Continual Test-Time Adaptation (CTTA), which aims to adapt the pre-trained model to ever-evolving target domains, emerges as an important task for vision models. As current vision models appear to be heavily biased towards texture, continuously adapting the model from one domain distribution to another can result in serious catastrophic forgetting. Drawing inspiration from the the encoding characteristics of neuron activation in neural networks, we propose the Mixture-of-Activation-Sparsity-Experts (MoASE) for the CTTA task. Given the distinct reaction of neurons with low and high activation to domain-specific and agnostic features, MoASE decomposes the neural activation into high-activation and low-activation components in each expert with a Spatial Differentiable Dropout (SDD). Based on the decomposition, we devise a Domain-Aware Router (DAR) that utilizes domain information to adaptively weight experts that process the post-SDD sparse activations, and the Activation Sparsity Gate (ASG) that adaptively assigns feature selection thresholds of the SDD for different experts for more precise feature decomposition. Finally, we introduce a Homeostatic-Proximal (HP) loss to maintain update consistency between the teacher and student experts to prevent error accumulation. Extensive experiments substantiate that MoASE achieves state-of-the-art performance in both classification and segmentation tasks.

Vision-Language-Action (VLA) models enable robotic systems to perform embodied tasks but face deployment challenges due to the high computational demands of the dense Large Language Models (LLMs), with existing early-exit-based sparsification methods often overlooking the critical semantic role of final layers in downstream tasks. Aligning with the recent breakthrough of the Shallow Brain Hypothesis (SBH) in neuroscience and the mixture of experts in model sparsification, we conceptualize each LLM layer as an expert and propose a Mixture-of-LayEr Vision Language Action model (MoLe-VLA or simply MoLe) architecture for dynamic LLM layer activation. Specifically, we introduce a Spatial-Temporal Aware Router (STAR) for MoLe to selectively activate only parts of the layers based on the robot’s current state, mimicking the brain's distinct signal pathways specialized for cognition and causal reasoning. Additionally, to compensate for the cognition ability of LLM lost during the layer-skipping, we devise a Cognitive self-Knowledge Distillation (CogKD) to enhance the understanding of task demands and generate task-relevant action sequences by leveraging cognition features. Extensive experiments in RLBench simulations and real-world environments demonstrate the superiority of MoLe-VLA in both efficiency and performance, improving the mean success rate by 9.7% across ten simulation tasks while accelerating inference by 36.8% over OpenVLA.

The computational demands of large-scale AI models raise significant concerns about their carbon footprint. Most carbon accounting methods for large-scale AI models suffer from three key limitations: they overlook embodied carbon (from hardware manufacturing) or model it simplistically, rely on location-based carbon attribution that fails to reflect individual corporate efforts to decarbonize (e.g., via Power Purchase Agreements (PPAs)), and are deterministic, ignoring inherent uncertainties. This paper proposes PUMA, a Probabilistic Uncertainty Market Attribution carbon accounting model for large-scale AI models. PUMA integrates market-based carbon intensity to accurately account for the impact of PPAs and employs probabilistic modeling to capture uncertainties in the carbon accounting for AI models arising from spatiotemporal variations in manufacturing and operation, as well as evolving efficiency. We make an effort to develop a comprehensive carbon dataset by aggregating related data from diverse sources, and then we implement a simple yet effective Kernel Density Estimate (KDE) on the distribution of the parameters from the collected dataset. We compare PUMA with LLMCarbon, the state-of-the-art carbon accounting model for large AI models. The deviation of the accounting result is significant, reaching up to around 201%.

Carbon intensity forecasting is vital for optimizing carbon-aware systems to reduce their carbon footprint. Existing methods focus on data-rich regions, yet most areas lack sufficient carbon data due to carbon intensity's inherent measurement limitations (i.e., reliance on indirect estimation rather than direct sensor metering) and dependence on upstream entities like grid operators for data disclosure. We propose DGCFM, a Dual Graph empowered Carbon-domain Foundation Model, enabling cross-regional general carbon intensity forecasting, especially for data-scarce regions. DGCFM builds on a pre-trained Time Series Foundation Model (TSFM) with strong generalization capabilities to capture temporal patterns under data constraints, empowered by metadata-driven carbon hypergraph fine-tuning and a spatiotemporal graph to capture spatial dependency in carbon networks. Evaluated on real-world datasets, DGCFM achieves 20.04% average accuracy improvement in low-data scenarios.

The growing complexity of Artificial Internet of Things (AIoT) applications, driven by massive data generation and intricate interaction demands, necessitates intelligent decision-making that surpasses basic automation. While traditional machine learning is already integrated into edge devices, the evolving AIoT paradigm still faces severe challenges in complex tasks. Large language models (LLMs) emerge as a promising solution for diverse AIoT tasks, yet their deployment on resource-constrained edge networks faces two major challenges: heterogeneous resource limitations and the complexity of dynamic model selection. To address these issues, we propose an edge computing framework that leverages a compatibility evaluation mechanism and a diffusion-based deep reinforcement learning (DRL) algorithm for dynamic LLM service node selection. This approach enhances deployment efficiency and adaptability in heterogeneous environments. Our diffusion-based DRL algorithm continuously learns from environmental states to make robust, real-time selection decisions, assigning each task to the most suitable LLM instance to maximize long-term task completion quality (TCQ). Experimental results demonstrate that our algorithm significantly outperforms baseline methods in TCQ, convergence speed, and stability. Furthermore, sensitivity and robustness analyses demonstrate the effectiveness of the proposed method under resource fluctuations, varying node scales, and different reward preferences.

The rapid evolution of Large Language Models (LLMs) into multi-modal LLM agents has enabled autonomous systems capable of complex reasoning and action. While deploying these Multi-Agent Systems (MAS) at the edge promises reduced latency and enhanced privacy, it introduces significant challenges due to the heterogeneity of edge resources, the massive computational overhead of multimodal LLMs, and the volatility of multimodal data transmission. Existing service placement strategies often overlook the intricate dependencies within agent workflows and the substantial configuration latency required for model switching. In this paper, we propose a latency-aware algorithm for placing heterogeneous multimodal LLM agents at the edge. We model agent collaboration as a Directed Acyclic Graph (DAG), and then formulate the placement problem to maximize the number of satisfied user requests under strict resource constraints. We introduce a dynamic programming-based algorithm with theoretical performance bounds for single-request optimization, explicitly accounting for dynamic model configuration costs. For multi-request scenarios, we develop an online list-scheduling algorithm leveraging B-level heuristics to maximize request completion rates. Extensive experiments using five state-of-the-art agent systems (e.g., AWorld, OpenManus) across four benchmarks (e.g., GAIA, OSWorld) demonstrate that our approach reduces average latency by up to 42.62% compared to baselines, and therefore significantly improves throughput in high-load environments.

Currently, with the impressive capabilities of large language models (LLMs), various services are provided, including flexible and personalized Embedding-as-a-service (EaaS), however, raising privacy concerns, especially Model Inversion Attacks. To address this, we proposeP4LLM, a novelPrivacy-Preserving framework to protect embedding privacy for EaaS LLM serving via token-wisePartition andPerturbation without model retraining. The core idea is to token-wise partition Transformers on the user side and transmit the privacy-preserved hidden states instead of raw embeddings. We design an Insensitive Tokens Filter that discards non-sensitive tokens, a Partitioning Scheduler that balances user computation resource within service latency, and a Hidden-State Perturbator that injects calibrated noise to meet privacy guarantees. We also provide a formal privacy analysis and guarantee proof for P4LLM. We evaluate P4LLM with four state-of-the-art approaches and show improvements of up to 40.06% in privacy protection capability, 23.21% in accuracy, and a latency reduction of 38.01%.

With the rapid development of edge devices and growing awareness of privacy protection, recent developers have transformed to fine-tune LLMs via federated learning instead of centralized training. Federated fine-tuning can leverage distributed data sources and computation power, but it also suffers from system and statistical heterogeneity. Client selection is an effective tool to solve the system and statistical heterogeneity in FL, but existing client selection schemes that involve online measurement will not be as effective in LLM fine-tuning as in conventional FL due to the huge LLM size and fewer fine-tuning rounds. In this paper, to the best of our knowledge, we are the first to consider both system and statistical heterogeneity in federated LLM fine-tuning, and we formulate a new latency minimization problem. We propose to measure client data overlap via knowledge graph offline to assist client selection in federated LLM fine-tuning. Our client selection scheme excels in both model accuracy and fine-tuning latency. We evaluate our scheme via two LLMs and two applications via four datasets. The experiment results illustrate that our scheme achieves the highest accuracy while 2.05x faster than the baselines.

In real-time edge-cloud video analytics systems, the edge conducts initial analytics on the video frames to a split layer of a trained neural network model. Then, it sends intermediate results to the cloud for follow-up analytics. In this article, we first show that malicious attackers can perform reconstruction attacks and attribute inference attacks on those intermediate results. We present Preva, a new P rivacy-preserving R eal-time E dge-cloud V ideo A nalytics system that defends against both attacks while respecting latency constraints. Preva first applies a lightweight, policy-based video frames transformation scheme generated on the fly by PrevaNetv2, with a lightweight backbone and an early-exit mechanism that adapts the computation resources of users. To guarantee end-to-end latency, a contextual multi-armed-bandit algorithm (BBSplit) dynamically chooses the optimal split layer, and a frame-similarity filter (PPReuse) amortizes policy generation across video frames. We present a formal privacy analysis and show that Preva can guarantee privacy leakage under reconstruction attacks and attribute inference attacks. We evaluate Preva through three video analytics applications and show that Preva outperforms existing systems by 41.6% in analytics accuracy and 52.7% in privacy leakage.

With the popularity of 3D volumetric video applications, such as Autonomous Driving, Virtual Reality, and Mixed Reality, current developers have turned to deep learning for compressing volumetric video frames, i.e., point clouds for video upstreaming. The latest deep learning-based solutions offer higher efficiency, lower distortion, and better hardware support compared to traditional ones like MPEG and JPEG. However, privacy threats arise, especially reconstruction attacks targeting to recover the original input point cloud from the intermediate results. In this paper, we design VVRec, to the best of our knowledge, which is the first targeting DL-based Volumetric Video Reconstruction attack scheme. VVRec demonstrates the ability to reconstruct high-quality point clouds from intercepted transmission intermediate results using four well-trained neural network modules we design. Leveraging the latest latent diffusion models with Gamma distribution and a refinement algorithm, VVRec excels in reconstruction quality and color recovery and surpasses existing defenses. We evaluate VVRec using three volumetric video datasets. The results demonstrate that VVRec achieves 64.70dB reconstruction accuracy, with an impressive 46.39% reduction of distortion over baselines.

Large Language Models (LLMs) have shown great promise in automating data analytics tasks by interpreting natural language queries and generating multi-operation execution plans. However, existing LLM-agent-based analytics frameworks operate under the assumption of centralized data access, offering little to no privacy protection. In contrast, federated analytics (FA) enables privacy-preserving computation across distributed data sources, but lacks support for natural language input and requires structured, machine-readable queries. In this work, we present LAFA, the first system that integrates LLM-agent-based data analytics with FA. LAFA introduces a hierarchical multi-agent architecture that accepts natural language queries and transforms them into optimized, executable FA workflows. A coarse-grained planner first decomposes complex queries into sub-queries, while a fine-grained planner maps each sub-query into a Directed Acyclic Graph of FA operations using prior structural knowledge. To improve execution efficiency, an optimizer agent rewrites and merges multiple DAGs, eliminating redundant operations and minimizing computational and communicational overhead. Our experiments demonstrate that LAFA consistently outperforms baseline prompting strategies by achieving higher execution plan success rates and reducing resource-intensive FA operations by a substantial margin. This work establishes a practical foundation for privacy-preserving, LLM-driven analytics that supports natural language input in the FA setting.

The solar photovoltaic power forecasting (SPPF) of a PV system is vital for the downstream power estimation. While approaches for recent decentralized PV systems require customized models for each PV installation, this method is labor-intensive and not scalable. Therefore, developing a general SPPF model for a decentralized PV system is essential. The primary challenge in developing such a model is accounting for regional weather variations. Recent advancements in weather foundation models (WFMs) offer a promising opportunity, providing accurate forecasts with reduced computational demands. However, integrating WFMs into SPPF models remains challenging due to the complexity of WFMs. This paper introduces a novel approach, spatio-temporal knowledge distillation (STKD), to efficiently adapt WFMs for SPPF. The proposed STKD-PV models leverage regional weather and PV power data to forecast power generation from six hours to a day ahead. Globally evaluated across six datasets, STKD-PV models demonstrate superior performance compared to state-of-the-art (SOTA) time-series models and fine-tuned WFMs, achieving significant improvements in forecasting accuracy. This study marks the first application of knowledge distillation from WFMs to SPPF, offering a scalable and cost-effective solution for decentralized PV systems.

Volumetric videos are essential for immersive applications due to their engaging and realistic experiences. However, streaming them in real time over constrained, fluctuating networks remains challenging. Progressive streaming is an effective method to mitigate this issue by gradually enhancing video quality through incremental data transmission. However, existing progressive volumetric streaming solutions often rely on specific compression algorithms or require codec modifications, leading to poor compatibility with standard codecs. In this paper, we propose P2VS, a progressive partition-based volumetric video streaming framework, to achieve codec-independent progressive streaming. Specifically, P2VS leverages the unique structure of point cloud-based volumetric video to incrementally enhance video quality without being constrained by specific compression algorithms. Moreover, we propose adaptive streaming algorithms under this framework to enhance the quality of experience (QoE). Extensive simulations demonstrate that P2VS improves QoE by 21% on average compared to non-progressive streaming schemes. It also achieves better bandwidth efficiency and full compatibility with standard codecs. A prototype is built to verify the feasibility of P2VS.

The rapid growth of large AI models has raised significant environmental concerns due to their substantial carbon footprint. Existing carbon accounting methods for AI models are fundamentally deterministic and fail to account for inherent uncertainties in embodied and operational carbon emissions. Our work aims to investigate the effect of these uncertainties on embodied and operational carbon footprint estimates for large AI models. We propose a Probabilistic Carbon Accounting Model (PCAM), which quantifies uncertainties in the carbon accounting of large AI models. We develop parameter models to quantify key components (processors, memory, storage) in the carbon footprint of AI models. To characterize the distribution of the parameters, we develop a carbon dataset by aggregating related data from various sources. Then, we generate the probabilistic distribution of the parameters from the collected dataset. We compare the performance of PCAM with LLMCarbon, the state-of-the-art carbon accounting method for large AI models. PCAM achieves $\leq7.44\%$ error compared to LLMCarbon’s $\leq108.51\%$.

3D Gaussian Splatting (3DGS) has emerged as a groundbreaking 3D scene representation technique, offering unprecedented visual quality and rendering efficiency. However, the substantial data volume of 3DGS scenes poses significant challenges for streaming applications. Existing research on 3DGS has primarily focused on compression and rendering efficiency, neglecting the specific requirements of streaming transmission. Moreover, the Spherical Harmonics color representation in 3DGS complicates viewport-based transmission partitioning. Achieving hierarchical Gaussian streaming without noticeable quality degradation also remains a significant challenge.To address these challenges, we propose SRBF-Gaussian, a new paradigm that revolutionizes the traditional 3DGS format. Our approach introduces viewport-dependent color encoding based on Spherical Radial Basis Functions (SRBFs) and HSL color space, enabling selective transmission of viewport-relevant color data. This reduces data transmission while maintaining visual quality. We implement adaptive Gaussian pruning and transmission, optimized for current viewports and network conditions. Additionally, we develop coherent multi-level Gaussian representations for smooth transitions between quality levels. Our system incorporates user-behavior-aware streaming strategies to anticipate and pre-fetch relevant data. In cloud VR scenarios, our approach demonstrates substantial improvements, achieving a 5.63% - 14.17% increase in PSNR, a 7.61% - 59.16% reduction in latency, and a 10.45% - 30.12% improvement in overall Quality of Experience (QoE).

The explosive growth in video services has significantly strained current mobile network infrastructure, leading to spectrum scarcity, backhaul congestion, and degraded quality of experience. While edge caching has emerged as a promising solution to address these challenges and deliver seamless video playback experience, multiversion edge caching for heterogeneous clients remains challenging due to varying client requirements and network conditions. This article proposes a transcoding-enabled edge caching framework for mobile edge-cloud computing networks. Specifically, we combine video transcoding with edge caching to support both “direct cache hits” and “soft cache hits” for reducing transmission latency. We model this joint caching and resource allocation problem as a Stackelberg game to minimize video transmission latency. To solve this problem, we develop a novel dual timescale model agnostic meta-learning (MAML)-based Stackelberg game (DTMSG) optimization approach that determines the delay-optimal Stackelberg equilibrium (SE) and accelerates convergence. Simulation results demonstrate that our DTMSG optimization algorithm efficiently converges to the SE point, maximizing the utility function of the MVNO and BSs while reducing the average video transmission delay.

3-D Gaussian splatting (3DGS) has emerged as a promising technique for high-quality 3-D scene representation. However, streaming 3DGS scenes poses significant challenges due to large data volumes and complex spatial structures, resulting in nonsmooth scene loading, inferior visual quality, and ineffective streaming adaptation, ultimately impacting user experience adversely. To tackle these challenges and enhance user Quality of Experience (QoE), this article introduces a novel adaptive streaming framework called 3DGStreaming. Our framework comprises three key components: 1) smart Spatial Partitioning for efficient scene division, enabling selective streaming and seamless scene merging; 2) two-step Progressive Scene Generation, involving content-aware downsampling and attribute compression to create multibitrate 3DGS scenes; and 3) Field of View (FoV)-based Bitrate Adaptation using a decision transformer for viewport-based bitrate selection. Extensive experiments demonstrate the superiority of 3DGStreaming over existing state-of-the-art solutions. 3DGStreaming achieves a greater rendering quality with a 5.7%–25.5% increase in PSNR, a 27.8%–69.2% reduction in latency, a 54.7% reduction in training time, and a 17.2%–68.3% improvement in overall QoE.

