Within this context, the the synchronization of multidomain controllers should be considered. To overcome the drawbacks of existing PS-based synchronization algorithm, Zou et al propose an active synchronization algorithm, which is flexibly triggered by events without considering the effect of time on synchronization. The evaluation results show that the active synchronization algorithm can achieve better load balance with less overhead. Zhao et al focus on the problem of SDN controller placement in their paper “Scalable SDN Architecture with Distributed Placement of Controllers for WAN.”2 A large number of SDN controllers are usually required to construct a scalable SDN architecture for WAN. In this architecture, the network performance should be affected by the number and the location of controllers. To minimize the number of controllers and the link delay between switches and controllers, Zhao et al propose a novel method of controller placement. In the proposed method, they formulate an Integer Linear Program (ILP) model for controller placement in WAN. Then they propose a heuristic algorithm as the solution. The numerical simulation results show that the proposed method can achieve better results. Huang et al study the problem of traffic scheduling for SDN with Deep Packet Inspection (DPI) proxies in their paper “Traffic Scheduling for Deep Packet Inspection in Software Defined Networks.”3 With the development of SDN, more and more network function proxies are deployed for flow processing, such as DPI, Firewall, and Load balancer. To ensure acceptable performance of DPI proxies in SDN, Huang et al propose an Integrated DPI Proxy Allocation and routing Determining (IPAD) problem and then formulate this problem as an Integer Linear Programming (ILP) model with the goal of minimizing the overall latency in DPI. They design a two-phase algorithm to solve the IPAD problem. The simulation results show that the proposed algorithm outperforms existing benchmark algorithms. Nowadays, cloud computing has become a highly demanded service. More and more people accept the “pay as you go” model of computing resources. More and more companies prefer using the cloud services to avoid up-front infrastructure costs. However, there are some new challenges rising with the development of cloud computing, such as the model of cloud broker, the concept of fog computing. Chen et al focus on the reservation schemes for IaaS cloud broker in their paper “Reservation Schemes for IaaS Cloud Broker: A Time-multiplexing Way for Different Rental Time.”4 As a novel service model, the cloud broker coordinates the demand of users with IaaS providers to gain high profit by leveraging the time multiplexing of VM instances and the pricing gap between on-demand billing and reservation billing. To achieve this goal, an effective reservation scheme for cloud broker is necessary. Chen et al propose an offline and an online reservation algorithms to reduce the cost of cloud broker without partitioning users' rental time. The offline algorithm is based on the demand graph while the online algorithm is based on the history data. The experimental results show that two kinds of proposed algorithms both can reduce the cost of cloud broker. Yao et al focus on the solution of cloudlet deployment for cost-effective fog computing in their paper “Heterogeneous Cloudlet Deployment and User-Cloudlet Association towards Cost Effective Fog Computing.”5 With the development of mobile cloud computing, the concept of fog computing is proposed to bring the cloud facilities closer to mobile users, thereby achieving low communication latency between mobile devices and the cloud. In fog computing research community, it is a big challenge to deploy the cloudlet servers in a cost-effective manner. To address this issue, Yao et al propose a method of heterogeneous cloudlet deployment. They formulate the deployment problem into an Integer Linear Programming model and then design a heuristic algorithm to solve it. The evaluation results validate the high efficiency of their method. Chen et al focus on the storage architecture with nonvolatile memory (NVM) device for virtual machines in their paper “MBSA: A Lightweight and Flexible Storage Architecture for Virtual Machines.”6 In cloud computing, virtualization technology is a good solution to provide a powerful computing platform. However, virtualization technology is not suitable for data-intensive workloads due to high I/O virtualization overhead. To address this problem, Chen et al propose a memory bus-based storage architecture named MBSA, in which the performance of NVM devices is fully explored to improve the storage performance of VMs. Compared with the conventional storage architecture, the MBSA can take full advantage of NVM devices. Experimental results show that the MBSA provides good performance on balancing the write operations to NVM devices with low overhead. Device-to-device (D2D) is a local short-range communication mode, by which files can be cached and shared among mobile devices within a certain range. Because D2D can be used to reduce the traffic explosion through direct communication without traversing the base station or core network, D2D sharing is viewed as a promising application of mobile social networks. To provide better quality of experience for D2D sharing, Wang et al try to design and implement a big data processing platform to analyze the empirical trace data from D2D sharing application in their paper “A Measurement Study of Device-to-Device Sharing in Mobile Social Networks Based on Spark.”7 They get a large-scale trace dataset from Xender, which is one world's leading D2D sharing application. Then they build a big data processing platform based on Spark to explore the characteristics of application and users. Finally, they discuss the potential methods to improve Xender's quality of service. Multicast is usually used to deliver the same content from a single source to a set of destinations. By multicast, the unnecessary duplicated transmissions can be reduced to save bandwidth efficiently. Although it is a traditional group communication method, multicast faces new challenges from some emerging Internet applications. For example, the problem of uncertain multicast is usually caused by content replica design in content distribution network (CDN) and datacenter network (DCN). Ren et al try to address the packing problem of uncertain multicast to minimize the total transmission cost in their paper “The Packing Problem of Uncertain Multicasts.”8 To give effective solution of network resources sharing when a set of uncertain multicast occupy the network simultaneously, they formally present the packet problem of uncertain multicast. Then they prove this problem is NP-hard and design two greedy algorithms approximating the optimal solution. Finally, they conduct large-scale simulations to verify the effectiveness and efficiency of the proposed algorithms. In this special issue, there are eight high-quality papers about recent advances in future network technologies. From these papers, we can see that SDN, Cloud, and D2D remain promising technologies of future network. With the development of these technologies, there are many important issues that need to be addressed. Moreover, we also can see that innovative Internet applications bring new challenges to traditional network technology, such as uncertain multicast in CDN or DCN. These new challenges should be addressed to promote the development of future network. We hope that the readers can benefit from this special issue. This work is supported by the JSPS KAKENHI grant number JP16F16349, JSPS KAKENHI grant number JP16K00117, KDDI Foundation, and the Dalian High-level Talent Innovation Program (No. 2015R049).