Federated multi-armed bandits extend the multi-armed bandits framework to the federated learning setting where multiple clients in the same exploration space collaboratively identify the optimal arm. While previous studies demonstrated its efficiency like classical federated learning, in this paper, we present the first work that reveals the serious fairness problem in federated multi-armed bandits when clients have heterogeneous and overlapping armsets. The fairness problem happens because clients with different trial requirements should conduct different numbers of trials summing up to a global requirement, but they wish to minimize their exploration effort. To address the novel fairness concern, we formally formulate the fairness-aware trial assignment as a coalitional game. Based on the theoretically derived trial requirements, we devise a Shapley value-based trial assignment mechanism to guarantee fairness. Regardless of the #P-hard complexity when deriving the general Shapley value, we achieve an accurate computation of trial assignment with polynomial complexity by exploiting its unique characteristic. We further carefully control the numbers of trials in each iteration to resolve the communication bottleneck and minimize the wasted trials. Experiment results show that, compared to the naïve federated scheme, our design outperforms with both high fairness metrics and high efficiency in total trials and communication.

Embodied carbonis the carbon emissions in the manufacturing process of products, which dominates the overall carbon footprint in many industries. Existing studies derive the embodied carbon through life cycle analysis (LCA) reports. Current LCA reports only provide the carbon emission of aproduct class, e.g. 28nm CPU, whereas aproduct instancecan be made in various regions and time periods. Carbon emissions depend on the electricity generation process, which has spatial-temporal dynamics. Therefore, the embodied carbon of a product instance can differ from its product class. Additionally, different carbon attribution methods (e.g., location-based and market-based) can affect the carbon emissions of electricity, thus further affecting the embodied carbon of products. In this paper, we present new Spatial-Temporal Embodied Carbon (STEC) accounting models with dual attribution methods. We observe significant differences between STEC and current models, e.g., for 7nm CPU the difference is 13.69%. We further examine the impact of STEC models on existing embodied carbon accounting schemes on computer applications, such as Large Language Model (LLM) training and LLM inference. We observe that using STEC results in much greater differences in the embodied carbon of certain applications as compared to others (e.g., 32.26% vs. 6.35%).

Real-time video analytics demand intensive computing resources, often exceeding device capabilities. Heterogeneous computing resources like CPU and GPU, usually work collaboratively to ensure real-time performance. GPU manage data-intensive computing, while CPU handle instruction-intensive tasks. However, analytics accuracy can degrade because of the imbalance between CPU-GPU workloads and resources, and static resources in most systems limit adaptability to dynamic workloads. In addition to CPU-GPU resource control, accuracy is highly dependent on video analytics configuration including resolution, frame rate, and model selection. In this paper, we propose Gemini+, a hardware-accelerated video analytics pipeline empowered by dual-image FPGAs. It enables flexible CPU-GPU resource adaptation by providing near-instantaneous switching between two pre-configured images. We investigate CPU-GPU resource control and video analytics configuration adaptation in a dual-image FPGA-based video analytics pipeline. We study and formulate two problems of practical importance, a single-camera and multi-camera dual computing resource control problem, i.e., SC-DCRC and MC-DCRC. We analyze the problem complexity and develop optimal and sub-optimal algorithms. We evaluate our algorithms through simulation and build a prototype for verification. The results show that Gemini+ can improve analytic accuracy by 25% compared to fixed CPU-GPU resource systems and 88% compared to GPU-dominant systems.

Vision-centric Bird’s Eye View (BEV) perception holds considerable promise for autonomous driving. Recent studies have prioritized efficiency or accuracy enhancements, yet the issue of domain shift has been overlooked, leading to substantial performance degradation upon transfer. We identify major domain gaps in real-world cross-domain scenarios and initiate the first effort to address the Domain Adaptation (DA) challenge in multi-view 3D object detection for BEV perception. Given the complexity of BEV perception approaches with their multiple components, domain shift accumulation across multi-geometric spaces (e.g., 2D, 3D Voxel, BEV) poses a significant challenge for BEV domain adaptation. In this paper, we introduce an innovative geometric-aware teacher-student framework, BEVUDA++, to diminish this issue, comprising a Reliable Depth Teacher (RDT) and a Geometric Consistent Student (GCS) model. Specifically, RDT effectively blends target LiDAR with dependable depth predictions to generate depth-aware information based on uncertainty estimation, enhancing the extraction of Voxel and BEV features that are essential for understanding the target domain. To collaboratively reduce the domain shift, GCS maps features from multiple spaces into a unified geometric embedding space, thereby narrowing the gap in data distribution between the two domains. Additionally, we introduce a novel Uncertainty-guided Exponential Moving Average (UEMA) to further reduce error accumulation due to domain shifts informed by previously obtained uncertainty guidance. To demonstrate the superiority of our proposed method, we execute comprehensive experiments in four cross-domain scenarios, securing state-of-the-art performance in BEV 3D object detection tasks, e.g., 12.9% NDS and 9.5% mAP enhancement on Day-Night adaptation.

This paper proposes ElastiCast, a novel Bluetooth Low Energy (BLE) broadcast mode that reduces the neighbor discovery latency in offline finding networks (OFNs). ElastiCast adapts the broadcast mode of the lost devices to the scan modes of the finder devices, considering their diversity. We start with an overview of OFNs, followed by a detailed analysis of the issues and challenges of existing solutions, which motivates the design of ElastiCast. Then we provide Blender, a simulator that models the neighbor discovery behavior of different broadcasters and scanners. By adopting Blender, ElastiCast can be implemented with three components: Local Optima Estimation, Common Interest Extraction, and Interval Multiplexing, in which we capture the key features of BLE neighbor discovery and globally optimize the broadcast mode interacting with diverse scan modes. Experimental evaluation results and commercial product deployment experience demonstrate that ElastiCast is effective in achieving stable and bounded neighbor discovery latency within the power budget.

Federated Learning (FL) enables edge devices to collaboratively train a global model using local data. However, the increasing prevalence of watermarks in datasets presents a new challenge to efficient FL. While watermarks assert data ownership and copyright, they introduce complexities that can lead to shortcut learning problems and mislead utility measurements for client selection. These issues are further exacerbated by batch size variations in efficient FL frameworks, ultimately undermining their time-to-accuracy performance. We introduceLotusFL, an FL system designed to address the challenges posed by watermarked datasets in efficient FL. Specifically, it tackles the increased time-to-accuracy due to erroneous client selection and the accuracy degradation observed with larger batch sizes.LotusFLfirst estimates the characteristics of watermarks through statistical estimation and then adjusts the batch size using this estimated watermark information to balance the negative impact of the watermark against device idle waiting time. Additionally, its client selection mechanism, based on historical information, avoids the misleading utility signals from watermarks. This mechanism, working in conjunction with batch size adjustment, aims to accurately predict device runtime and identify potentially valuable devices. We evaluatedLotusFLthrough a real-world deployment on 40 edge devices. Compared to state-of-the-art efficient FL frameworks,LotusFLachieves superior performance, enhancing accuracy by up to 8.2% and reducing training time by 1.97×.

In the Computer Vision, the Large Vision Models (LVM) based on Vision Transformer (ViT) achieve advanced performance on general complex visual tasks. However, deploying these resource-intensive models on edge devices with limited computational power and memory is challenging, especially for those mobile edge devices. Existing model compression works downgrade the prediction accuracy and fail to adapt to dynamic network bandwidth or hardware changes. Besides, the split inference for typical Deep Neural Networks (DNN) is inefficient for LVM’s large intermediate result size. To address the computation and bandwidth limitation of edge devices of LVM, we design a new split inference acceleration LVMScissor by leveraging model parallelism and meta-heuristic algorithm. We first implement an inter-layer ViT parallelism strategy. After modeling the parallelized ViT split problem into a Multi-task three-processor scheduling (MTS) problem, we propose LVM-MSSA, a scheduling algorithm based on a famous meta-heuristic algorithm, the Multi-objective Salp Swarm Algorithm (MSSA) to schedule model parallelism and split strategy. The evaluation results show that compared to state-of-the-art inference acceleration approaches, our solution is faster from$1.6\times$to$6.92\times$under variant LVMs, edge devices, datasets, and network traces.

Multimodal learning has significantly advanced the extraction of features from varied data sources, enhancing model performance. Federated learning (FL) complements this by enabling collaborative training while maintaining data privacy. The fusion of these two fields, multimodal federated learning, offers considerable promise. Yet, standard methods often incorrectly assume that each node in the FL network has a full complement of multimodal data, which is rare in real-world applications. In our study, we present a novel architecture designed to surmount these challenges, termed the Unimodal Training - Multimodal Prediction (UTMP) framework, positioned within the multimodal federated learning paradigm. Our proposed model, the HA-Fedformer, is a transformer-based model crafted to facilitate unimodal training on the client-side using exclusively unimodal datasets and to execute multimodal inference by synthesizing insights from multiple clients. Our HA-Fedformer model effectively handles non-IID data through a novel uncertainty-aware aggregation technique and layer-wise Markov Chain Monte Carlo sampling in local encoders. It also resolves misaligned language sequences via cross-modal decoder aggregation, capturing correlations between decoders trained on different modalities. Our comprehensive evaluations conducted on widely recognized sentiment analysis benchmarks demonstrate the superiority of the HA-Fedformer. The results show that our model achieves a substantial uplift in performance.

Recently, content-aware methods have been employed to reduce bandwidth and enhance the quality of Internet video delivery. These methods involve training distinct content-aware super-resolution (SR) models for each video chunk on the server, subsequently streaming the low-resolution (LR) video chunks with the SR models to the client. Prior research has incorporated additional partial parameters to customize the models for individual video chunks. However, this leads to parameter accumulation and can fail to adapt appropriately as video lengths increase, resulting in increased delivery costs and reduced performance. In this paper, we introduce RepCaM++, an innovative framework based on a novel Re- parameterization Content-aware Modulation (RepCaM) module that uniformly modulates video chunks. The RepCaM framework integrates extra parallel-cascade parameters during training to accommodate multiple chunks, subsequently eliminating these additional parameters through re- parameterization during inference. Furthermore, to enhance RepCaM's performance, we propose the Transparent Visual Prompt (TVP), which includes a minimal set of zero-initialized image-level parameters (e.g., less than 0.1%) to capture fine details within video chunks. We conduct extensive experiments on the VSD4K dataset, encompassing six different video scenes, and achieve state-of-the-art results in video restoration quality and delivery bandwidth compression.

With the widespread deployment of Low Earth Orbit (LEO) satellites, they generate a vast amount of data. This data has been instrumental in supporting machine learning (ML) in various terrestrial services to address global challenges such as monitoring climate change and natural disasters. However, many national regulations restrict the direct transmission of satellite data to ground stations (GSs). Therefore, ground-assisted satellite federated learning (FL) has emerged as a paradigm to safeguard data privacy by transferring model parameters instead of raw data for collaborative training. At present, the existing groundassisted satellite FL methods encounter practical challenges: 1) The dynamic environment of LEO satellites results in continuous changes in the types of data collected by satellites, making it difficult for traditional FL models to adapt to these changes. This can lead to a deterioration in model accuracy over extended periods of model training. 2) Communication between satellites and GS is affected by atmospheric interference and weather factors, resulting in increased transmission delays and affecting the realtime efficiency of the FL system. In response to these challenges, we propose a dynamic, efficient, and distributed ground-assisted LEO satellite federated learning (DEDFL) framework to improve model accuracy and reduce satellite communication delays. In DEDFL, we design a Balanced Class Memory Extraction and an information playback strategy that enables the onboard FL model to adapt to changing satellite data types, thus achieving a performance balance across different classes. Additionally, we propose an adaptive fine coding method for parameter adoption prior to satellite transmission, effectively reducing the delay caused by satellites and ground-specific environmental variations. Experimental results demonstrate that the DEDFL method offers better accuracy and communication efficiency than other baseline algorithms.

Due to its advantage of protecting data privacy and reducing communication overhead, Federated Learning (FL) is becoming a promising machine learning paradigm. However, resource limitations and unstable communication connections on the participating client end can lead to unintentional failures that degrade FL performance. Moreover, as FL systems scale and interconnect increasingly, they face growing exposure to intentional network risks. Furthermore, the assumption of continued trust in historically benign clients introduces vulnerabilities to potential internal betrayal within FL systems. In this paper, we enhance the robustness of FL by incorporating the zero-trust principle, which eliminates implicit trust in clients and mitigates unintentional failures, intentional attacks, and strategic betrayal risks. The framework incorporates dynamic client selection and aggregation weight allocation through trustworthiness evaluation and sustained skepticism toward each potential betrayal behavior. Specifically, a Dirichlet-based trust evaluation technique is presented to update clients' trustworthiness with evolving observations. Then, to reduce potential betrayal loss, we formulate a min-max optimization problem that minimizes the worst-case betrayal loss. Next, we transform the formulation into a convex programming problem for solution. Extensive simulations are conducted to demonstrate the efficacy of the zero-trust based FL in the accurate trust assessment and the system's betrayal-aware robustness enhancement.

The rise of the metaverse has driven the rapid development of various applications, such as Virtual Reality (VR) and Augmented Reality (AR). As a form of multimedia in the metaverse, VR video streaming (a.k.a., VR spherical video streaming and 360$^{\circ }$video streaming) can provide users with a 360$^{\circ }$immersive experience. Generally, transmitting VR video requires far more bandwidth than regular videos, which greatly strains existing network transmission. Predicting and selectively streaming VR video in the users' viewports in advance can reduce bandwidth consumption and system latency. However, existing methods either consider only historical viewport-based prediction methods or predict viewports by correlations between visual features of video frames, making it hard to adapt to the dynamics of users and video content. In the meantime, spurious correlations between visual features lead to inaccurate and unreliable prediction results. Hence, we propose an unsupervised multiscale causal representation learning (UMCRL)-based method to predict viewports in VR video streaming, including user preference-based and video content-based viewport prediction models. The former is designed by a position predictor to predict the future users' viewports based on their historical viewports in multiple video frames to adapt to users' dynamic preferences. The latter achieves unsupervised multiscale causal representation learning through an asymmetric causal regressor, used to infer the causalities between local and global-local visual features in video frames, thereby helping the model understand the contextual information in the videos. We embed the causalities in the transformer decoder via causal self-attention for predicting the users' viewports, adapting to the dynamic changes of video content. Finally, combining the results of the two aforementioned models yields the final prediction of the users' viewports. In addition, the QoE of users is satisfied by assigning different bitrates to the tiles in the viewport through a pyramid-based bitrate allocation. The experimental results verify the effectiveness of the method.

Federated learning (FL) addresses privacy concerns by training models without sharing raw data, overcoming the limitations of traditional machine learning paradigms. However, the rise of smart applications has accentuated the heterogeneity in data and devices, which presents significant challenges for FL. In particular, data skewness among participants can compromise model accuracy, while diverse device capabilities lead to aggregation bottlenecks, causing severe model congestion. In this article, we introduce Spread+, a hierarchical system that enhances FL by organizing clients into clusters and delegating model aggregation to edge devices, thus mitigating these challenges. Spread+ leverages hedonic coalition formation game to optimize customer organization and adaptive algorithms to regulate aggregation intervals within and across clusters. Moreover, it refines the aggregation algorithm to boost model accuracy. Our experiments demonstrate that Spread+ significantly alleviates the central aggregation bottleneck and surpasses mainstream benchmarks, achieving performance improvements of 49.58% over FAVG and 22.78% over Ring-allreduce.

Nowadays, Federated Learning (FL) has emerged as a promising and critical machine learning scheme to protect data privacy and reduce communication overhead. As the scale and connectivity expand in the FL system, enhancing the model’s robustness against security threats from malicious clients grows ever more critical. An effective defensive solution involves selecting benign clients appropriately, thereby mitigating the vulnerability of the FL system to malicious attacks. However, clients exhibit varying behaviors over time, which complicates the task of accurately modeling their future trustworthiness. Moreover, blindly trusting clients with high trust values poses risks, given the potential for severe losses from betrayal. To tackle these problems, we propose a zero-trust policy in FL aimed at establishing continuous trust in each client while maintaining skepticism towards potential betrayal attacks. Specifically, we develop a Dirichlet-based trust evaluation technique to enable a comprehensive selection of trustworthy participants. This technique leverages the posterior distribution to estimate clients’ trust values from their evolving behavior records over time. Then, we anticipate potential betrayal from a selected client and formulate a min-max optimization problem to minimize the worst-case betrayal loss, thereby boosting the system’s betrayalaware robustness. Next, we convert this problem into a convex optimization problem and utilize the interior point method for resolution. We conduct extensive simulations to validate the efficacy of our proposed zero-trust policy in accurately assessing trust and enhancing the model’s robustness to betrayal.

The emerging Web 3.0 paradigm aims to decentralize existing web services, enabling desirable properties such as transparency, incentives, and privacy preservation. However, current Web 3.0 applications supported by blockchain infrastructure still cannot support complex data analytics tasks in a scalable and privacy-preserving way. This paper introduces the emerging federated analytics (FA) paradigm into the realm of Web 3.0 services, enabling data to stay local while still contributing to complex web analytics tasks in a privacy-preserving way. We propose FedWeb, a tailored FA design for important frequent pattern mining tasks in Web 3.0. FedWeb remarkably reduces the number of required participating data owners to support privacy-preserving Web 3.0 data analytics based on a novel distributed differential privacy technique. The correctness of mining results is guaranteed by a theoretically rigid candidate filtering scheme based on Hoeffding’s inequality and Chebychev’s inequality. Two response budget saving solutions are proposed to further reduce participating data owners. Experiments on three representative Web 3.0 scenarios show that FedWeb can improve data utility by ∼25.3% and reduce the participating data owners by ∼98.4%.

Federated learning is a promising privacy-preserving learning paradigm in which multiple clients can collaboratively learn a model with their image data kept local. For protecting data ownership, personalized watermarks are usually added to the image data by each client. However, the introduced watermarks can lead to a shortcut learning problem, where the learned model performs predictions over-rely on the simple watermark-related features and represents a low accuracy on real-world data. Existing works assume the central server can directly access the predefined shortcut features during the training process. However, these may fail in the federated learning setting as the shortcut features of the heterogeneous watermarked data are difficult to obtain. In this paper, we propose a federated Morozov regularization technique, where the regularization parameter can be adaptively determined based on the watermark knowledge of all the clients in a privacy-preserving way, to eliminate the shortcut learning problem caused by the watermarked data. Specifically, federated Morozov regularization firstly performs lightweight local watermark mask estimation in each client to obtain the locations and intensities knowledge of local watermarks. Then, it aggregates the estimated local watermark masks to generate the global watermark knowledge with a weighted averaging. Finally, federated Morozov regularization determines the regularization parameter for each client by combining the local and global watermark knowledge. With the regularization parameter determined, the model is trained as normal federated learning. We implement and evaluate federated Morozov regularization based on a real-world deployment of federated learning on 40 Jetson devices with real-world datasets. The results show that federated Morozov regularization improves model accuracy by 11.22% compared to existing baselines.