In large-scale Internet applications running on geographically distributed datacenters, such as video streaming, it is important to efficiently allocate requests among datacenters. To the best of our knowledge, existing approaches, however, either solely focus on minimizing total cost for provider, or guaranteeing QoS for end-users. In this paper, we apply the software defined network (SDN) controller to enable the central control of the entire network, and propose a joint optimization model to consider high bandwidth utilization for provider and low delay for users. We present the Nash bargaining solution (NBS) based method to model both requirements of provider's high bandwidth utilization and end-users' low delay. Specifically, we formulate the design of request allocation under those requirements as an optimization problem, which is NP-hard. To solve such hard optimization problem, we develop an efficient algorithm blending the advantages of Logarithmic Smoothing technique and the auxiliary variable method. According to the theoretical analysis, we verify the existence and uniqueness of our solution and the convergence of our algorithm. We conduct a large amount of experiments based on real-world workload traces and demonstrate the efficiency of our algorithm compared to both greedy and locality algorithms.

For large-scale RFID systems, this paper studies the practically important problem of target tag information collection, which aims at collecting information from a specific set of target tags instead of all. However, the existing solutions are of low time-efficiency because of two reasons. First, the serious collisions among tags due to hashing randomness seriously reduce the frame utilization, whose upper bound is just 36.8 percent. Second, they cannot efficiently distinguish the target tags from the non-target tags and thus inevitably collect a lot of irrelevant information on non-target tags, which further deteriorates the effective utilization of the time frame. To overcome the above two drawbacks, this paper proposes the minimal Perfect hashing-based Information Collection (PIC) protocol, which first leverages lightweight indicator vectors to establish a one-to-one mapping between target tags and slots, thereby improving the frame utilization to nearly 100 percent; and then uses the novel data structure called Minimal Perfect Hashing based Filter (MPHF) to filter out the non-target tags, thereby preventing them from interfering with the process of collecting information from target tags. Sufficient theoretical analyses are also presented in this paper to minimize the execution time of the proposed PIC protocol. Extensive simulations are conducted to compare the proposed PIC protocol with prior works side-by-side. The simulation results demonstrate that PIC significantly outperforms the state-of-the-art protocols in terms of time-efficiency.

With the increasing popularization of radio frequency identification (RFID) technology in the retail and logistics industry, RFID privacy concern has attracted much attention, because a tag responds to queries from readers no matter they are authorized or not. An effective solution is to use a commercially available blocker tag that behaves as if a set of tags with known blocking IDs are present. However, the use of blocker tags makes the classical RFID estimation problem much more challenging, as some genuine tag IDs are covered by the blocker tag and some are not. In this paper, we propose RFID estimation scheme with blocker tags (REB), the first RFID estimation scheme with the presence of blocker tags. REB uses the framed slotted Aloha protocol specified in the EPC C1G2 standard. For each round of the Aloha protocol, REB first executes the protocol on the genuine tags and the blocker tag, and then virtually executes the protocol on the known blocking IDs using the same Aloha protocol parameters. REB conducts statistical inference from the two sets of responses and estimates the number of genuine tags. Rigorous theoretical analysis of parameter settings is proposed to guarantee the required estimation accuracy, meanwhile minimizing the time cost and energy cost of REB. We also reveal a fundamental tradeoff between the time cost and energy cost of REB, which can be flexibly adjusted by the users according to the practical requirements. Extensive experimental results reveal that REB significantly outperforms the state-of-the-art identification protocols in terms of both time efficiency and energy efficiency.

In large-scale radio frequency identification (RFID)-enabled applications, we sometimes only pay attention to a small set of key tags, instead of all. This paper studies the problem of key tag population tracking, which aims at estimating how many key tags in a given set exist in the current RFID system and how many of them are absent. Previous work is slow to solve this problem due to the serious interference replies from a large number of ordinary (i.e., non-key) tags. However, time-efficiency is a crucial metric to the studied key tag tracking problem. In this paper, we propose a singleton slot-based estimator, which is time-efficient, because the RFID reader only needs to observe the status change of expected singleton slots corresponding to key tags instead of the whole time frame. In practice, the ratio of key tags to all current tags is small, because key members are usually rare. As a result, even when the whole time frame is long, the number of expected singleton slots is limited and the running of our protocol is very fast. To obtain good scalability in large-scale RFID systems, we exploit the sampling idea in the estimation process. A rigorous theoretical analysis shows that the proposed protocol can provide guaranteed estimation accuracy to end users. Extensive simulation results demonstrate that our scheme outperforms the prior protocols by significantly reducing the time cost.

Organizing shapes by convex parts is a fundamental procedure for many shape-related applications. However, convexity is sensitive to noise and shape variations. Recent publications in the field concentrated on decomposing shapes into near-convex parts. Although a variety of methods have been presented, there is still a need for a robust and versatile method, especially when a shape possesses long curved branches such as a lizard with a long curved tail. It is difficult to capture the tail as a whole part because its concavity is too high based on classic measures. To address this issue, we propose a `Visibility Range', novel shape signature in this paper. Visibility range reaches low values for points in concave regions and high values in convex regions. Moreover, a novel concavity measure based on visibility range is presented. Compared to previous measures, the novel measure describes long curved branches better. With these, a simple but effective shape decomposition algorithm is designed. The decomposition is formulated as a problem of detecting points with extreme visibility range in a visibility matrix. Extensive experiments have been done on shapes with various kinds of near-convex parts, demonstrating that the proposed method is more robust and effective than the state-of-art methods based on other concave-convex features.