As a result of its low search cost, Differentiable Architecture Search (DARTS) has recently received a lot of interest. Nowadays, most methods based on DARTS only focus on improving a single indicator (e.g., accuracy), making the search process more inclined to complex networks with more robust representational capabilities. Hence, the architectures searched by these methods tend to own high latency, leading to DARTS in some low-latency scenarios or edges with limited computing power, and on-device deployment becomes difficult. To deal with this challenge, we propose a Hardware Latency-Aware Differentiable Search (HL-DARTS) algorithm. This algorithm designs a multi-layer regression network that uses the soft attention mechanism to predict the latency on the corresponding hardware devices, thus adding a differentiable latency loss term based on the DARTS algorithm. We further propose an adaptive constraint amplitude—a mechanism for balancing accuracy and latency while searching for a latency-friendly architecture for a given hardware device. We conduct ablation experiments on different datasets and different hardware devices. The experimental results show that HL-DARTS can find the ideal architecture for different hardware devices and that this architecture is also broadly applicable to various datasets.

Federated Learning (FL), as a distributed machine learning approach, makes it possible to realize global training and knowledge sharing of models while protecting data privacy. However, existing FL methods still face some challenges. First, since the local client model has only a limited parameter capacity when the model learns a new task or new data, previously learned information can be lost quickly and in large amounts, leading to catastrophic forgetting. Second, the private information may still be leaked by analyzing the parameters uploaded by the clients. Besides, the data distributions of FL clients are usually non-independent and identically distributed, leading to a degradation of the quality of the global aggregation model. These challenges pose obstacles to the effective implementation of federated learning. To address the above issues, we propose a novel differentially private federated class-incremental learning (DPFCIL) method. Specifically, it includes three aspects: 1) We design an elastic memory strategy that mitigates the class incremental model’s forgetting of old knowledge by dynamically expanding new modules and applying network regularization at the local client. 2) Before uploading parameters to the local client, we introduce differential privacy to prevent information leakage effectively. 3) We propose to use a dynamic adaptive aggregation method that relies on model parameter differences in the server aggregation phase to mitigate the loss of global accuracy due to the heterogeneity of the client’s data. Extensive experiments are designed to compare with state-of-the-art baseline methods on several datasets, and the experimental results show that DPFCIL outperforms other methods by an average of at least 5.6% in accuracy.

In recent years, volumetric videos have brought immersive experiences to users. Existing viewport-based volumetric video streaming (VVS) systems prune the point cloud according to visibility to reduce bandwidth consumption, leading to a better responsiveness. They also predict bandwidth and allocate bitrate to different parts of the video to enhance Quality-of-Experience (QoE). However, such designs sometimes result in drastic quality fluctuations in real-world deployment, due to limited generalization performance. Our measurement notes that these systems tend to have a significant accuracy loss under an unseen Out-of-Distribution (OoD) environments. On the other hand, open world prediction/adaptation problem have been addressed in the recent reinforcement learning advances, particularly through prompt-based few-shot and zero-shot learning. Inspired by this development, in this work, we first reformulate the volumetric bitrate adaptation (volumetric ABR) into a sequence prediction problem, then we design a volumetric causal transformer algorithm to solve it. We train our model on a large action trajectory data set, then evaluate it on various OoD scenarios. The result show that FewVV consistently outperforms the existing systems on both performance and generalization.

Volumetric video is becoming a popular application among various multimedia services, which is envisioned as a fundamental technology for VR, AR, and the emerging metaverse. The commodity RGB-D cameras provide an affordable solution for volumetric video capture, and the VR headset allows immersive and interactive display. From the networking perspective, the primary challenge lies in the smooth and high-quality transmission over the Internet, given the enormous data volume and limited bandwidth. MPEG V-PCC standard has stood out recently as a promising compression and streaming solution that can effectively reduce video size while maintaining high-visual quality. Since MPEG V-PCC is largely backward compatible with the 2-D video compression standard like H.264/AVC, it is natural to use DASH, the most widely used 2-D streaming framework, to stream it. We first propose an integrated framework based on DASH to support MPEG V-PCC Internet streaming. During this, we faced three challenges. First, the lack of a rate-distortion model for MPEG V-PCC encoding parameters. Second, the need for a new bitrate adaptation controller that not only considers the rate of the chunks but also chooses the chunks with proper frame rate. We align the decision transformer to this problem, which expands the success of transformer-based models in natural language processing to the decision problems. Finally, to solve the stalling time issue inherited from DASH, we use a frame-dropping mechanism to eliminate the stalling in DASH playback. Our evaluations show that VSAS achieves an average acrlong QoE improvement of$1.67\times $over a range of network conditions.

With the increasing pervasiveness of smart earphones, it is appealing to propose more unobtrusive and convenient wearable authentication methods. Researchers have designed earphone-based authentication systems which utilize high-frequency audio signals to scan ear canal structure. Nevertheless, they possess shortcomings of low unobtrusiveness and robustness. In this article, we put forward an earphone-based passive authentication system which makes use of physiological and behavioral acoustic signals caused by a user’s natural actions, including putting on earphones and inner organs’ activities, respectively. By introducing attention mechanism into the network design, our method adaptively weighs two channel signals, and extracts stable fingerprints for different people, which relieves model retraining for unseen users and improves its scalability. We have built a real-time prototype called EarPrint by designing the earphones and a mobile application, and conducted comprehensive experiments under diverse settings. Experimental results demonstrate that EarPrint has low false acceptance rate (FAR) and equal error rate (EER) less than 1% and 5% in most cases, respectively.

In this paper, we reference the previous conference version and complete the grant number mentioned in the acknowledgments of the conference version.

The incompleteness of knowledge graphs (KGs) negatively impacts the performance of KGs in downstream applications (e.g., recommendation systems and information retrieval). This phenomenon has brought an increasing rise in research related to knowledge graph reasoning. Recently, emerged reinforcement learning (RL)-based multi-hop reasoning methods can infer missing information through multi-hop reasoning according to the existing information in KGs, which has better reasoning performance and interpretability. However, these methods always use relation-entity pairs that have been pre-cropped as the action space of agents for path reasoning, which leads to two problems: 1) insufficient learning and reasoning ability of reasoning models and 2) the hard convergence of the training process of agents. To address these problems, we propose aHierarchicalSequenceMultiHop (HSMH) reasoning framework, which consists of the interactive search reasoning model, local-global knowledge fusion mechanism, and action optimization mechanism. We use interactive search reasoning models to select relations and entities independently, thus fully mining the semantic information of relations and entities and improving the learning and reasoning ability of reasoning models. In the HSMH framework, we design the local-global knowledge fusion and action optimization mechanisms for path reasoning, which can enhance agents' state information and action space. Specifically, the local-global knowledge fusion mechanism is designed to acquire the local knowledge of entities and neighboring relations and the global knowledge about KG structure. This local-global knowledge can improve the learning ability of reasoning models. In addition, the action optimization mechanism can combine the filtered action space and the additional action space for efficient path reasoning for agents. Experimental results on five benchmark datasets show that our proposed HSMH framework comprehensively outperforms the state-of-the-art multi-hop reasoning model.

Network traffic classifiers of mobile devices are widely learned with federated learning(FL) for privacy preservation. Noisy labels commonly occur in each device and deteriorate the accuracy of the learned network traffic classifier. Existing noise elimination approaches attempt to solve this by detecting and removing noisy labeled data before training. However, they may lead to poor performance of the learned classifier, as the remaining traffic data in each device is few after noise removal. Motivated by the observation that the data feature of the noisy labeled traffic data is clean and the underlying true distribution of the noisy labeled data is statistically close to the clean traffic data, we propose to utilize the noisy labeled data by normalizing it to be close to the clean traffic data distribution. Specifically, we first formulate a distributionally robust federated network traffic classifier learning problem (DR-NTC) to jointly take the normalized traffic data and clean data into training. Then we specify the normalization function under Wasserstein distance to transform the noisy labeled traffic data into a certified robust region around the clean data distribution, and we reformulate the DR-NTC problem into an equivalent DR-NTC-W problem. Finally, we design a robust federated network traffic classifier learning algorithm, RFNTC, to solve the DR-NTC-W problem. Theoretical analysis shows the robustness guarantee of RFNTC. We evaluate the algorithm by training classifiers on a real-world dataset. Our experimental results show that RFNTC significantly improves the accuracy of the learned classifier by up to 1.05 times.

High Definition (HD) maps have become core supporting components for autonomous driving. To date, their updates heavily depend on the vehicle fleets of the map vendors, which cannot scale and timely reflect the highly dynamic environment. To ensure the HD map quality, it is advocated social vehicles should be used. Nevertheless, there are privacy concerns and a lack of incentives for social vehicles to contribute data. In this paper, we leverage federated analytics (FA), a newly developed collaborative data analytics paradigm, where raw data are kept local and only the insights generated from local analytics are sent to a server for aggregation. We present a new Federated Analytics based HD map Updating model (FAUMap) to protect the privacy of social vehicles. To motivate social vehicles to contribute data and improve the HD map quality, we formulate an overlapping coalition formation game, OCFUMap, and develop an algorithm to find feasible coalitions. Simulations show that our approach can improve the quality of the updated HD map by 1.56 times. To study an end-to-end operation of the FAUMap model and OCFUMap game, we present a case of HD map updates of the Powell street in San Francisco using the autonomous driving simulator CarLA.

Realtime video streaming (RVS) services are gaining popularity in various applications such as video conferencing, online education, and mixed reality. However, adverse network conditions can significantly damage video transmission, leading to a decline in users' Quality of Experience (QoE). Existing approaches have made considerable efforts to address these problems, including bitrate adaptation, FEC (forward error correction) encoding, and super-resolution techniques. Nevertheless, these methods either focus solely on adjusting transmission configurations (ABR) or consume additional network and computational resources to enhance QoE (FEC or super-resolution), making them suboptimal for adverse network conditions. In this paper, we analyze the limitations of conventional RVS systems when confronted with adverse network conditions and proposeTrimStream, a novel RVS solution based on intelligent frame retrospection, to effectively handle such scenarios. Our approach leverages the high similarity observed between frames in realtime video streaming. The core idea is to store a subset of correctly received frames and exploit frame similarity to minimize transmission while breaking down frame-level dependencies. We formulate the frame caching problem to maximize QoE in RVS and present an online frame cache algorithm. Furthermore, we design a vision-transformer-based, cost-effective frame matching framework that combines different levels of frame information. Our evaluation results demonstrate thatTrimStreamoutperforms state-of-the-art solutions by$14.8\% \sim 21.1\%$improvement in overall QoE.

The widespread availability of convenient wireless network connection and video capture have fueled the development of multi-party realtime video streaming (MRVS) services, such as Zoom or Microsoft Teams. These services have transformed the generation and distribution of realtime streaming content and offer a new way of online communication, striving to provide high Quality-of-Experience (QoE) for individuals. However, delivering high QoE in MRVS is more challenging than in traditional video scenarios due to the stringent delay requirements and complex multi-party interactive architectures. In this paper, we propose DSJA, a distributed server-driven multi-party realtime video streaming framework that conquers the challenges. We first design an appropriate QoE model for MRVS services to capture the interplay among perceptual quality, variations, bitrate mismatch, loss damage, and streaming delay. We then model the QoE maximization problem in MRVS as a route scheduling and streaming adaptation problem. Afterward, we design DSJA which seamlessly integrates multiple selective forwarding units (SFU) architecture and server-driven approaches based on a two-step solution of route scheduling and streaming adaptation. DSJA first determines the most suitable SFU and streaming routes for each video session based on SFUs' job queuing delay and path latency. Then, the server conducts joint loss and bitrate adaptation decisions to optimize the streaming configuration of all clients, considering network conditions and QoE preferences. Our evaluations show that our framework outperforms state-of-the-art solutions by$23.1\% \sim 41.7\%$from the perspective of QoE, and reduces the backbone network transmission by$14.0\% \sim 36.6\%$.

The outbreak of COVID-19 has dramatically promoted the explosive proliferation of multi-party realtime video streaming (MRVS) services, represented by Zoom and Microsoft Teams. Different from Video-on-Demand (VoD) or live streaming, MRVS enables all-to-all realtime video communication, bringing significant challenges to service providing. First, unreliable network transmission can cause network loss, resulting in delay increase and visual quality degradation. Second, the transformation from two-party to multi-party communication makes resource scheduling much more difficult. Moreover, optimizing the overall QoE requires a global coordination, which is quite challenging given the various impact factors such as bitrate and loss.In this paper, we propose the SJA framework, which is, to our best knowledge, the first server-driven joint loss and bitrate adaptation framework in multi-party realtime video streaming services towards maximized QoE. We comprehensively design an appropriate QoE model for MRVS services to capture the interplay among perceptual quality, variations, bitrate mismatch, loss damage, and streaming delay. We mathematically formulate the QoE maximization problem in MRVS services. A Lyapunov-based relaxation and the SJA algorithm are further designed to address the optimization problem with close-to-optimal performance. Evaluations show that our framework can outperform the SOTA solutions by 18.4% ∼ 46.5%.

With the increasing popularity of 3D volumetric video applications, e.g., metaverse, AR/VR, etc., there is a growing need to protect users' privacy while sharing their experiences during streaming. In this paper, we show that the existing privacy-preserving approaches for dense point clouds suffer a massive computation cost and degrade the quality of the streaming experience. We design Pagoda, a new PrivAcy-preservinG VOlumetric ViDeo StreAming incorporating the MPEG V-PCC standard, which protects different domain privacy information of dense point cloud, and maintains high throughput. The core idea is to content-aware transform the privacy attribute information to the geometry domain and content-agnostic protect the geometry information by adding Poisson noise perturbations. These perturbations can be denoised through a Poisson diffusion probabilistic model we design to deploy on the cloud. Users only need to encrypt a small amount of high-sensitive information and achieve secure streaming. Our designs ensure the dense point clouds can be transmitted in high quality and the attackers cannot reconstruct the original one. We evaluate Pagoda using three volumetric video datasets. The results show that Pagoda outperforms existing privacy-preserving baselines for 75.6% protection capability improvement, 4.27 times streaming quality, and 26 times latency reduction.

Device-to-device (D2D) caching is becoming prevalent in relieving network congestion. However, there remain challenges in exploring efficient D2D caching strategies due to the diverse user requirements. In this article, we propose a social-aware D2D caching scheme that integrates the concept of social incentive and recommendation with D2D caching decision making. First, we investigate federated learning (FL)-based prediction method to achieve the social-aware in a privacy-preserving manner. Then, the predicted social relationship provides prior knowledge for deep reinforcement learning (DRL) to make optimal D2D caching decisions. The optimization problem of this article is to maximize the data offloading probability, which can be formulated as a Markov decision process. To solve it, we propose a double deep$Q$-learning network (DDQN)-based D2D caching algorithm. Finally, simulation results validate the prediction and convergence performance of the proposed scheme. Besides, the scheme also shows superior caching performance in reducing the average delay and improving overall offloading probability.

Network traffic classification is the foundation for many network security and network management applications. Recently, to preserve the privacy of the data which are generated in the mobile ends, federated learning (FL)-based classification methods are being proposed. Unfortunately, the performance of FL-based methods can seriously degrade when the client data have skewness. This is particularly true for mobile network traffic classification where the environments in the mobile ends are highly heterogeneous. In this article, we first conduct a measurement study on traffic classification accuracy through FL using real-world network traffic trace and we observe serious accuracy degradation due to heterogeneous environments. We propose a novel federated analytics (FA) approach, FEAT, to improve the accuracy. Note that FL emphasizes on model training, yet our FA performs local analytic tasks that can estimate traffic data skewness and select appropriate clients for FL model training. Our analytics tasks are performed locally and in a federated manner; thus, we preserve privacy as well. Our approach has strong theoretical properties where we exploit Hoeffding inequality to infer traffic data skewness and we leverage the Thompson Sampling for client selection. We evaluate our approach through extensive experiments using real-world traffic data sets QUIC and ISCX. The extensive experiments demonstrate that FEAT can improve traffic classification accuracy in heterogeneous environments.

With the growing demand for latency-sensitive and compute-intensive services in the Internet of Things (IoT), multiaccess edge computing (MEC)-enabled IoT is envisioned as a promising technique that allows network nodes to have computing and caching capabilities. In this article, we propose a cache-aided MEC (CA-MEC) offloading framework for joint optimization of communication, computing, and caching (3C) resources in the MEC-enabled IoT. Our goal is to optimize the offloading decision and resource allocation strategy to minimize the system latency subject to dynamic cache capacities and computing resource constraints. We first formulate this optimization problem as a multiagent decision problem, a partially observable Markov decision process (POMDP). Then, the deep graph convolution reinforcement learning (DGRL) method is applied to motivate the agents to learn optimal strategies cooperatively in a highly dynamic environment. Simulations show that our method is highly effective for computation offloading and resource allocation and performs superior results in a large-scale network.

Nowadays, emerging sixth-generation (6G) mobile networks, the Internet of Things (IoT), and mobile-edge computing (MEC) technologies have played significant roles in developing a sustainable computing network. In sustainable computing networks, with the increasing scale of data-driven applications, massive privacy-sensitive data are generated. How to effectively process such data on resource-limited IoT devices is challenging. Although edge intelligence (EI) is designed to maintain an appropriate level of ultradelay reliability, low-latency communication (URLLC), real-time data processing, and security and privacy are concerning. In this article, we propose a novel blockchain-supported hierarchical digital twin IoT (HDTIoT) framework, which combines the digital twin to edge network and adopts blockchain technology to achieve secure and reliable real-time computation. We first propose a data and knowledge dual-driven learning solution to ensure real-time interaction and efficient optimization between the physical and the digital worlds. To improve communication and computation efficiency with data and knowledge dual-driven learning, the optimization goal is to minimize the system delay and energy consumption and ensure system reliability and the learning accuracy of IoT devices. Moreover, we propose a proximal policy optimization (PPO)-based multiagent reinforcement learning (MARL) algorithm to solve the resource allocation (RA) problem. Experimental results show that the proposed RA scheme can improve the efficiency of the HDTIoT system, guarantee learning accuracy, reliability, and security, and make a balance between system delay and energy consumption.