In data centers, a lot of cluster computing applications follow a coflow pattern. On the other hand, network function virtualization (NFV) sufficiently improves the performance of the data center network. However, coflows encounter extremely different processing delays under diverse network functions. Traditional coflow scheduling schemes become insufficient in this situation. Based on the observation that the benefit of coflows is closely related to the data rates of flows, we propose DRGC (Data Rate Guarantee for Coflow) to guarantee the data rate requirements of coflows in the NFV environment. We prioritize the scheduling sequence of coflows, precisely allocate data rates for individual flows, and design an efficient scheduling algorithm. DRGC maintains the desired data rates of coflows with higher priorities at middleboxes and leaves more scheduling opportunities to the ones with lower priorities. In the large-scale trace-driven experiment, DRGC efficiently guarantees the data rate requirements of coflows and supports more 15% workload, compared with other scheduling schemes.

This paper studies the problem of data collection in sensor-augmented RFID networks: how to quickly obtain the error-bounded data from sensor-augmented RFID tags. Existing data collection protocols require each tag to transmit the sensor data to the reader through a low-rate channel. However, in large-scale RFID system, they take too long time and block other time-sensitive operations. By exploring the correlation of sensor data, our Sampling-based Information Collection (SIC) protocol significantly reduces the number of responding tags. Specifically, SIC obtains an error bound based on the estimation model by using some randomly-sampled data. The error bound is expected to maximize the number of data within it. These data can be seen as a cluster and be approximated by one value within the error bound. Then, SIC only needs to collect the data of out this cluster, thereby significantly reducing the data transmission. It minimizes the execution time by optimizing the sample size and estimating the number of tags out of the error bound. We conduct extensive simulations to evaluate the performance of SIC and compare it with three major related work. The results demonstrate that SIC is 1 to 10 times faster than the state-of-the-art solution.

As a novel architecture, software-defined networking (SDN) is viewed as the key technology of future networking. The core idea of SDN is to decouple the control plane and the data plane, enabling centralized, flexible, and programmable network control. Although local area networks like data center networks have benefited from SDN, it is still a problem to deploy SDN in wide area networks (WANs) or large-scale networks. Existing works show that multiple controllers are required in WANs with each covering one small SDN domain. However, the problems of SDN domain partition and controller placement should be further addressed. Therefore, we propose the spectral clustering based partition and placement algorithms, by which we can partition a large network into several small SDN domains efficiently and effectively. In our algorithms, the matrix perturbation theory and eigengap are used to discover the stability of SDN domains and decide the optimal number of SDN domains automatically. To evaluate our algorithms, we develop a new experimental framework with the Internet2 topology and other available WAN topologies. The results show the effectiveness of our algorithm for the SDN domain partition and controller placement problems.

In Wireless Sensor Networks (WSNs), end-to-end delay is an important metric because the sensed information is necessary to be transmitted to the sink node within a predefined time threshold. Therefore, opportunistic routing protocols are proposed to reduce the end-to-end delay. As a matter of fact, increasing the number of wake-up slots will certainly reduce the transmission delay, however, also consumes more energy. Hence, it is interesting to control the number of wake-up slots to investigate the trade-off between the end-to-end delay and the energy-efficiency. To the best of our knowledge, no existing work takes both of end-to-end delay and energy-efficiency into consideration in the opportunistic routing networks. Therefore, this paper studies how to minimize the energy-consumption while guaranteeing that the expected end-to-end delay is below a given threshold. To solve this problem, we propose an energy-based Dynamic Duty Cycle Scheme(D2CS) in opportunistic routing network. Specifically, we first present an analytical model to measure the expected end-to-end delay. Then, we decompose the studied problem into a set of single-hop delay guarantee problems and using the feedback controller to approximate the optimal solution. Finally, extensive simulations are conducted to evaluate the performance of the proposed D2CS algorithm. The experimental results reveal that our D2CS can guarantee the delay requirement, meanwhile, significantly reduce the energy consumption compared with prior schemes.