Video analytics systems conduct video preprocessing to filter out unnecessary frames and model inference using appropriately selected neural networks for high analytics speed. Video preprocessing is instruction-intensive computing (IIC) executed by CPU, and model inference is data-intensive computing (DIC) executed by GPU. In this paper, we show the analytics accuracy of existing systems can largely vary in fields, caused by thedynamicIIC and DIC workloads of differentcontentsin applications. Unfortunately, cameras havefixedCPU/GPU resources and cannot effectively adapt to workload dynamics. We develop Gemini, a new edge-side real-time video analytics system enhanced by a dual-image FPGA. We take the advantage of negligible image switching time of dual-image FPGAs, pre-configure one CPU image and one GPU image and elastically multiplex the dual CPU-GPU resources intimedimension. Gemini requires both hardware and software revisions. In hardware, we overcome challenges of hardware-dependent application development, low communication efficiency between the microprocessor and FPGA, and high programming complexity by hardware abstraction, asynchronous data transfer mechanism and stub-skeleton middleware. In software, we overcome the challenge of adapting to the dynamic workloads by a bandit learning approach. We implement Gemini and show that Gemini can improve the analytics accuracy to 90.35%.

Recent advances in deep neural networks have substantially improved the accuracy and speed of various intelligent applications. Nevertheless, one obstacle is that DNN inference imposes a heavy computation burden on end devices, but offloading inference tasks to the cloud causes a large volume of data transmission. Motivated by the fact that the data size of some intermediate DNN layers is significantly smaller than that of raw input data, we designed the DNN surgery, which allows partitioned DNN to be processed at both the edge and cloud while limiting the data transmission. The challenge is twofold: (1) Network dynamics substantially influence the performance of DNN partition, and (2) State-of-the-art DNNs are characterized by a directed acyclic graph rather than a chain, so that partition is incredibly complicated. To solve the issues, We design a Dynamic Adaptive DNN Surgery(DADS) scheme, which optimally partitions the DNN under different network conditions. We also study the partition problem under the cost-constrained system, where the resource of the cloud for inference is limited. Then, a real-world prototype based on the selif-driving car video dataset is implemented, showing that compared with current approaches, DNN surgery can improve latency up to 6.45 times and improve throughput up to 8.31 times. We further evaluate DNN surgery through two case studies where we use DNN surgery to support an indoor intrusion detection application and a campus traffic monitor application, and DNN surgery shows consistently high throughput and low latency.

This paper studies the problem of cross-network node classification to overcome the insufficiency of labeled data in a single network. It aims to leverage the label information in a partially labeled source network to assist node classification in a completely unlabeled or partially labeled target network. Existing methods for single network learning cannot solve this problem due to the domain shift across networks. Some multi-network learning methods heavily rely on the existence of cross-network connections, thus are inapplicable for this problem. To tackle this problem, we propose a novel graph transfer learning framework AdaGCN by leveraging the techniques of adversarial domain adaptation and graph convolution. It consists of two components: a semi-supervised learning component and an adversarial domain adaptation component. The former aims to learn class discriminative node representations with given label information of the source and target networks, while the latter contributes to mitigating the distribution divergence between the source and target domains to facilitate knowledge transfer. Extensive empirical evaluations on real-world datasets show that AdaGCN can successfully transfer class information with a low label rate on the source network and a substantial divergence between the source and target domains.

The core idea of cross-domain recommendation is to alleviate the problem of data scarcity. Previous methods have made brilliant successes. However, many of them mainly focus on learning an ideal mapping function across-domains, ignoring the user preferences within a specific domain, which leads to suboptimal results. In this paper, we propose a Cross-Domain Recommendation Variational AutoEncoder framework (CDRVAE), a novel extension of a variational autoencoder on cross-domain recommendations for user behaviour distribution modeling. It applies a new hybrid architecture of VAE as the backbone and simultaneously constructs two information flows, within-domain and cross-domain modeling. For the former, an asymmetric codec structure is designed to reconstruct preference distribution from domain-specific latent factors. To relieve the posterior collapse dilemma, a combined prior is employed to increase the distribution complexity. The equivalent transition by a transformation matrix and the unobserved interaction generation by cross-domain reconstruction contribute to the latter. We combine all the above components for the more accurate and reliable user features. Extensive experiments are conducted on three public benchmark datasets to validate the effectiveness of the proposed CDRVAE. Experimental results demonstrate that CDRVAE is consistently superior to other state-of-the-art alternative baseline models.

The increasing awareness of privacy and the adoption of data regulations challenge the traditional trajectory publication framework in which a trusted server has access to the raw data from mobile clients. In the new untrusted environment, the clients call for much stronger data privacy preservation locally without sharing their raw data. Based on the emerging paradigm of federated analytics, we propose a Federated Analytics-based Secure Trajectory PUBlication (FASTPub) mechanism to operate in such untrusted environments. Compared with existing local differential privacy (LDP) methods, FASTPub guarantees LDP and loss-bounded$k$-anonymity simultaneously with greatly improved data utility. Specifically, FASTPub works interactively between the server and clients and iteratively builds up the trajectory without exposing raw data. Sampled clients only respond to selected trajectory fragments with randomized answers to preserve privacy as much as possible. The server then intelligently aggregates these randomized responses leveraging the intrinsic Apriori property and a Markov independent assumption of trajectory data to guide further iterations. Extensive experiments on synthetic and real-world datasets on two downstream tasks demonstrate that FASTPub gains a remarkably improved data utility compared to the existing state-of-the-art solutions.

Image-text matching has become a challenging task in the multimedia analysis field. Many advanced methods have been used to explore local and global cross-modal correspondence in matching. However, most methods ignore the importance of eliminating potential irrelevant features in the original features of each modality and cross-modal common feature. Moreover, the features extracted from regions in images and words in sentences contain cluttered background noise and different occlusion noise, which negatively affects alignment. Different from these methods, we propose a novel DCT-Transformer Adversarial Network (DTAN) for image-text matching in this paper. This work can obtain an effective metric based on two aspects: i) DCT-Transformer uses DCT (Discrete Cosine Transform) method based on a transformer mechanism to extract multi-domain common representations and eliminate irrelevant features from different modalities (inter-modal). Among them, DCT divides multi-modal content into chunks of different frequencies and quantifies them. ii) The adversarial network introduces an adversary idea by combining the original features of various single modalities and the multi-domain common representation, alleviating the background noise within each modality (intra-modal). The proposed adversarial feature augmentation method can easily obtain the common representation that is only useful for alignment. Extensive experiments are completed on the benchmark datasets Flickr30K and MS-COCO, demonstrating the superiority of the DTAN model over the state-of-the-art methods.

With the tremendous growth of mobile social networks (MSNs), people are highly relying on it to connect with friends and further expand their social circles. However, the conventional friendship inference techniques have issues handling such a large yet sparse multi-source data. The related friend recommendation systems are therefore suffering from reduced accuracy and limited scalability. To address this issue, we propose a Two-stage Deep learning framework for Friendship Inference, namely TDFI. This approach enables MSNs to exploit multi-source information simultaneously, rather than hierarchically. Therefore, there is no need to manually set which information is more important and the order in which the various information is applied. In details, we apply an Extended Adjacency Matrix (EAM) to represent the multi-source information. We then adopt an improved Deep Auto-Encoder Network (iDAEN) to extract the fused feature vector for each user. Our framework also provides an improved Deep Siamese Network (iDSN) to measure user similarity. To provide a substantial description and evaluation of the proposed methodology, we evaluate the effectiveness and robustness on three large-scale real-world datasets. Trace-driven evaluation results demonstrate that TDFI can effectively handle the sparse multi-source data while providing better accuracy for friendship inference. Through the comparison with numerous state-of-the-art methods, we find that TDFI can achieve superior performance via real-world multi-source information. Meanwhile, it demonstrates that the proposed pipeline can not only integrate structural information and attribute information, but also be compatible with different attribute information, which further enhances the overall applicability of friend-recommendation systems under information-rich MSNs.

To meet the user demand for high-speed and low-latency video services, this paper envisions a cooperative caching and video transcoding architecture in the multi-access edge computing (MEC)-enabled heterogeneous network. Under this architecture, we propose a novel knowledge graph (KG)-based video caching scheme. Specifically, KG reveals the relation between videos, which acts as an external knowledge to reflect user preferences thus guiding caching decisions. The goal of this paper is to minimize the service delay within the caching and computing resource constraints. To this end, we combine KG with the deep reinforcement learning (DRL) method and design a KG-deep Q network (DQN) based caching algorithm. KG is designed to select the related videos as candidate actions for DQN's caching decision. This way improves the convergence performance of DRL while providing rich external references for caching decisions. Numerous simulation results demonstrate that the proposed algorithm outperforms the traditional baselines on cache hit rate and delay performance.

Local differential privacy (LDP) is a prominent approach and widely adopted in federated learning (FL) to preserve the privacy of local training data. It also nicely provides a rigorous privacy guarantee with computational efficiency in theory. However, a strong privacy guarantee with local differential privacy can degrade the adversarial robustness of the learned global model. To date, very few studies focus on the interplay between LDP and the adversarial robustness of federated learning. In this paper, we observe that LDP adds random noise to the data to achieve privacy guarantee of local data, and thus introduces uncertainty to the training dataset of federated learning. This leads to decreased robustness. To solve this robustness problem caused by uncertainty, we propose to leverage the promising distributionally robust optimization (DRO) modeling approach. Specifically, we first formulate a distributionally robust and private federated learning problem (DRPri). While our formulation successfully captures the uncertainty generated by the LDP, we show that it is not easily tractable. We thus transform our DRPri problem to another equivalent problem, under the Wasserstein distance-based uncertainty set, which is named the DRPri-W problem. We then design a robust and private federated learning algorithm, RPFL, to solve the DRPri-W problem. We analyze RPFL and theoretically show it satisfies differential privacy with a robustness guarantee. We evaluate algorithm RPFL by training classifiers on real-world datasets under a set of well-known attacks. Our experimental results show our algorithm RPFL can significantly improve the robustness of the trained global model under differentially private data by up to 4.33 times.

Recent years have seen booming development of realtime streaming services, highly improving user experience in remote work, online education, and entertainment. Unlike video-on-demand (VoD) or live services, realtime streaming service has extremely stringent delay requirements, rendering the TCP-based transmission no longer applicable. Existing works based on UDP (or its variants) either suffer from the packet loss problem or only focus on improving several QoS metrics, which cannot achieve satisfactory user QoE. Our insight is to slightly sacrifice the bitrate and video quality to trade for the most significant delay to maximize the overall QoE. We propose Oppugno‡‡Oppugno is a spell in Harry Potter that makes magical creatures attack the caster. It is a metaphor that we use an additional mechanism to mitigate the influence of packet loss., an integrated framework that achieves joint loss adaptation and bitrate adaption towards maximized QoE in realtime streaming services. Oppugno leverages existing UDP mechanisms and employs an advanced deep reinforcement learning algorithm Proximal Policy Optimization (PPO), to adaptively select optimal actions based on network conditions. Trace-driven experiments demonstrate the superiority of our framework, which outperforms the SOTA work by 3.9% ∼ 11.6%.

Federated learning (FL) is a new distributed machine learning paradigm that enables machine learning on edge devices. One unique feature of FL is that edge devices belong to individuals; and since they are not “owned” by the FL coordinator, but can be “federated” instead, there can potentially be a huge number of edge devices. In the current distributed ML architecture, the parameter server (PS) architecture, model aggregation is centralized. When facing a large number of edge devices, the centralized model aggregation becomes the bottleneck and fundamentally restricts system scalability.

Edge-side real-time video analytics systems recognize spatial or temporal events (e.g., vehicle counting) in a video stream. To meet the delay requirement, existing systems in smart edge cameras conduct video preprocessing to filter out unnecessary frames and model inference using appropriately selected neural network (NN) models. Video preprocessing is instruction-intensive computing (IIC) and executed by the CPU of the edge camera, and model inference is data-intensive computing (DIC) and executed by the GPU of the edge camera. In this paper, we show that the analytics accuracy of existing systems can largely vary in fields. The root cause is that video analytics applications have different contents, which result in dynamic IIC and DIC workloads. Unfortunately, intelligent cameras in fields have fixed CPU and GPU resources and cannot effectively adapt to workload dynamics. We develop Gemini, a new real-time video analytics system enhanced by a dual-image FPGA. The newly developed dual-image FPGAs can be pre-configured with two FPGA images with a key advantage of negligible image switching time. We thus pre-configure one CPU image and one GPU image and elastically multiplex the dual CPU-GPU resources in the time dimension. The Gemini system design requires both hardware and software revisions. We overcame a challenge that the application development on different dual-image FPGAs is hardware-dependent. We develop a new abstraction of hardware functions to make the Gemini system hardware-agnostic. It is also a challenge to adapt to the dynamic workloads and optimize video analytics accuracy. We develop a bandit learning approach to capture content dynamics and conduct dual computing resource control. We implement Gemini and show that Gemini can improve the analytics accuracy to 90.35 %. We further evaluate Gemini by a case study where we use Gemini to support an intrusion detection application, and Gemini shows consistent high analytics accuracy.

Real-time edge-cloud video analytics systems have been widely used to support such applications as traffic counting, surveillance, autonomous driving, Metaverse, etc. In such a system, the edge and the cloud cooperatively conduct model inference of the video frames captured by the camera of the edge, using a trained DNN model of the video analytics application. The edge conducts initial analytics on the video frames to a split layer of the DNN model; and then sends intermediate results to the cloud for follow-up analytics. In this paper, we show that an attacker can perform reconstruction attacks to the intermediate results; and private information of the raw video frames, e.g., a plate number of a car, can be leaked. In this paper, we present Preva, a new Privacy preserving Real-time Edge-cloud Video Analytics system. The core idea of Preva is to conduct image transformation on the video frames, as preprocessing, prior to the video frames starting the edge-cloud video analytics process, so that during edge-cloud video analytics, the intermediate results will not leak private information under attack. We design a policy-based video-frame transformation scheme. Given the resource constraints of the edge, Preva ensures high accuracy in the final video analytics results and minimizes privacy leakage in any split layer. We present a formal privacy analysis and we show that Preva can guarantee privacy leakage under the reconstruction attacks of both outsider attackers and insider attackers. We evaluate Preva through three video analytics applications and we show that Preva outperforms existing systems for 64.4% in analytics accuracy and 59.2% in privacy leakage.

Frequent pattern mining is an important class of knowledge discovery problems. It aims at finding out high-frequency items or structures (e.g., itemset, sequence) in a database, and plays an essential role in deriving other interesting patterns, like association rules. The traditional approach of gathering data to a central server and analyze is no longer viable due to the increasing awareness of user privacy and newly established laws on data protection. Previous privacy-preserving frequent pattern mining approaches only target a particular problem with great utility loss when handling complex structures. In this paper, we take the first initiative to propose a unified federated analytics framework (FedFPM) for a variety of frequent pattern mining problems, including item, itemset, and sequence mining. FedFPM achieves high data utility and guarantees local differential privacy without uploading raw data. Specifically, FedFPM adopts an interactive query-response approach between clients and a server. The server meticulously employs the Apriori property and the Hoeffding’s inequality to generates informed queries. The clients randomize their responses in the reduced space to realize local differential privacy. Experiments on three different frequent pattern mining tasks demonstrate that FedFPM achieves better performances than the state-of-the-art specialized benchmarks, with a much smaller computation overhead.

The past decade has witnessed the rapid expansion of demands for mobile traffic, while the traditional mobile traffic pricing schemes cannot accommodate such demands. Sponsored data plan (SDP), which can increase the revenue of all stakeholders in the market through transferring some of the revenue from content providers (CPs) to end users (EUs), is more suitable. However, existing studies have focused more on Internet service providers (ISPs) and CPs, ignoring the influence of EUs (e.g., the inherent attribute differences of EUs and the interaction among EUs) on the market under SDP. Regarding the difficulty of modeling the abstract property about interaction among EUs, we utilize network congestion as the medium and construct the congestion-aware SDP model based on Stackelberg game. The newly proposed model can not only analyze how network congestion affects SDP mechanism, but also elucidate the impact of interactions among EUs. More specifically, through theoretical analysis, we prove that there is a unique dynamic equilibrium in the interaction among EUs (i.e., the traffic consumption of different EUs). By taking into account network congestion, the newly proposed model also more accurately and realistically describes the optimal strategies and computation methods of all stakeholders in the market. Moreover, simulation experiments demonstrate that the positive effect brought by SDP is not as obvious as before, and EUs influence each other instead of being independent of each other. Overall, this paper emphasizes the non-negligible influence of EUs and promotes a deeper understanding of SDP mechanism, which can guide the relevant stakeholders to optimize their own decision-making details.

For the widely used Bluetooth Low-Energy (BLE) neighbor discovery, the parameter configuration of neighbor discovery directly decides the results of the trade-off between discovery latency and power consumption. Therefore, it requires evaluating whether any given parameter configuration meets the demands. The existing solutions, however, are far from satisfactory due to unsolved issues. In this paper, we propose Blender, a simulation framework that produces a determined and full probabilistic distribution of discovery latency for a given parameter configuration. To capture the key features in practice, Blender provides adaption to the stochastic factors such as the channel collision and the random behavior of the advertiser. Evaluation results show that, compared with the state-of-art simulators, Blender converges closer to the traces from the Android-based realistic estimations. Blender can be used to guide parameter configuration for BLE neighbor discovery systems where the trade-off between discovery latency and power consumption is of critical importance.

Recent developments in the Internet of Vehicles (IoV) enable interconnected vehicles to support ubiquitous services. Various emerging service applications are promising to increase the Quality of Experience (QoE) of users. On-board computation tasks generated by these applications have heavily overloaded the resource-constrained vehicles, forcing it to offload on-board tasks to other edge intelligence (EI)-assisted servers. However, excessive task offloading can lead to severe competition for communication and computation resources among vehicles, thereby increasing the processing latency, energy consumption, and system cost. To address these problems, we investigate the transmission-awareness and computing-sense uplink resource management problem and formulate it as a time-varying Markov decision process. Considering the total delay, energy consumption, and cost, quantum-inspired reinforcement learning (QRL) is proposed to develop an intelligence-oriented edge offloading strategy. Specifically, the vehicle can flexibly choose the network access mode and offloading strategy through two different radio interfaces to offload tasks to multiaccess edge computing (MEC) servers through WiFi and cloud servers through 5G. The objective of this joint optimization is to maintain a self-adaptive balance between these two aspects. Simulation results show that the proposed algorithm can significantly reduce the transmission latency and computation delay.

The local model poisoning attack is an attack to manipulate the shared local models during the process of distributed learning. Existing defense methods are passive in the sense that they try to mitigate the negative impact of the poisoned local models instead of eliminating them. In this paper, we leverage the new federated analytics paradigm, to develop a proactive defense method. More specifically, federated analytics is to collectively carry out analytics tasks without disclosing local data of the edge devices. We propose a Federated Anomaly Analytics enhanced Distributed Learning (FAA-DL) framework, where the clients and the server collaboratively analyze the anomalies. FAA-DL firstly detects all the uploaded local models and splits out the potential malicious ones. Then, it verifies each potential malicious local model with functional encryption. Finally, it removes the verified anomalies and aggregates the remaining to produce the global model. We analyze the FAA-DL framework and show that it is accurate, robust, and efficient. We evaluate FAA-DL by training classifiers on MNIST and Fashion-MNIST under various local model poisoning attacks. Our experiment results show FAA-DL improves the accuracy of the learned global model under strong attacks up to 6.90 times and outperforms the state-of-the-art defense methods with a robustness guarantee.

It is well-known that the HVAC (heating, ventilation, and air conditioning) dominates electricity consumption in commercial buildings. In this paper, we focus on one of the core problems in building operation, namelychiller sequencingto reduce HVAC electricity consumption. Our contributions are threefold. First, we make a case for why it is important to quantify the performance profile of a chiller, namely coefficient of performance (COP), atrun-time, by developing a data-driven COP estimation methodology. Second, we show that predicting COP accurately is a non trivial problem, requiring considerable computation time. To overcome this barrier, we develop a data-driven COP prediction model and an edge-based chiller sequencing framework integrating the COP predictions, and show that they strike a good balance between electricity saving and ease of use for real-world deployment. Finally, we evaluate the performance of our scheme by applying it to real-world data, spanning four years, obtained from multiple chillers across three large commercial buildings in Hong Kong. The results show an electricity saving of over 30 percent compared to baselines. We offer our edge based data-driven chiller sequencing framework as a cost-effective and practical mechanism to reduce electricity consumption associated with HVAC operation in commercial buildings.

The emerging federated optimization paradigm performs data mining or artificial intelligence techniques locally on the edge devices, enabling scientists and engineers to utilize the blooming edge data with privacy protection. In such a paradigm, since data cannot be shared or gathered, data heterogeneity naturally emerges, which significantly degrades the performance of federated optimization, ultimately leading to poor quality of federated services. In this paper, we present the first work on characterizing the data heterogeneity in the framework of federated analytics, i.e., to collectively carry out analytics tasks without raw data sharing, and use the information to create a desirable data environment via intelligent client selection. Our proposed Analytics-driven Client Selection framework, named FedACS, tackles the data heterogeneity problem in three steps. First, clients are in charge of generating insights about local data without disclosure of sensitive information. Then, the server uses these insights to infer the situation of clients’ data heterogeneity based on the Hoeffding’s inequality. Finally, a dueling bandit is formulated to intelligently select clients with slighter data heterogeneity to form a desirable client pool. FedACS can be universally applied to all kinds of federated optimization tasks, and gains benefits including privacy protection, infrastructure reuse, and client load reduction. To test its efficiency, we further customize it to assist federated learning, a popular scenario of federated optimization. According to experiment results, FedACS reduces the accuracy degrading by up to 65.6%, and speeds up the convergence for up to 2.4 times.

We are now in an information era and the volume of data is growing explosively. However, due to privacy issues, it is very common that data cannot be freely shared among the data generating.Federated analyticswas recently proposed aiming at deriving analytical insights among data-generating devices without exposing the raw data, but the intermediate analytics results. Note that the computing resources at the data generating devices are limited, thus making on-device execution of computing-intensive tasks challenging. We thus propose to apply the digital twin technique, which emulates the resource-limited physical/end side, while utilizing the rich resource at the virtual/computing side. Nevertheless, how to use the digital twin technique to assist federated analytics while preserving distributed data privacy is challenging. To address such a challenge, this work first formulates a problem on digital twin-assisted federated distribution discovery. Then, we propose a federated Markov chain Monte Carlo with a delayed rejection (FMCMC-DR) method to estimate the representative parameters of the global distribution. We combine a rejection–acceptance sampling technique and a delayed rejection technique, allowing our method to be able to explore the full state space. Finally, we evaluate FMCMC-DR against the Metropolis–Hastings (MH) algorithm and random walk Markov chain Monte Carlo method (RW-MCMC) using numerical experiments. The results show our algorithm outperforms the other two methods by 50% and 95% contour accuracy, respectively, and has a better convergence rate.

Wireless virtual reality (VR)-enabled medical treatment (WVMT) system, integrating the VR technology and the platform of the Internet of Medical Things (IoMT), is a promising application in future medical industries. Multiaccess edge computing (MEC) is an effective approach to support the ubiquitous applications of WVMT systems. Due to the high requirements of medical services, the computation efficiency and security are two issues in WVMT systems. In this article, we propose a blockchain-enabled task offloading scheme, where the viewport rendering tasks of VR devices (VDs) can be offloaded to edge access points (EAPs). The blockchain is integrated into the system to reach the consensus of the global information of task offloading and data processing to resist malicious attacks. To reduce VDs’ computation load under the promise of high VR QoE, we formulate the computation offloading and resource allocation to be a Markov decision problem, considering block consensus, content correlation, and fluctuating channel conditions. Then, a novel collective reinforcement learning (CRL) algorithm is proposed to adaptively allocate resources based on the requirements of viewport rendering, block consensus, and content transmission. In the simulations, the convergence rate and the performance in terms of energy consumption and stalling rate are evaluated. simulation results demonstrate the effectiveness of the proposed scheme.

The evolution of the Internet of Things (IoT) makes an increased emphasis on extending their computing and storage capabilities by relying particularly on the cloud/edge computing (EC) for cyber-physical systems (CPSs). Especially, in software-defined CPS (SD-CPS), different software-defined networking (SDN) controllers share information and cooperate to make global decisions. To further enhance system security during the information sharing process, we introduce blockchain technology into SD-CPS. However, because many security-related decisions are sensitive to latency, it is vital to minimize the system latency in blockchain-empowered SD-CPS. In this article, a blockchain-empowered distributed SD-CPS framework is proposed to realize consensus and distributed resource management by offloading data in a hybrid network paradigm that combines cloud computing and EC. Moreover, to adaptively implement offloading and control strategies while guaranteeing data security, we design a resource management scheme for reducing system latency and provide the flexibility of cooperation. To foster intelligence, we formulate the joint communication, computation, and consensus problems as a Markov decision process and use deep reinforcement learning to balance resource allocation, reduce latency, and guarantee data security. Compared with other schemes, simulation results verify the effectiveness of the proposed scheme, which performs better on self-adaptation decision making and system delay reduction.

Wireless virtual reality (VR) is increasingly used in industrial Internet of Things (IIoTs). However, ultra-high viewport rendering demands and excessive terminal energy consumption restrict the application of wireless VR. Pervasive edge computing emerges as a promising method for wireless VR. In this article, we propose an energy-aware resource management scheme for wireless-VR-supported IIoTs. To reduce the energy consumption of VR equipments (VEs) while ensuring a smooth immersive VR experience, we formulate the viewport rendering offloading, computing, and spectrum resource allocation to be a joint optimization problem, considering content correlation between VEs, fluctuating channel conditions, and VR quality of experience. By applying dual approximation, the original problem is transformed to be a Markov decision process and an reinforcement learning (RL)-based online learning algorithm is designed to find the optimal policy. To improve the learning efficiency, the quantum parallelism is integrated into the RL to overcome “curse of dimensionality”. In the simulations, the convergence rate and the performance in terms of energy consumption and stalling rate are evaluated. Simulation results demonstrate the effectiveness of the proposed scheme.

Conversion rate (CVR) prediction is becoming increasingly important in the multi-billion dollar online display advertising industry. It has two major challenges: firstly, the scarce user history data is very complicated and non-linear; secondly, the time delay between the clicks and the corresponding conversions can be very large, e.g., ranging from seconds to weeks. Existing models usually suffer from such scarce and delayed conversion behaviors. In this paper, we propose a novel deep learning framework to tackle the two challenges. Specifically, we extract the pre-trained embedding from impressions/clicks to assist in conversion models and propose an inner/self-attention mechanism to capture the fine-grained personalized product purchase interests from the sequential click data. Besides, to overcome the time-delay issue, we calibrate the delay model by learning dynamic hazard function with the abundant post-click data more in line with the real distribution. Empirical experiments with real-world user behavior data prove the effectiveness of the proposed method.

Emotions show high significance on human health. Automatic emotion recognition is helpful for monitoring psychological disorders, mental problems and exploring behavioral mechanisms. Existing approaches adopt costly and bulky specialized hardware such as EEG/ECG helmet, possess privacy risks, or with low accuracy and user experience. With the increasing popularity of wearables, people tend to equip multiple smart devices, which provides potential opportunity for emotion perception. In this paper, we present a pervasive and portable system called MW-Emotion to recognize common emotional states with multi-modal wearable devices. However, ubiquitous wearable devices perceive shallow information which is not obviously related to human emotions. MW-Emotion excavates intrinsic mapping relationship between emotions and sensing data. Our experiments show that MW-Emotion can recognize different emotion states with a relatively high accuracy of 83.1%.

Strength training is becoming increasingly popular among all age groups, as it helps the participants increase muscle strength, improve body flexibility, reduce health risks, and reshape physical forms. However, strength training imposes strict regulations on gestures and requires professional instruction in real time for the sake of body-building efficiency and safety. For this purpose, in this article, we propose a novel low-cost system named iCoach, to provide real-time monitoring and coaching service for strength training participants. Specifically, we design and implement a smart fitness glove, which can be seamlessly equipped with a pervasive inertial unit. With this customized but low-cost device, we can recognize various training programs, detect nonstandard behaviors while exercising, and assess exercising qualities of a user. Our primary experimental results show that iCoach can recognize 15 sets of training programs, detect three common nonstandard behaviors, and assess the quality of training with high accuracy and reliability.

Sponsored Data Plan (SDP) is an emerging pricing model for the wireless data market where the Content Provider (CP) can sponsor the data usage for specific content on behalf of the users. This strategy sheds new light on the data pricing model and receives significant attention from the Internet Service Provider (ISP). However, the existing SDP studies consider traffic price (e.g., sponsorship) as the only factor that affects user decision. The impact of other classic market features, such as the demand for a variety of contents (i.e., love of variety), remains largely unclear. In this paper, we develop a new model to understand the love of variety in the wireless data market under SDPs. Our model has demonstrated that, such variety is important to understand the complex gaming between ISPs, CPs, and users in both short-run and long-run markets. For example, the analysis indicates that the advantage of CPs with higher revenue will be significantly reduced when users have a greater love of variety. Moreover, to help the ISP better adopt the proposed model in the real market, we also develop a practical method to calibrate the related parameters, which can also be applied to quantity the love of variety.

In this article, we study solder paste inspection (SPI), an important stage that is used in the semiconductor manufacturing industry, where abnormal boards should be detected. A highly accurate SPI can substantially reduce human expert involvement, as well as reduce the waste in disposing of the boards in good condition. A key difference today is that because of increasing demand in board customization, the number of board types increases substantially and quantity of the boards produced in each type decreases. Thus, the previous approaches where a fine-tuned model is developed for each board type are no longer viable. Intrinsically, our problem is an anomaly detection problem. A major specialty in today’s SPI is that the target tasks for prediction cannot be fully pre-determined due to context changes during the solder paste printing stage. Our experiences show that a conventional approach to first define a set of tasks and train these tasks offline will lead to low accuracy. Here, we propose a novel multi-task approach, where the performance of all target tasks is ensured simultaneously. We note that the SPI process is streamlined and automatic, allowing the SPI time for only a few seconds. We propose a fast clustering algorithm that reuses existing models to avoid retraining and fine tune in the inference phase. We evaluate our approach using 3-month data collected from production lines. We show that we can reduce 81.28% of false alarms. This can translate to annual savings of $11.3 million.

With increasing demands for more flexible services, the routing policies in enterprise networks become much richer. This has placed a heavy burden to the current router forwarding plane in support of the increasing number of policies, primarily due to the limited capacity in TCAM, which further hinders the development of new network services and applications. The scalable forwarding table structures for enterprise networks have therefore attracted numerous attentions from both academia and industry. To tackle this challenge, in this paper we present the design and implementation of a new forwarding table structure. It separates the functions of TCAM and SRAM, and maximally utilizes the large and flexible SRAM. A set of schemes are progressively designed, to compress storage of forwarding rules, and maintain correctness and achieve line-card speeds of packet forwarding. We further design an incremental update algorithm that allows less access to memory. The proposed scheme is validated and evaluated through a realistic implementation on a commercial router using real datasets. Our proposal can be easily implemented in the existing devices. The evaluation results show that the performance of forwarding tables under the proposed scheme is promising.

Urban Mobility Models (UMMs) are fundamental tools for estimating the population in urban sites and their spatial movements over time. Most existing UMMs were developed primarily in 2D. However, we argue that people’s movements and living patterns involve 3D space, i.e., buildings, which can heavily affect the accuracy of UMMs. In this article, we for the first time conduct a comprehensive study on the impacts of buildings on human movements and the effect on UMMs. We innovatively capture the impacts by developing a Semi-absorbing Urban Mobility model (SUM) and theoretically prove its properties on its difference from that of previous UMMs. We also show that calibrating our original SUM may need a large number of parameters. As such, we develop two SUM extensions with a substantially reduced number of parameters, making calibration practical. Our evaluation also demonstrates that, as a basis for supporting mobile applications in an intracity and hourly scale, the SUM is far superior to previous UMMs. In a case study, we also show that the performance of the resource allocation scheme in a cellular network substantially improves by using SUM, with a reduction in the packet loss probability of 3.19 times.

On edge devices, data scarcity occurs as a common problem where transfer learning serves as a widely-suggested remedy. Nevertheless, transfer learning imposes heavy computation burden to the resource-constrained edge devices. Existing task allocation works usually assume all submitted tasks are equally important, leading to inefficient resource allocation at a task level when directly applied in Multi-task Transfer Learning (MTL). To address these issues, we first reveal that it is crucial to measure the impact of tasks on overall decision performance improvement and quantify task importance. We then show that task allocation with task importance for MTL (TATIM) is a variant of NP-complete Knapsack problem, where the complicated computation to solve this problem needs to be conducted repeatedly under varying contexts. To solve TATIM with high computational efficiency, we propose a Data-driven Cooperative Task Allocation (DCTA) approach. Finally, we evaluate the performance of DCTA by not only a trace-driven simulation, but also a new comprehensive real-world AIOps case study which bridges model and practice via a new architecture and main components design within AIOps system. Extensive experiments show that our DCTA reduces 3.24 times of processing time, and saves 48.4 percent energy consumption compared with the state-of-the-art when solving TATIM.

Presents corrections to author information for the above named paper.

Edge computing for machine learning has become a heated research topic. On edge devices, data scarcity occurs as a common problem where transfer learning serves as a widely-suggested remedy. Nevertheless, one obstacle is that transfer learning imposes heavy computation burden to the resource-constrained edge devices. Motivated by the fact that only a few tasks of Multi-task Transfer Learning (MTL) have a higher potential for overall decision performance improvement, we design a novel task allocation scheme, which assigns more important tasks to more powerful edge devices to maximize the overall decision performance. In this paper, we focus on task allocation under multi-task scenarios by introducing task importance and make the following contributions. First, we reveal that it is important to measure the impact of tasks on overall decision performance improvement and quantify task importance. We also observe the long-tail property of task importance, i.e., only a few tasks are important, which facilitates more efficient task allocation. Second, we show that task allocation with task importance for MTL (TATIM) is in fact a variant of the NP-complete Knapsack problem, where the complicated computation to solve this problem needs to be conducted repeatedly under varying contexts. To solve TATIM with high computational efficiency, we innovatively propose a Data-driven Cooperative Task Allocation (DCTA) approach. Third, we evaluate the performance of our DCTA approach by applying it to a real-world industrial operation (e.g., AIOps) scenario. Experiments show that our DCTA approach can reduce 3.24 times of processing time compared with the state-of-the-art when solving TATIM. We offer our DCTA approach as an effective and practical mechanism for reducing the required resource associated with performing MTL on edge devices.

Task Relation Discovery (TRD), i.e., reveal the relation of tasks, has notable value: it is the key concept underlying Multi-task Learning (MTL) and provides a principled way for identifying redundancies across tasks. However, task relation is usually specifically determined by data scientist resulting in the additional human effort for TRD, while transfer based on brute-force methods or mere training samples may cause negative effects which degrade the learning performance. To avoid negative transfer in an automatic manner, our idea is to leverage commonly available context attributes in nowadays systems, i.e., the metadata. In this paper, we, for the first time, introduce metadata into TRD for MTL and propose a novel Metadata Clustering method, which jointly uses historical samples and additional metadata to automatically exploit the true relatedness. It also avoids the negative transfer by identifying reusable samples between related tasks. Experimental results on five real-world datasets demonstrate that the proposed method is effective for MTL with TRD, and particularly useful in complicated systems with diverse metadata but insufficient data samples. In general, this study helps in automatic relation discovery among partially related tasks and sheds new light on the development of TRD in MTL through the use of metadata as apriori information.

Recent advances in deep neural networks (DNNs) have substantially improved the accuracy and speed of a variety of intelligent applications. Nevertheless, one obstacle is that DNN inference imposes heavy computation burden to end devices, but offloading inference tasks to the cloud causes transmission of a large volume of data. Motivated by the fact that the data size of some intermediate DNN layers is significantly smaller than that of raw input data, we design the DNN surgery, which allows partitioned DNN processed at both the edge and cloud while limiting the data transmission. The challenge is twofold: (1) Network dynamics substantially influence the performance of DNN partition, and (2) State-of-the-art DNNs are characterized by a directed acyclic graph (DAG) rather than a chain so that partition is greatly complicated. In order to solve the issues, we design a Dynamic Adaptive DNN Surgery (DADS) scheme, which optimally partitions the DNN under different network condition. Under the lightly loaded condition, DNN Surgery Light (DSL) is developed, which minimizes the overall delay to process one frame. The minimization problem is equivalent to a min-cut problem so that a globally optimal solution is derived. In the heavily loaded condition, DNN Surgery Heavy (DSH) is developed, with the objective to maximize throughput. However, the problem is NP-hard so that DSH resorts an approximation method to achieve an approximation ratio of 3. Real-world prototype based on self-driving car video dataset is implemented, showing that compared with executing entire the DNN on the edge and cloud, DADS can improve latency up to 6.45 and 8.08 times respectively, and improve throughput up to 8.31 and 14.01 times respectively.