Traditional networks are surprisingly fragile and difficult to manage. Software Defined Networking (SDN) gained significant attention from both academia and industry, as if simplify network management through centralized configuration. Existing work primarily focuses on networks of limited scope such as data-centers and enterprises, which makes the development of SDN hindered when it comes to large-scale network environments. One way of enabling communication between data-centers, enterprises and ISPs in a large-scale network is to establish a standard communication mechanism between these entities. In this paper, we propose Zebra, a framework for enabling communication between different SDN domains. Zebra has two modules: Heterogeneous Controller Management (HCM) module and Domain Relationships Management (DRM) module. HCM collects network information from a group of controllers with no interconnection and generate a domain-wide network view. DRM collects network information from other domains to generate a global-wide network view. Moreover, HCM supports different SDN controllers, such as floodlight, maestro and so on. To test this framework, we develop a prototype system, and give some experimental results.

The widely used RFID tags impose serious privacy concerns as a tag responds to queries from readers no matter they are authorized or not. The common solution is to use a commercially available blocker tag which behaves as if a set of tags with known blocking IDs are present. The use of blocker tags makes RFID estimation much more challenging as some genuine tag IDs are covered by the blocker tag and some are not. In this paper, we propose REB, the first RFID estimation scheme with the presence of blocker tags. REB uses the framed slotted Aloha protocol specified in the C1G2 standard. For each round of the Aloha protocol, REB first executes the protocol on the genuine tags and the blocker tag, and then virtually executes the protocol on the known blocking IDs using the same Aloha protocol parameters. The basic idea of REB is to conduct statistically inference from the two sets of responses and estimate the number of genuine tags. We conduct extensive simulations to evaluate the performance of REB, in terms of time-efficiency and estimation reliability. The experimental results reveal that our REB scheme runs tens of times faster than the fastest identification protocol with the same accuracy requirement.

Summary This paper focuses on the topology structure of interconnection networks. To overcome drawbacks in the existing hypercube structure, we present an exchanged folded hypercube (EFH) structure, which is an improvement of the exchanged hypercube. Compared with the existing hypercube structures, EFH shows better performance in terms of many metrics such as smaller diameter, lower cost factor, and constant node degree. In this paper, we first introduce the structure of an EFH; then, we propose a routing algorithm and a load‐balancing algorithm for EFHs. Finally, we analyze the fault tolerance characteristics of EFHs including fault diameter and the cost effectiveness factor. Copyright © 2015 John Wiley & Sons, Ltd.

Unknown RFID tags appear when the unread tagged objects are moved in or tagged objects are misplaced. This paper studies the practically important problem of unknown tag detection while taking both time-efficiency and energy-efficiency of battery-powered active tags into consideration. We first propose a Sampling Bloom Filter which generalizes the standard Bloom Filter. Using the new filtering technique, we propose the Sampling Bloom Filter-based Unknown tag Detection Protocol (SBF-UDP), whose detection accuracy is tunable by the end users. We present the theoretical analysis to minimize the time and energy costs. SBF-UDP can be tuned to either the time-saving mode or the energy-saving mode, according to the specific requirements. Extensive simulations are conducted to evaluate the performance of the proposed protocol. The experimental results show that SBF-UDP considerably outperforms the previous related protocols in terms of both time-efficiency and energy-efficiency. For example, when 3 or more unknown tags appear in the RFID system with 30000 known tags, the proposed SBF-UDP is able to successfully report the existence of unknown tags with a confidence more than 99%. While our protocol runs 9 times faster than the fastest existing scheme and reducing the energy consumption by more than 80%.

Large-scale data centers are the key infrastructures for hosting and running a variety of applications. Besides traditional L2/L3 devices, middleboxes are widely deployed in data centers and perform many important functions, e.g., the intrusion detection and firewall. Middleboxes are equipped with multiple kinds of resources, such as CPU and memory. Data flows undergoing different functions have heterogeneous processing time requirements on diverse resources. Researchers are in a dilemma as to how to provide fair service for flows and efficiently utilize those scarce resources. To address this problem, we propose a novel packet scheduling method, active time fairness queuing (ATFQ), for multi-resource environments. Prior packet scheduling methods usually focus on pursuing the fairness among flows, resulting in enormous waste of those scarce resources. ATFQ overcomes this essentially by redefining the fairness and can maximize the resource utilization with the guarantee of fairness. We conduct extensive simulations to evaluate the performance of ATFQ. The evaluation results demonstrate that flows get better service in many aspects under ATFQ. Meanwhile, the resource utilization rises up by about 10% than the traditional DRFQ, which is one of the mainstream involved methods.