Due to the explosive growth of social network services, friendship inference has been widely adopted by Online Social Service Providers (OSSPs) for friend recommendation. The conventional techniques, however, have limitations in accuracy or scalability to handle such a large yet sparse multi-source data. For example, the OSSPs will be required to manually give the order in which the various information is applied. This unavoidably reduces the applicability of existing friend recommendation systems. To address this issue, we propose a Two-stage Deep learning framework for Friendship Inference (TDFI). This approach can utilize multi-source information simultaneously with low complexity. In particular, we apply an Extended Adjacency Matrix (EAM) to represent the multi-source information. We then adopt an improved Deep AutoEncoder Network (iDAEN) to extract the fused feature vector for each user. The TDFI framework also provides an improved Deep Siamese Network (iDSN) to measure user similarity from iDAEN. Finally, we evaluate the effectiveness and robustness of TDFI on three large-scale real-world datasets. It shows that TDFI can effectively handle the sparse multi-source data while providing better accuracy for friend recommendation.

In this paper, we develop a new model to study the competition among Content Providers (CPs) under Sponsored Data Plans (SDPs). SDP is an emerging pricing model for the wireless data market where Internet Service Providers (ISPs) allow a CP to compensate the traffic volume of users when users access the contents of this CP. Studies have shown that SDPs create a triple-win situation, where users consume more contents and the revenue of both CPs and ISPs increases. Currently, a main concern of SDPs is on whether SDPs may bring about unfair competition among CPs. Studies have shown that big CPs have an advantage over small CPs. We observe that such conclusions are derived because in all previous models, traffic price is the only factor that affects user decisions. We argue that it is not precise. Nowadays, people conduct a large variety of activities online, and users have an intrinsic demand for a variety of contents. To reflect this, we for the first time characterize the variety demand as an intrinsic parameter of users, and integrate such variety into a new model to help us drive some novel insights into SDPs, especially the competition among CPs. Our model shows that variety matters for understanding SDPs more thoroughly and comprehensively. For example, under SDPs, the advantage of CPs with higher revenue will be significantly reduced if users have a greater love for variety. Overall, our new model leads to a set of completely new results and rectifies some past conclusions.

Recently, wearable devices have become increasingly popular in our lives because of their neat features and stylish appearance. However, due to the tiny size, it is inconvenient for users to interact with a device using conventional methods, especially for text entry. Although some methods have been proposed to handle this problem, they have different limitations and are not applicable to many existing mobile devices. As a result, we take the first step to propose a digits-entry system, i.e., AcouDigits, in which digits can be entered in the air using a finger without taking help from any additional hardware. We implement AcouDigits on two commercial devices and conduct experiments to evaluate its performance in recognizing ten basic digits. Experimental results show that AcouDigits can achieve average accuracies of 91.7% and 87.4% in recognizing basic digits and 26 English alphabets, respectively.

Network Embedding is the task of learning continuous node representations for networks, which has been shown effective in a variety of tasks such as link prediction and node classification. Most of existing works aim to preserve different network structures and properties in low-dimensional embedding vectors, while neglecting the existence of noisy information in many real-world networks and the overfitting issue in the embedding learning process. Most recently, generative adversarial networks (GANs) based regularization methods are exploited to regularize embedding learning process, which can encourage a global smoothness of embedding vectors. These methods have very complicated architecture and suffer from the well-recognized non-convergence problem of GANs. In this paper, we aim to introduce a more succinct and effective local regularization method, namely adversarial training, to network embedding so as to achieve model robustness and better generalization performance. Firstly, the adversarial training method is applied by defining adversarial perturbations in the embedding space with an adaptive L2 norm constraint that depends on the connectivity pattern of node pairs. Though effective as a regularizer, it suffers from the interpretability issue which may hinder its application in certain real-world scenarios. To improve this strategy, we further propose an interpretable adversarial training method by enforcing the reconstruction of the adversarial examples in the discrete graph domain. These two regularization methods can be applied to many existing embedding models, and we take DeepWalk as the base model for illustration in the paper. Empirical evaluations in both link prediction and node classification demonstrate the effectiveness of the proposed methods.

Containerization has been used in many applications for isolation purposes due to its lightweight, scalable, and highly portable properties. However, to apply containerization in large-scale Internet data centers faces a big challenge. Services in data centers are always instantiated as a group of containers, which often generate heavy communication workloads and therefore resulting in inefficient communications and downgraded service performance. Although assigning the containers of the same service to the same server can reduce the communication overhead, this may cause heavily imbalanced resource utilization since containers of the same service are usually intensive to the same resource. To reduce communication cost as well as balance the resource utilization in large-scale data centers, we further explore the container distribution issues in a real industrial environment and find that such conflict lies in two phases-container placement and container reassignment. The objective of this paper is to address the container distribution problem in these two phases. For the container placement problem, we propose an efficient communication aware worst fit decreasing algorithm to place a set of new containers into data centers. For the container reassignment problem, we propose a two-stage algorithm called Sweep&Search to optimize a given initial distribution of containers by migrating containers among servers. We implement the proposed algorithms in Baidu's data centers and conduct extensive evaluations. Compared with the state-of-the-art strategies, the evaluation results show that our algorithms perform better up to 70% and increase the overall service throughput up to 90% simultaneously.

Learning low-dimensional representations of networks has proved effective in a variety of tasks such as node classification, link prediction and network visualization. Existing methods can effectively encode different structural properties into the representations, such as neighborhood connectivity patterns, global structural role similarities and other high-order proximities. However, except for objectives to capture network structural properties, most of them suffer from lack of additional constraints for enhancing the robustness of representations. In this paper, we aim to exploit the strengths of generative adversarial networks in capturing latent features, and investigate its contribution in learning stable and robust graph representations. Specifically, we propose an Adversarial Network Embedding (ANE) framework, which leverages the adversarial learning principle to regularize the representation learning. It consists of two components, i.e., a structure preserving component and an adversarial learning component. The former component aims to capture network structural properties, while the latter contributes to learning robust representations by matching the posterior distribution of the latent representations to given priors. As shown by the empirical results, our method is competitive with or superior to state-of-the-art approaches on benchmark network embedding tasks.

Nowadays, manufacturers want to collect the data of their sold-products to the cloud, so that they can conduct analysis and improve the operation, maintenance and services of their products. Manufacturers are looking for a self-contained solution for data transmission since their products are typically deployed in a large number of different buildings, and it is neither feasible to negotiate with each building to use the building's network (e.g., WiFi) nor practical to establish its own network infrastructure. ISPs are aware of this market. Since the readily available 3G/4G is over costly for most IoT devices, ISPs are developing new choices. Nevertheless, it can be expected that the choices from ISPs will not be fine-grained enough to match hundreds or thousands of requirements on different costs and data volumes from the IoT applications. To address this problem, we for the first time propose IoT communication sharing (ICS). We first clarify the ICS scenarios. We then formulate the IoT communication sharing (ICS) problem, and develop a set of algorithms with provable performance. We further present our implementation of a fully functioning system. Our evaluations show that ICS and our algorithms can lead to a cost reduction of five times and eight times respectively for the two real-world cases.

Recent advances in high speed rails (HSRs) are propelling the need for acceptable network service in high speed mobility environments. However, previous studies show that the performance of traditional single-path transmission degrades significantly during high speed mobility due to frequent handoff. Multi-path transmission with multiple carriers is a promising way to enhance the performance, because at any time, there is possibly at least one path not suffering a handoff. In this paper, for the first time, we measure multi-path TCP (MPTCP) with two cellular carriers on HSRs with a peak speed of 310km/h. We find a significant difference in handoff time between the two carriers. Moreover, we observe that MPTCP can provide much better performance than TCP in the poorer of the two paths. This indicates that MPTCP's robustness to handoff is much higher than TCP's. However, the efficiency of MPTCP is far from satisfactory. MPTCP performs worse than TCP in the better path most of the time. We find that the low efficiency can be attributed to poor adaptability to frequent handoff by MPTCP's key operations in sub-flow establishment, congestion control and scheduling. Finally, we discuss possible directions for improving MPTCP for such scenarios.

Nowadays, manufacturers want to send the data of their products to the cloud so that they can conduct analysis and improve their operation, maintenance, and services. Manufacturers are looking for a self-contained solution. This is because their products are deployed in a large number of different buildings, and it is neither feasible for a vendor to negotiate with each building to use the building’s network (e.g., WiFi) nor practical to establish its own network infrastructure. The vendor can rent a dedicated channel from an ISP to act as a thing-to-cloud communication (TCC) link for each of its IoT devices. The readily available choices, e.g., 3G, is over costly for most IoT devices. ISPs are developing cheaper choices for TCC links, yet we expect that the number of choices for TCC links will be small as compared to hundreds or thousands of requirements on different costs and data rates from IoT applications. We address this issue by proposing a communication sharing architecture sTube+, sharing tube . The objective of sTube+ is to organize a greater number of IoT devices, with heterogeneous data communication and cost requirements, to efficiently share fewer choices of TCC links and transmit their data to the cloud. We take a design of centralized price optimization and distributed network control. More specifically, we architect a layered architecture for data delivery, develop algorithms to optimize the overall monetary cost, and prototype a fully functioning system of sTube+. We evaluate sTube+ by both experiments and simulations. In addition, we develop a case study on smart maintenance of chillers and pumps, using sTube+ as the underlying network architecture.

Containers have been used in many applications for isolation purposes due to the lightweight, scalable and highly portable properties. However, to apply containers in virtual network functions (VNFs) faces a big challenge because high-performance VNFs often generate frequent communication workloads among containers while the container communications are generally not efficient. Compared with hardware modification solutions, properly distributing containers among hosts is an efficient and low-cost way to reduce communication overhead. However, we observe that this approach yields a trade-off between the communication overhead and the overall throughput of the cluster. In this paper, we focus on the communication-aware container redistribution problem to optimize the communication overhead and the overall throughput jointly for VNF clusters. We propose a solution called FreeContainer which utilizes a novel two-stage algorithm to re-distribute containers among hosts. We implement FreeContainer in Baidu clusters with 6000 servers and 35 services deployed. Extensive experiments on real networks are conducted to evaluate the performance of the proposed approach. The results show that FreeContainer can increase the overall throughput up to 90% with significant reduction on communication overhead.

Interdependent networks, where two networks depend on each other, are becoming more and more significant in modern systems, e.g., the control and transmission networks in the smart grid. From previous work, it can be concluded that interdependent networks are more vulnerable than a single network. The robustness in interdependent networks deserves special attention. In this paper, we propose a metric of robustness from a new perspective — the balance. We define the balance-coefficient of the interdependent system. Some evaluations verify the impact of the balance property.

Recent advances in high speed rails (HSRs), coupled with user demands for communication on the move, are propelling the need for acceptable quality of communication services in high speed mobility scenarios. This calls for an evaluation of how well popular voice/video call applications, such as Skype, can perform in such scenarios. This paper presents the first comprehensive measurement study on Skype voice/video calls in LTE networks on HSRs with a peak speed of 310 km/h in China. We collected 50 GB of performance data, covering a total HSR distance of 39,900 km. We study various objective performance metrics (such as RTT, sending rate, call drop rate, etc.), as well as subjective metrics such as quality of experience of the calls. We also evaluate the efficiency of Skype's algorithms regarding the level of utilization of network resources. We observed that the quality of Skype calls degrades significantly on HSRs. Moreover, it was discovered that Skype significantly under-utilizes the network resources, such as available bandwidth. We discovered that the root of these inefficiencies is the poor adaptability of Skype in many aspects, including overlay routing, rate control, state update and call termination. These findings highlight the need to develop more adaptive voice/video call services for high speed mobility scenarios.

In this paper, we show a design and implementation of a (partial) wireless building management system (BMS). Compared to the existing wired BMS, a wireless system can be much cheaper and more flexible in deployment. There are existing studies on smart and wireless BMS. Our design differs from others as the latter usually takes a re-arch approach and develops a brand new suite of protocols. However, it can take a considerably long time for re-standardization and adoption by vendors. Our design does not intend to tear down the full suite of upper layer protocols. We thus face difficulties as we need to maintain the upper layer protocols in operation and support their data traffic. The key ideas of our approach are an asynchronous-response framework to maintain the control plane of the upper layer protocols intact, and a modular design to prioritize and schedule data flow to handle link quality and throughput variations. We implemented the proposed design into a real system and evaluated the system by comprehensive experiments with real BMS controllers and software. In addition, we conducted a field deployment by integrating our system with the BMS in FG-building of The Hong Kong Polytechnic University. The system operated smoothly during 5-h deployment.

While TCP has been extensively studied in static and low speed mobility situations, it has not yet been well explored in high speed mobility scenarios. Given the increasing deployment of high speed transport systems (such as high speed rails), there is an urgent need to understand the performance and behavior of TCP in such high speed mobility environments. In this paper, we conduct a comprehensive study to investigate the performance and behavior of TCP in a high speed environment with a peak speed of 310 km/h. Over a 16-month period spanning four years, we collect 500 GB of performance data on 3/4G networks in high speed trains in China, covering a distance of 108,490 km. We start by analyzing performance metrics, such as RTT, packet loss rate, and throughput. We then evaluate the challenges posed on the main TCP operations (establishment, transmission, congestion control, flow control, and termination) by such high speed mobility. This paper shows that RTT and packet loss rate increase significantly and throughput drops considerably in high speed situations. Moreover, TCP fails to adapt well to such extremely high speed leading to abnormal behavior, such as high spurious retransmission time out rate, aggressive congestion window reduction, long delays during connection establishment and closure, and transmission interruption. As we prepare to move into the era of 5G, and as the need for high speed travel continues to increase, our findings indicate a critical need for efforts to develop more adaptive transport protocols for such high speed environments.

Traffic prediction, particularly in urban regions, is an important application of tremendous practical value. In this paper, we report a novel and interesting case study of urban traffic prediction in Central, Hong Kong, one of the densest urban areas in the world. The novelty of our study is that we make good second use of inexpensive big data collected from the Hong Kong International Commerce Centre (ICC), a 118-story building in Hong Kong where more than 10,000 people work. As building environment data are much cheaper to obtain than traffic data, we demonstrate that it is highly effective to estimate building occupancy information using building environment data, and then to further use the information on occupancy to provide traffic predictions in the proximate area. Scientifically, we investigate how and to what extent building data can complement traffic data in predicting traffic. In general, this study sheds new light on the development of accurate data mining applications through the second use of inexpensive big data.

We present a study on minimizing non-renewable energy for the Internet. The classification of renewable and non-renewable energy brings in several challenges. First, it is necessary to understand how the routing system can distinguish the two types of energy in the power supply. Second, the routing problem changes due to renewable energy; and so do the algorithm designs and analysis. We first clarify the model of how routers can distinguish renewable and non-renewable energy supporting their power supply. This cannot be determined by the routing system alone, and involves modeling the energy generation and supply of the grid. We then present the router power consumption model, which has a fixed startup power and a dynamic traffic-dependent power. We formulate a minimum non-renewable energy routing problem, and two special cases representing either the startup power dominates or the traffic-dependent power dominates. We analyze the complexity of these problems, develop optimal and sub-optimal algorithms, and jointly consider QoS requirements such as path stretch. We evaluate our algorithms using real data from both National and European centers. As compared to the algorithms minimizing the total energy, our algorithms can reduce the non-renewable energy consumption for more than 20% under realistic assumptions.

Mobile data demand is increasing tremendously, and thus new pricing models are in urgent need. One promising new pricing scheme is the “sponsored data plan”, i.e., end users may enjoy free access to contents from certain content providers, while these content providers will pay ISPs for corresponding traffic consumed by end users. Proven a number of advantages, the sponsored data plan is still in its infancy. In this paper, we explore some potential of further development of this plan. We extend the design space and propose the idea of time-dependent, sponsoring, i.e, content providers can decide when to sponsor how much fractions of traffic. The key intuition is by migrating some traffic consumption from peak to valley times, bandwidth resources can be better utilized. We formulate a game model to study the interactions between the ISP, CPs and users, and derive the optimal sponsoring fractions over various times under this new plan. We show that all parties involved can benefit from this plan, and social welfare increases. We believe our proposal, i.e., time-dependent sponsoring, provides important insights to potential development of the sponsored data plan.

Recently, the cloud systems face an increasing number of big data applications. It becomes an important issue for the cloud providers to allocate resources so as to accommodate as many of these big data applications as possible. In current cloud service, e.g., Amazon EMR, a job runs on a fixed cluster. This means that a fixed amount of resources (e.g. CPU, memory) is allocated to the life cycle of this job. We observe that the resources are inefficiently used in such services because of resources usage unbalance. Therefore, we propose a runtime elastic VM approach where the cloud system can increase or decrease the number of CPUs at different time periods for the jobs. There is little change to such services as Amazon EMR, yet the cloud system can accommodate many more jobs. In this paper, we first present a measurement study to show the feasibility and the quantitative impact of adjusting VM configurations dynamically. We then model the task and job completion time of big data applications, which are used for elastic VM adjustment decisions. We validate our models through experiments. We present Tetris, an elastic VM strategy based on cloud system that can better optimize resource utilization to support big data applications. We further implement a Tetris prototype and comprehensively evaluate Tetris on a real private cloud platform using Facebook trace and Wikipedia dataset. We observe that with Tetris, the cloud system can accommodate 31.3% more jobs.

MapReduce-like frameworks have achieved tremendous success for large-scale data processing in data centers. A key feature distinguishing MapReduce from previous parallel models is that it interleaves parallel and sequential computation. Past schemes, and especially their theoretical bounds, on general parallel models are therefore, unlikely to be applied to MapReduce directly. There are many recent studies on MapReduce job and task scheduling. These studies assume that the servers are assigned in advance. In current data centers, multiple MapReduce jobs of different importance levels run together. In this paper, we investigate a schedule problem for MapReduce taking server assignment into consideration as well. We formulate a MapReduce server-job organizer problem (MSJO) and show that it is NP-complete. We develop a 3-approximation algorithm and a fast heuristic design. Moreover, we further propose a novel fine-grained practical algorithm for general MapReduce-like task scheduling problem. Finally, we evaluate our algorithms through both simulations and experiments on Amazon EC2 with an implementation with Hadoop. The results confirm the superiority of our algorithms.

Satellite networks are drawing more and more attention, since they can provide various services to everywhere on the earth. Communication devices in satellites are typically powered by solar panels and battery cells, which are carefully designed to guarantee power supply and avoid deficiency. However, we find that unrestrained use of energy will cause a satellite to age quickly, because the number of recharge/discharge of battery cells is limited. Due to the extremely high cost of satellites, the development of energy-efficient satellite routing to save energy and prolong satellite lifetimes has become significantly important. In this paper, we do comprehensive studies. First, we model the power consumption of a space router, power supply by solar panels, and aging of battery cells formally. Second, we define the energy-efficient satellite routing (EESR) problem, and prove that the EESR problem is NP-hard. Then, we develop three algorithms to gradually solve the EESR problem. GreenSR-B is a baseline algorithm which computes link costs iteratively to compute a routing that minimizes the total recharge/discharge cycle number. GreenSR-A selects space routers to switch into sleep mode to improve energy conservation. GreenSR jointly considers energy efficiency and QoS requirements of path length and the maximum link utilization ratio. We evaluate our algorithms by simulations on a low earth orbit satellite network with real Internet usage traces. The results show that GreenSR can prolong the lifetime of satellite battery cells by more than 40%, with little increment in path length and a small link utilization ratio.

With the advances of the information and communications technology, and smart meters in particular, fine grained user electricity usage of households is available for analyzing electricity usage behaviors. The information makes it possible for utility companies to provide differentiated user services from the time-of-use perspective, i.e., different pricing for users based upon when and how users consume power. In this paper, we present a methodology on differentiated user services based on extracted characteristic consumer load shapes (usage profiles as a function of time) from a large smart meter data set. We identify distinct user subgroups based upon their actual historic usage patterns, which are represented by the proposed electricity usage distributions. Since the big electricity user data cover millions of users and for each user the data are multi-dimensional and in fine-time granularity, we thus propose a sublinear algorithm to make the computation of the differentiated user service model efficient. The algorithm requests an input of only a small portion of users, and a sublinear amount of the electricity data from each of these selected users. We prove that the algorithm provides performance guarantees. Our simulated evaluation demonstrates the effectiveness of our algorithm and the differentiating user service model.

In this paper we study energy conservation in the Internet. We observe that different traffic volumes on a link can result in different energy consumption; this is mainly due to such technologies as trunking (IEEE 802.1AX), adaptive link rates, etc. We design a green Internet routing scheme, where the routing can lead traffic in a way that is green. We differ from previous studies where they switch network components, such as line cards and routers, into sleep mode. We do not prune the Internet topology. We first develop a power model, and validate it using real commercial routers. Instead of developing a centralized optimization algorithm, which requires additional protocols such as MPLS to materialize in the Internet, we choose a hop-by-hop approach. It is thus much easier to integrate our scheme into the current Internet. We progressively develop three algorithms, which are loop-free, substantially reduce energy consumption, and jointly consider green and QoS requirements such as path stretch. We further analyze the power saving ratio, the routing dynamics, and the relationship between hop-by-hop green routing and QoS requirements. We comprehensively evaluate our algorithms through simulations on synthetic, measured, and real topologies, with synthetic and real traffic traces. We show that the power saving in the line cards can be as much as 50 percent.

TCP has been the dominant transport protocol for mobile internet since its origin. Its behaviors play an essential role in determining quality of service/experience (QoS and QoE) for mobile apps. While TCP has been extensively studied in a static, walking, or driving mobility, it has not been well explored in highspeed (> 200 km/h) mobility cases. With increasing investment and deployment of high speed rails (HSRs), a critical demand of understanding TCP performance under extremely high-speed mobility arises. In this paper, we conduct an in-depth study to investigate TCP behaviors on HSR. We collect 90 GB of measurement data on HSPA+ networks in Chinese high-speed trains with a peak speed of 310 km/h, along various routes (covering 5,000 km) during an 8-month period. We analyze the impacts of high-speed mobility and handoff on performance metrics including RTT, packet loss and network disconnection. Then we demystify the grand challenges posed on TCP operations (TCP establishment, transmission, congestion control and termination). Our study shows that performance greatly declines in HSR, where RTT spikes, packet drops and network disconnections are more significant and occur more frequently, compared with static, slowly moving or driving mobility cases. Moreover, TCP fails to adapt well to such extremely high-speed and yields severely abnormal behaviors, such as high spurious RTO rate, aggressive congestion window reduction, long delay of connection establishment and closure, and transmission interruption. All these findings indicate that extremely high-speed indeed poses a big threat to today's TCP and it calls for urgent efforts to develop HSR-friendly protocols and wireless networks to address even more complicated challenges raised by faster trains/aircrafts in the foreseeable future.

Radio Frequency based Wireless Power Transfer (RF-WPT) technology is recognized as a promising way to charge low-power wireless devices. But the application of RF-WPT in wireless sensor networks also introduces charging interference to wireless communications. The network lifetime maximization by jointly considering wireless charging and data transmission under interference concerns, however, has seldom been examined. In this paper, we take initial steps to consider communication and charger scheduling together in wireless sensor networks. We propose a smart interference-aware scheduling to maximize the network lifetime and avoid potential data loss caused by charging interference. The evaluation result indicates that the proposed design can guarantee 99% optimality and significantly improve network lifetime.

We propose an "α%-Green" network benchmark, whose spirit is that α% of the network power should come from the renewable energy. We argue that, as long as such a benchmark is satisfied, the network should always set its primary objective to its own optimization concerns.

Data traffic demand over the Internet is increasing rapidly, and it is changing the pricing model between Internet service providers (ISPs), content providers (CPs) and end users. One recent pricing proposal is sponsored data plan, i.e., when accessing contents from a particular CP, end users do not need to pay for that volume of traffic consumed, but the CP will sponsor for this data consumption. In this paper, our goal is to understand the rationale behind this new pricing model, as well as its impacts to the wireless data market, in particular, who will benefit and who will be hurt from this scheme. We build a two-class service model to analyze the consumers' traffic demand under the sponsored data plan with consideration of QoS. We use a two-stage Stackelberg game to characterize the interaction between CPs and the ISP and reveal a number of important findings. Our conclusions include: 1) When the ISP's capacity is sufficient, the sponsored data plan benefits consumers and CPs in the short run, but the ISP does not have incentives to further improve its service in the long run. 2) When ISP's capacity is insufficient, the ISP and end users may achieve a win- win trade, while the ISP and CPs always compete for the revenue. 3) The sponsored data plan may enlarge the un- balance in revenue distribution between different CPs; CPs with higher unit income and poorer technology support are more likely to prefer the sponsored data plan.

Accurate localization in wireless sensor networks is the foundation for many applications, such as geographic routing and position-aware data processing. In this paper, we develop a new localization protocol based on approximate convex decomposition (ACDL), with reliance on network connectivity information only. ACDL can calculate the node virtual locations for a large-scale sensor network with a complex shape. We first examine one representative localization algorithm and study the influential factors on the localization accuracy, including the sharpness of the angle at the concave point and the depth of the concave valley. We show that after decomposition, the depth of the concave valley becomes irrelevant. We thus define the concavity according to the angle at a concave point, which reflects the localization error. We then propose ACDL protocol for network localization. It consists of four main steps. First, convex and concave nodes are recognized and network boundaries are segmented. As the sensor network is discrete, we show that it is acceptable to approximately identify the concave nodes to control the localization error. Second, an approximate convex decomposition is conducted. Our convex decomposition requires only local information and we show that it has low message overhead. Third, for each convex section of the network, an improved MDS algorithm is proposed to compute a relative location map. Fourth, a fast and low complexity merging algorithm is developed to construct the global location map. Besides, by slight modification on the third step, we propose a variant of ACDL, denoted by ACDL-Tri, which is fully distributed and scalable while the localization accuracy is still comparable. We finally show the efficiency of ACDL by extensive simulations.

Recent years have witnessed cloud computing as an efficient means for providing resources as a form of utility. Driven by the strong demands, industrial pioneers have offered commercial cloud platforms, mostly datacenter-based, which are known to be powerful and effective. Yet, as the cloud customers are pure consumers, their local resources, though abundant, have been largely ignored. In this paper, We present SpotCloud, a real working system that seamlessly integrates the customers' local resources into the cloud platform, enabling them to sell, buy, and utilize these resources. We also investigate the potentials and challenges towards enabling customer-provided resources for cloud computing. Given that these local resources are highly heterogeneous and dynamic, we closely examine two critical challenges in this new context: (1) How can the customers be motivated to contribute or utilize such resources? and (2) How can high service availability be ensured out of the dynamic resources? We demonstrate a distributed market for potential sellers to flexibly and adaptively determine their resource prices through a repeated seller competition game. We also present an optimal resource provisioning algorithm that ensures service availability with minimized lease and migration costs. The evaluation results indicate it as a flexible and less expensive complement to the pure datacenter-based cloud.

Information-Centric Networking (ICN) provides substantial flexibility for users. One of the most important features of ICN is the universal in-network caching. The characteristics of ICN make it substantially different from traditional caching systems. In this paper we propose an age-based cooperative caching scheme in response to the special characteristics of ICN. We leverage the coupling between routing and caching in ICN to develop a light-weight collaboration mechanism that adaptively pushes popular contents to the network edge. We evaluate our approach using real traces and realistic network topology. Results show that our approach can significantly reduce network delay and traffic, and outperforms existing schemes.

Recently, the performance of mobile data networks has been evaluated from many aspects, e.g., TCP/IP protocols, comparison with WiFi or even satellite communication, under different movements within a metropolis area. Nevertheless, the result is still unknown in high-speed mobility scenarios and in a scale that crosses different metropolis and geographic areas. To fill in this blank, we carry out a comprehensive measurement study on the performance of mobile data networks under high-speed mobility, i.e., 300 km/h or above. Such speed is the current de facto standard of the China Railway High speed (CRH) network, the largest commercial high-speed railway network in the world so far. We first present an overview on the TCP performance over LTE networks. We observe that decent throughput may exist under high-speed mobility. However, comparing to the stationary and driving (100 km/h) scenarios, the throughput and RTT not only are worse, but also have a large variance. We then take an in-depth investigation into two key factors affecting the performance, i.e., The wireless channel and handoff. We believe our study on these factors is useful not only for TCP, but also for other upper-layer protocols.

MapReduce has emerged as a powerful tool for distributed and scalable processing of voluminous data. In this paper, we, for the first time, examine the problem of accommodating data skew in MapReduce with online operations. Different from earlier heuristics in the very late reduce stage or after seeing all the data, we address the skew from the beginning of data input, and make no assumption about a priori knowledge of the data distribution nor require synchronized operations. We examine the input in a continuous fashion and adaptively assign tasks with a load-balanced strategy. We show that the optimal strategy is a constrained version of online minimum makespan and, in the MapReduce context where pairs with identical keys must be scheduled to the same machine, there is an online algorithm with a provable 2-competitive ratio. We further suggest a sample-based enhancement, which, probabilistically, achieves a 3/2-competitive ratio with a bounded error.

With increasing demands for more flexible services, the routing policies in enterprise network becomes much richer. This has placed a heavy burden to the current router forwarding plane to support the increasing number of policies, primarily due to the limited capacity in TCAM. This hinders the development of new network services. In this paper, we present the design and implementation of a new forwarding table structure. It separates the functions of TCAM and SRAM and maximally utilizes the large & flexible SRAM. We progressively design a set of schemes, to maintain correctness, compress storage, and achieve line-card speeds. We also design incremental update algorithms that bring less accesses to memory. We present implementation designs and evaluate our scheme with a real implementation on a commercial router using real data sets. Our design does not require new devices. The evaluation results show that the performance of our forwarding tables is promising.

MapReduce has achieved tremendous success for large-scale data processing in data centers. A key feature distinguishing MapReduce from previous parallel models is that it interleaves parallel and sequential computation. Past schemes, and especially their theoretical bounds, on general parallel models are therefore, unlikely to be applied to MapReduce directly. There are many recent studies on MapReduce job and task scheduling. These studies assume that the servers are assigned in advance. In current data centers, multiple MapReduce jobs of different importance levels run together. In this paper, we investigate a schedule problem for MapReduce taking server assignment into consideration as well. We formulate a MapReduce server-job organizer problem (MSJO) and show that it is NP-complete. We develop a 3-approximation algorithm and a fast heuristic. We evaluate our algorithms through both simulations and experiments on Amazon EC2 with an implementation in Hadoop. The results confirm the advantage of our algorithms.

With the advances of bandwidth-intensive mobile devices, we see severe congestion problems in wireless data networks. Recently, research emerges to solve this problem from a pricing point of view. Time dependent pricing has been introduced, and initial investigations have shown its advantages over the conventional time independent pricing. Nevertheless, much is unknown in how a practical and effective time dependent pricing scheme can be designed. In this paper, we explore the design space of time dependent pricing. In particular, we focus on a number of schemes, e.g., the usage-based scheme, the flat-rate scheme, and a mixture of them which we called a cap scheme. Our findings include: 1) the ISP obtains a higher profit with usage-based (or flat-rate) scheme if the capacity is insufficient (or sufficient); 2) the usage-based scheme usually achieves a higher consumer surplus and more efficient traffic utilization than the flat-rate scheme; and 3) the cap scheme is strongly preferred by the ISP to further increase its revenue. We believe our findings provide important insights for ISPs to design effective pricing schemes.

In the application field using sensor networks to monitor valuable asset, source-location anonymity is a serious concern. As a series of event packets are reported to the base station, adversaries eavesdropping on the network can backtrack to the source through traffic analysis and the RF localization techniques. This leakage of contextual information will expose sensitive or precious objects and bring down the effectiveness of sensor networks. Existing techniques such as phantom routing or source simulation are proposed to discourage the adversaries, both of which trade energy for security. In this paper, we propose a new scheme, called path extension method (PEM), providing strong protection for source-location privacy. It performs quite well even though an object occurs near the base station, while other methods cannot protect the source well in this case. In PEM, fake sources are generated dynamically after the source sends event messages to the base station, which makes it much more flexible. Fake sources form several fake paths in the network and an adversary will be induced farther away from the source if it is entrapped by any of them. The theoretical and simulation results show that PEM is efficient in protecting source-location privacy with minimal message delivery delay and acceptable overhead.

In this paper we study energy conservation in the Internet. We observe that different traffic volumes on a link can result in different energy consumption; this is mainly due to such technologies as trunking (IEEE 802.1AX), adaptive link rates, etc. We design a green Internet routing scheme, where the routing can lead traffic in a way that is green. We differ from previous studies where they switch network components, such as line cards and routers, into sleep mode. We do not prune the Internet topology. We first develop a power model, and validate it using real commercial routers. Instead of developing a centralized optimization algorithm, which requires additional protocols such as MPLS to materialize in the Internet, we choose a hop-by-hop approach. It is thus much easier to integrate our scheme into the current Internet. We progressively develop three algorithms, which are loop-free, maximize energy conservation, and jointly consider green and QoS requirements such as path stretch. We comprehensively evaluate our algorithms through simulations on synthetic and real topologies and traffic traces. We show that the power saving in the line cards can be as much as 50%.

Recent advances in cloud-based big data analysis offers a convenient mean for providing an elastic and cost-efficient exploration of voluminous data sets. Following such a trend, industry leaders as Amazon, Google and IBM deploy various of big data systems on their cloud platforms, aiming to occupy the huge market around the globe. While these cloud systems greatly facilitate the implementation of big data analysis, their real-world applicability remains largely unclear. In this paper, we take the first steps towards a better understanding of the big data system on the cloud platforms. Using the typical MapReduce framework as a case study, we find that its pipeline-based design intergrades the computational-intensive operations (such as mapping/reducing) together with the I/O-intensive operations (such as shuffling). Such computational-intensive and I/O-intensive operations will seriously affect the performance of each other and largely reduces the system efficiency especially on the low-end virtual machines (VMs). To make the matter worse, our measurement also indicates that more than 90 % of the task-lifetime is in the shadow of such interference. This unavoidably reduces the applicability of cloud-based big data processing and makes the overall performance hard to predict. To address this problem, we re-model the resource provisioning problem in the cloud-based big data systems and present an interference-aware solution that smartly allocates the MapReduce jobs to different VMs. Our evaluation result shows that our new model can accurately predict the job completion time across different configurations and significantly improve the user experience for this new generation of data processing service.

No abstract available.

Energy is a global concern and the electricity bills nowadays are leading to unprecedented costs. Electricity price is market-based and dynamic. In this paper, we investigate how to cut the electricity bills of commercial buildings in a dynamic power market. The building thermal systems (e.g., air-conditioning), which dominate electricity bills, has a special property of thermal storage, i.e., the energy will not immediately dissipate from thermal air/water. Intuitively, with storage, the energy can be "stored" in the thermal system, making it possible to purchase electricity in low price and use it at appropriate time. The building thermal supply and electricity purchasing surely depends on human activities that the building should support such as class and meeting schedules. To minimize electricity bills, we develop a holistic planning of electricity purchasing schedule with thermal storage management, and appropriate room assignment schedules for classes/meetings usage. The computing algorithms require inputs of physical modeling on energy consumption. We develop wireless sensing systems to collect fine-grained data which are used to assist the cross-disciplinary physical modeling. We conduct validation through real experiments. We formulate an optimization problem and show that it is NP-complete. Our primary focus is to minimize electricity bills, which matches the incentives of the commercial buildings. We show that this does not coincide with energy conservation. We further investigate the relationship of minimization of electricity bills and minimization of energy consumption. We develop algorithms for our problem and our evaluation shows that we can achieve a 40% cost reduction.

The goal of this paper is to improve wireless AP caching by leveraging in-network caching. We observe that by treating routers as an in-network storage extension, we can relieve the storage limitation of APs. The unique challenge is that APs and routers cannot have a full collaboration, which makes the problem different from traditional cooperative caching problems. We study how APs can optimize caching decisions by using in-network caching information without controlling routers.

No abstract available.

Cloud offloading has recently attracted a substantial amount of attention from both industry and academia. This new generation of service, beyond conventional task scheduling and management, utilizes the abundant computation capacity from the public clouds and takes it as a part of external resource on the mobile devices. In this paper, we find that the cloud offloading can potentially increase the running latency of the offloaded tasks. Our measurement shows that the cloud offloading systems, such as Gaikai, can introduce up to 400ms communication latency between the customers and the remote offloading servers. This creates a severe bottleneck to offload the delay sensitive tasks. To mitigate such a challenge, we suggest Cloud Offloading on Customer-Provided Resources, which utilizes the local resources of the customers, such as their home PCs, to minimize the latency during the task offloading. We discuss the framework design based on our commercial system SpotCloud and propose a travel-aware protocol to further address the latency problems when the customers are traveling far from their home PCs. Our model analysis and trace-based simulation indicate that the task execution latency can be reduced by 60% when using the customer-provided offloading service. It is reasonable to believe that such a framework can serve as a useful complement to the conventional cloud offloading on pubic clouds.