Shape retrieval is still a challenging problem. To address this problem, there is a growing interest in context-sensitive shape retrieval. In such methods, when computing the similarity between any two shape objects, the influence of their neighbors is propagated by a diffusion process on contextual space. We can re-evaluate the pairwise similarities to get better retrieval results. Although the context-sensitive shape retrieval methods are very promising in theory, but the time efficiency of these methods is a big problem in practice. Because it is a time-consuming to diffuse on the contextual space formed by all shape objects in the database. To improve the time efficiency, we propose a novel framework for context-sensitive shape retrieval. This framework consists of the off-line and the on-line stages. In the off-line stage, we conduct context-sensitive based shape clustering to build index. In the on-line stage, we quickly find candidate shape objects for the query by the index. Then, by diffusion on the context space consisting of candidate objects, we can get the final retrieval results. Compared with the state-of-the-art methods, the proposed framework achieves the same bull's-eye retrieval score with more than 80% reduction of searching time on the standard shape database.

In RFID-enabled applications, we may pay more attention to key tags instead of all tags. This paper studies the problem of key tag counting, which aims at estimating how many key tags in a given set exist in the current RFID system. Previous work is slow to solve this new problem because of the serious interference replies from the large number of ordinary (i.e., Nonkey) tags. However, time-efficiency is an important metric for the fast tag cardinality estimation in a large-scale RFID system. In this paper, we propose a singleton slot-based estimator, which is time-efficient because the RFID reader only needs to observe the status change of expected singleton slots of key tags instead of the whole time frame. In practice, the ratio of key tags to all current tags is small for "key" members should be rare. As a result, even when the whole time frame is long, the expected singleton slot number is limited and the running of our protocol is fast to achieve estimation accuracy. Rigorous theoretical analysis shows that the proposed protocol can provide guaranteed estimation accuracy to end users. We conduct simulations and implement a prototype of our protocol to verify its efficiency and deployability.

As a kind of interesting mobile application, Mobile Visual Search (MVS) has attracted extensive research efforts from both academy and industry. Most of the MVS systems adopt the client-server framework, in which transmission latency caused by the limited bandwidth in wireless network is a big problem. To address this problem, the state-of-the-art work focuses on designing low bit-rate descriptors for MVS. However, few work focuses on reducing the number of descriptors. To further reduce the latency, we propose a novel framework of MVS based on the weighted matching of dominant descriptor. Firstly, we present an affinity propagation based algorithm for dominant descriptor selection. Secondly, we propose a weighted feature matching method to consider the differences of dominant descriptors in feature matching. By the proposed framework, we not only reduce the network latency in MVS, but also avoid transmitting useless descriptors to improve the retrieval accuracy of MVS. The experimental results on Stanford MVS data set show that when using CHoG descriptors, the proposed framework outperforms the existing framework by reducing more than 40% of the amount of data transmission and increasing 5% of the average retrieval accuracy.

Due to the bandwidth limitation in wireless networks, transmission overhead is a big problem in Mobile Visual Search (MVS). Existing work proposes transmitting the compressed local feature descriptors instead of the query image to reduce the transmission overhead. Although many kinds of compressed descriptors are proposed, designing a suitable lossless compressed descriptor has proven elusive. In this paper, we propose a novel framework for MVS with low transmission overhead rather than focusing on compressed descriptors. The key point of the proposed framework is to migrate the vector quantization in the bag of visual words model from the server to the client. In this framework, no matter what descriptors are used, the client only transmits the ID numbers of the visual words to the server, thereby reaching the minimal possible transmission overhead. To achieve this goal, we present vocabulary decomposition by which we can decompose the large vocabulary into several small ones satisfying storage constraints on mobile devices. In this paper, we first formulate vocabulary decomposition as an optimization problem. We then present Joint Product Quantization (JPQ) and Joint Optimized Product Quantization (JOPQ) to address the proposed optimization problem. Finally , we conduct a large number of simulation experiments and real experiments. The experimental results show that the proposed framework outperforms the existing framework by reducing more than 95% of the transmission overhead.

A shape signature is defined as any 1-D function on a shape, which is a compact and concise representation for some essence of the shape. Although a variety of shape signatures are proposed and utilized in shape retrieval and recognition tasks, the existing signatures cannot yet provide entirely satisfactory solutions to describe the shape variations well, especially when significant noise or articulation occurs. Motivated by the fact that electrical charge distributions are almost the same for similar shapes but not vice versa when shapes reach their electrical equilibrium condition, we propose a novel shape signature based on the electrical charge distribution on the shape (ECDS). Compared to other shape descriptors, ECDS is more intuitively and robust, which is computed in a global manner. Furthermore, as well as being invariant to translation, scale and rotation, ECDS is articulation insensitive and therefore exhibits better performance by the introduction of generalized coulomb potentials. This allows it to better match shapes whose parts can move independently, such as scissors. Finally, numerous experiments have done on public databases, demonstrating that ECDS has the above properties and compares well with other shape descriptors in many kinds of shape retrieval and recognition tasks.

Radio Frequency Identification (RFID) technology is widely used in the the retail, warehouse and supply chain management. However, unknown RFID tags appear when the unregistered tagged objects are moved in or tagged objects are misplaced, which leads to huge economic losses (e.g., misplaced chilled food in a warehouse may quickly decay). This paper studies the practically important problem of unknown tag detection. To the best of our knowledge, this is the first piece of work taking both time-efficiency and energy-efficiency into consideration, where the energy-efficiency is very important when the battery-powered active tags are used. This paper proposes two efficient protocols to address the problem of unknown tag detection. Specifically, the Basic Unknown Tag Detection (B-UTD) protocol leverages a cost-effective filter vector to detect the unknown tags, based on which we then propose a Sampling based Unknown Tag Detection (SUTD) protocol by adopting the well-known sampling idea. We present theoretical analysis to optimize the performance of the proposed protocols. Extensive simulations are conducted to evaluate the performance of the proposed protocols. And the experimental results show that the proposed S-UTD protocol considerably outperforms the most related protocol by reducing more than 90% of the required execution time and energy consumption.

To eliminate the influence of subjective factors when evaluating the quality of website, this paper proposes a novel system using link analysis. In the proposed system, the indexes are classified into two categories: research index and contrast index. Principal component analysis (PCA) and fuzzy comprehensive evaluation method are used to analyze them, respectively. After analyzing, we can find the subjective factors by computing the consistency of indexes. In the evaluation of websites, we ignore these subjective factors to improve the authority of evaluation results. We conduct a large number of experiments on a group of websites. The experimental results show that the proposed system is effective.

In object-based image retrieval, there are two important issues: an effective image representation method for representing image content and an effective image classification method for processing user feedback to find more images containing the user-desired object categories. In the image representation method, the local-based representation is the best selection for object-based image retrieval. As a kernel-based classification method, Support Vector Machine (SVM) has shown impressive performance on image classification. But SVM cannot work on the local-based representation unless there is an appropriate kernel. To address this problem, some representative kernels are proposed in literatures. However, these kernels cannot work effectively in object-based image retrieval due to ignoring the spatial context and the combination of local features. In this article, we present Adjacent Matrix (AM) and the Local Combined Features (LCF) to incorporate the spatial context and the combination of local features into the kernel. We propose the AM-LCF feature vector to represent image content and the AM-LCF kernel to measure the similarities between AM-LCF feature vectors. According to the detailed analysis, we show that the proposed kernel can overcome the deficiencies of existing kernels. Moreover, we evaluate the proposed kernel through experiments of object-based image retrieval on two public image sets. The experimental results show that the performance of object-based image retrieval can be improved by the proposed kernel.