Recently, many video sharing sites provide external links so that their video or audio contents can be embedded into external web sites. For example, users can copy the embedded URLs of the videos of YouTube and post the URL links on their own blogs. Clearly, the purpose of such function is to increase the distribution of the videos and the associated advertisement. Does this function fulfill its purpose and what is the quantification? In this paper, we provide a comprehensive measurement study and analysis on these external links to answer these two questions. With the traces collected from two major video sharing sites, YouTube and Youku of China, we show that the external links have various impacts on the popularity of the video sharing sites. More specifically, for videos that have been uploaded for eight months in Youku, around 15% of views can come from external links. Some contents are densely linked. For example, comedy videos can attract more than 800 external links on average. We also study the relationship between the external links and the internal links. We show that there are correlations; for example, if a video is popular itself, it is likely to have a large number of external links. Another observation we find is that the external links usually have a higher impact on Youku than that of YouTube. We conjecture that it is more likely that the external links have higher impact for a regional site than a worldwide site.

Providing field coverage is a key task in many sensor network applications. With unevenly distributed static sensors, quality coverage with acceptable network lifetime is often difficult to achieve. Fortunately, recent advances on embedded and robotic systems make mobile sensors possible, and we suggest that a small set of mobile sensors can be leveraged toward a cost-effective solution for field coverage. There are, however, a series of fundamental questions to be answered in such a hybrid network of static and mobile sensors: (1) Given the expected coverage quality and system lifetime, how many mobile sensors should be deployed? (2) What are the necessary coverage contributions from each type of sensors? (3) What working and moving patterns should the sensors adopt to achieve the desired coverage contributions? In this article, we offer an analytical study on these problems, and the results lead to a practical system design. Specifically, we present an optimal algorithm for calculating the contributions from different types of sensors, which fully exploits the potentials of the mobile sensors and maximizes the network lifetime. We then present a random walk model for the mobile sensors. The model is distributed with very low control overhead. Its parameters can be fine-tuned to match the moving capability of different mobile sensors and the demands from a broad spectrum of applications. A node collaboration scheme is then introduced to further enhance the system performance. We demonstrate through analysis and simulation that, in our mobile assisted design, a small set of mobile sensors can effectively address the uneven distribution of the static sensors and significantly improve the coverage quality.

We are in an age where people are paying increasing attention to energy conservation around the world. The heating and air-conditioning systems of buildings introduce one of the largest chunks of energy expenses. In this article, we make a key observation that after a meeting or a class ends in a room, the indoor temperature will not immediately increase to the outdoor temperature. We call this phenomenon thermal inertia . Thus, if we arrange subsequent meetings in the same room rather than in a room that has not been used for some time, we can take advantage of such undissipated cool or heated air and conserve energy. Though many existing energy conservation solutions for buildings can intelligently turn off facilities when people are absent, we believe that understanding thermal inertia can lead system designs to go beyond on-and-off-based solutions to a wider realm. We propose a framework for exploring thermal inertia in room management. Our framework contains two components. (1) The energy-temperature correlation model captures the relation between indoor temperature change and energy consumption. (2) The energy-aware scheduling algorithms: given information for the relation between energy and temperature change, energy-aware scheduling algorithms arrange meetings not only based on common restrictions, such as meeting time and room capacity requirement, but also energy consumptions. We identify the interface between these components so further works towards same on direction can make efforts on individual components. We develop a system to verify our framework. First, it has a wireless sensor network to collect indoor, outdoor temperature and electricity expenses of the heating or air-conditioning devices. Second, we build an energy-temperature correlation model for the energy expenses and the corresponding room temperature. Third, we develop room scheduling algorithms. In detail, we first extend the current sensor hardware so that it can record the electricity expenses in re-heating or re-cooling a room. As the sensor network needs to work unattendedly, we develop a hardware board for long-range communications so that the Imote2 can send data to a remote server without a computer relay close by. An efficient two-tiered sensor network is developed with our extended Imote2 and TelosB sensors. We apply laws of thermodynamics and build a correlation model of the energy needed to re-cool a room to a target temperature. Such model requires parameter calibration and uses the data collected from the sensor network for model refinement. Armed with the energy-temperature correlation model, we develop an optimal algorithm for a specified case, and we further develop two fast heuristics for different practical scenarios. Our demo system is validated with real deployment of a sensor network for data collection and thermodynamics model calibration. We conduct a comprehensive evaluation with synthetic room and meeting configurations, as well as real class schedules and classroom topologies of The Hong Kong Polytechnic University, academic calendar year of Spring 2011. We observe 20% energy savings as compared with the current schedules.

Data gathering in sensor networks is required to be efficient, adaptable and robust. Recently, compressive sensing (CS) based data gathering shows promise in meeting these requirements. Existing CS-based data gathering solutions require that a transform that best sparsifies the sensor readings should be used in order to reduce the amount of data traffic in the network as much as possible. As a result, it is very likely that different transforms have to be determined for varied sensor networks, which seriously affects the adaptability of CS-based schemes. In addition, the existing schemes result in significant errors when the sampling rate of sensor data is low (equivalent to the case of high packet loss rate) because CS inherently requires that the number of measurements should exceed a certain threshold. This paper presents STCDG, an efficient data gathering scheme based on matrix completion. STCDG takes advantage of the low-rank feature instead of sparsity, thereby avoiding the problem of having to be customized for specific sensor networks. Besides, we exploit the presence of the short-term stability feature in sensor data, which further narrows down the set of feasible readings and reduces the recovery errors significantly. Furthermore, STCDG avoids the optimization problem involving empty columns by first removing the empty columns and only recovering the non-empty columns, then filling the empty columns using an optimization technique based on temporal stability. Our experimental results indicate that STCDG outperforms the state-of-the-art data gathering algorithms in terms of recovery error, power consumption, lifespan, and network capacity.

ABSTRACT Network coding (NC) has showed to be beneficial to improve transmission performance in wireless mesh networks. Random linear coding is usually applied as the default coding schema. However, random linear coding causes significant decoding delay and jitter at receiver. Further, current NC does not support weight assignment to original packets, which is however indispensable for popular applications such as quality of service control and multipath media streaming in wireless mesh networks. Partial network coding (PNC) can largely reduce decoding delay and receiving fluctuation while keeping the benefit of NC. However, PNC does not support weight‐based data replacement and weight assignment to original packets. In this work, we propose weighted partial network coding (WPNC), which is a generalized coding schema of PNC. WPNC inherits all merits of PNC and part of NC. With WPNC, both decoding delay and receiving fluctuation will be reduced as observed in PNC. Also, WPNC is quite suitable for those applications that require weight assignment to original packets. After providing the whole framework of WPNC and thorough theoretical analysis to its performance, we have demonstrated how WPNC can be integrated with quality of service control and multipath routing supported media streaming in wireless mesh networks. Performance of WPNC is inter‐validated by both theoretical analysis and numeric evaluations. Copyright © 2011 John; Wiley & Sons, Ltd.

The Internet is a great hit in human history. However, it has evolved greatly from its original incarnation. Content distribution is playing a central role in today's Internet, which makes it difficult for the conventional host-to-host communication to meet the ever-increasing demands. In this paper, we present a novel “service and information oriented network architecture” (SIONA). The key aspect of SIONA is the name-based two-dimensional routing paradigm that provides scalable routing, caching and content delivery. We argue that SIONA solves the problems of mobile Internet by naturally supporting mobility, and provides network layer P2P for massive data distribution. Evaluation is conducted to investigate its caching and mobility performance.

Source address filtering is used as an important mechanism to prevent malicious traffic. Currently, most networks store filters in hardware such as TCAM, which has limited capacity, high power consumption and high cost. Although software can accommodate large number of filters, it needs multiple accesses to memory on the border router, which bears much more additional burden than other routers. In this paper, we propose a software-based mechanism for source address filtering. In our mechanism, we only need to check a few bits in source addresses on each router, rather than checking all bits on the ingress router. Through cooperation among routers, our mechanism ensures that malicious traffic will be filtered in the network. We formulate this problem as finding a cooperative scheme such that the loads on all routers are optimally balanced. We show that the problem can be optimally solved by dynamic programming. We evaluate our algorithms using comprehensive simulations with BRITE generated topologies and real world topologies. We conduct a case study on China Education and Research Network 2 (CERNET2) configurations, a large IPv6 network. Compared to checking 128-bit IP addresses on ingress routers, our algorithm checks at most 40 bits on each router.

Accurate localization in wireless sensor networks is the foundation for many applications, such as geographic routing and position-aware data processing. An important research direction for localization is to develop schemes using connectivity information only. These schemes primary apply hop counts to distance estimation. Not surprisingly, they work well only when the network topology has a convex shape. In this paper, we develop a new Localization protocol based on Approximate Convex Decomposition (ACDL). It can calculate the node virtual locations for a large-scale sensor network with arbitrary shapes. The basic idea is to decompose the network into convex subregions. It is not straight-forward, however. We first examine the influential factors on the localization accuracy when the network is concave such as the sharpness of concave angle and the depth of the concave valley. We show that after decomposition, the depth of the concave valley becomes irrelevant. We thus define concavity according to the angle at a concave point, which can reflect the localization error. We then propose ACDL protocol for network localization. It consists of four main steps. First, convex and concave nodes are recognized and network boundaries are segmented. As the sensor network is discrete, we show that it is acceptable to approximately identify the concave nodes to control the localization error. Second, an approximate convex decomposition is conducted. Our convex decomposition requires only local information and we show that it has low message overhead. Third, for each convex subsection of the network, an improved Multi-Dimensional Scaling (MDS) algorithm is proposed to compute a relative location map. Fourth, a fast and low complexity merging algorithm is developed to construct the global location map. Our simulation on several representative networks demonstrated that ACDL has localization error that is 60%-90% smaller as compared with the typical MDS-MAP algorithm and 20%-30% smaller as compared to a recent state-of-the-art localization algorithm CATL.

We are in an age where people are paying increasing attention to energy conservation around the world. The heating and air-conditioning systems of buildings introduce one of the largest chunk of energy expenses. In this paper, we make a key observation that after a meeting or a class ends in a room, the indoor temperature will not immediately increase to the outdoor temperature. We call this phenomenon Thermal Inertia. Thus, if we arrange subsequent meetings in the same room; than a room that has not been used for some time, we can take advantage of such un-dissipated cool or heated air and conserve energy. We develop a green room management system with three main components. First, it has a wireless sensor network to collect indoor, outdoor temperature and electricity expenses of the air-conditioning devices. Second, we build an energy-temperature correlation model for the energy expenses and the corresponding room temperature. Third, we develop room scheduling algorithms. Our system is validated with real deployment of a sensor network for data collection and thermodynamics model calibration. We conduct a comprehensive evaluation with synthetic room and meeting configurations. We observe a 30% energy saving as compared with the current schedules.

Wireless sensor networks have been widely suggested to be used in Cyber-Physical Systems for Structural Health Monitoring. However, for nowadays high-rise structures (e.g., the Guangzhou New TV Tower, peaking at 600m above ground), the extensive vertical dimension creates enormous challenges toward sensor data collection, beyond those addressed in state-of-the-art mote-like systems. One example is the data transmission from the sensor nodes to the base station. Given the long span of the civil structures, neither a strategy of long-range one-hop data transmission nor short-range hop-by-hop communication is cost-efficient. In this paper, we propose EleSense, a novel high-rise structure monitoring framework that uses elevators to assist data collection. In EleSense, an elevator is attached with the base station and collects data when it moves to serve passengers; as such, the communication distance can be effectively reduced. To maximize the benefit, we formulate the problem as a cross-layer optimization problem and propose a centralized algorithm to solve it optimally. We further propose a distributed implementation to accommodate the hardware capability of sensor nodes and address other practical issues. Through extensive simulations, we show that EleSense has achieved a significant throughput gain over the case without elevators and a straightforward 802.11 MAC scheme without the cross-layer optimization. Moreover, EleSense can greatly reduce the communication costs while maintaining good fairness and reliability. We also conduct a case study with real experiments and data sets on the Guangzhou New TV Tower, which further validates the effectiveness of our EleSense.

Sensor networks nowadays are widely used for structural health monitoring; for example, the sensor monitoring system deployed on the Guangzhou New TV Tower, China. While wired systems still dominate, it is commonly believed that wireless sensors will play a key role in the near future. One key difficulty for such systems is the data transmission from the sensor nodes to the base station. Given the long span of the civil structures, neither a strategy of long-range one-hop data transmission nor short-range hop-by-hop communication is cost-efficient. In this paper, we propose a novel scheme of using the elevators to assist data collection. A base station is attached to an elevator. A representative node on each floor collects and transmits the data to the base station using short range communication when the elevator stops at or passes by this floor. As such, communication distance can be minimized. To validate the feasibility of the idea, we first conduct an experiment in an elevator of the Guangzhou New TV Tower. We observe steady transmission when elevator is in movement. To maximize the gain, we formulate the problem as an optimization problem where the data traffic should be transmitted on time and the lifetime of the sensors should be maximized. We show that if we know the movement pattern of the elevator in advance, this problem can be solved optimally. We then study the online version of the problem and show that no online algorithm has a constant competitive ratio against the offline algorithm. We show that knowledge of the future elevator movement will intrinsically improve the data collection performance. We discuss how the information could be collected and develop online algorithms based on different level of knowledge of the elevator movement patterns. Theoretically, given that the links capacity assumptions we made, we can prove that our online algorithm can guarantee data delivery on time. In practice, we may set a buffer zone to minimize the possible data delivery violation. A comprehensive set of simulations and MicaZ testbed experiments have demonstrated that our algorithm substantially outperforms conventional multihop routing and naive waiting for elevator scheme. The performance of our online algorithm is close to the optimal offline solution.

Top-k query has long been a crucial problem in multiple fields of computer science, such as data processing and information retrieval. In emerging cyber-physical systems, where there can be a large number of users searching information directly into the physical world, many new challenges arise for top-k query processing. From the client's perspective, users may request different sets of information, with different priorities and at different times. Thus, top-k search should not only be multidimensional, but also be across time domain. From the system's perspective, data collection is usually carried out by small sensing devices. Unlike the data centers used for searching in the cyber-space, these devices are often extremely resource constrained and system efficiency is of paramount importance. In this paper, we develop a framework that can effectively satisfy demands from the two aspects. The sensor network maintains an efficient dominant graph data structure for data readings. A simple top-k extraction algorithm is used for user query processing and two schemes are proposed to further reduce communication cost. Our methods can be used for top-k query with any linear convex query function. The framework is adaptive enough to incorporate some advanced features; for example, we show how approximate queries and data aging can be applied. To the best of our knowledge, this is the first work for continuous multidimensional top-k query processing in sensor networks. Simulation results show that our schemes can reduce the total communication cost by up to 90 percent, compared with a centralized scheme or a straightforward extension from previous top-k algorithm on 1D sensor data.

We are interested in the sensor networks for scientific applications to cover and measure statistics on the sea surface. Due to flows and waves, the sensor nodes may gradually lose their positions; leaving the points of interest uncovered. Manual readjustment is costly and cannot be performed in time. We argue that a network of mobile sensor nodes which can perform self-adjustment is the best candidate to maintain the coverage of the surface area. In our application, we face a unique double mobility coverage problem. That is, there is an uncontrollable mobility, U-Mobility, by the flows which breaks the coverage of the sensor network. Moreover, there is also a controllable mobility, C-Mobility, by the mobile nodes which we can utilize to reinstall the coverage. Our objective is to build an energy efficient scheme for the sensor network coverage issue with this double mobility behavior. A key observation of our scheme is that the motion of the flow is not only a curse but should also be considered as a fortune. The sensor nodes can be pushed to some locations by the U-Mobility that potentially help to improve the overall coverage. With that taken into consideration, more efficient movement decision can be made. To this end, we present a dominating set maintenance scheme to maximally exploit the U-Mobility and balance the energy consumption among all the sensor nodes. We prove that the coverage is guaranteed in our scheme. We further propose a fully distributed protocol that addresses a set of practical issues. Through extensive simulation, we demonstrate that the network lifetime can be significantly extended, compared to a straightforward back-to-original reposition scheme.

No abstract available.

In the past few years, we have witnessed a flourish of online social network sites (OSNs). In this kind of websites, the users are not only information consumers, but also actively upload contents of their own to the OSNs. Being sharply different from the conventional sites, OSNs have attracted many studies recently. These studies, however, mainly focus on the social behaviors within OSNs, e.g., the user-user interaction, and the distribution of the time users spend on certain OSNs, etc. There are much fewer studies on the characteristics of the traffic patterns of OSNs upon the Internet infrastructure, even though it is important for the ISPs and the OSNs alike. One major difficulty is that it is not easy to obtain data from the private ISP backbone routers. In this paper, we collect data from two backbone routers of China Telecom and present the traffic patterns of several OSNs from the angle of the network layer. As well, we also compare the OSN traffic with three types of traditional websites like forums, search engines and news websites. We find that the traffic pattern in network layer is quite intuitively similar as what we can see in the application layer: users are the most active in OSNs, and are slightly less active in forums, and they are the least active in News websites. Besides that, we also can see that the traffic pattern in search engines is quite different from the other three.

Top-k query has long been an important topic in many fields of computer science. Efficient implementation of the top-k queries is the key for information searching. With the new frontier such as the cyber-physical systems, where there can be a large number of users searching information directly into the physical world, many new challenges arise for top-k query processing. From the client's perspective, different users may request different set of information, with different priorities and at different times. Thus, the top-k search not only should be multi-dimensional, but also across time domain. From the system's perspective, the data collection is usually carried out by small sensing devices. Unlike the data centers used for searching in the cyber-space, these devices are often extremely resource-constrained and system efficiency is of paramount importance. In this paper, we develop a framework that can effectively satisfy the two ends. The sensor network maintains an efficient dominant graph data structure for data readings. A simple top-k extraction algorithm is used for the user query processing and two schemes are proposed to further reduce communication cost. Our proposed methods can be used for top-k query with any linear convex query function. To the best of our knowledge, this is the first work for continuous multi-dimensional top-k query processing in sensor networks; and our simulation results show that our schemes can reduce the total communication cost by up to 90%, compared with the centralized scheme or a straightforward extension from previous top-k algorithm on one-dimensional sensor data.

In recent years, the core-net routing table, e.g., Forwarding Information Base (FIB), is growing at an alarming speed and this has become a major concern for Internet Service Providers. One effective solution for this routing scalability problem, which requires only upgrades on individual routers, is FIB aggregation. Intrinsically, IP prefixes with numerical prefix matching and the same next hop can be aggregated. Very commonly, all previous studies assume that each IP prefix has one corresponding next hop, i.e., towards one optimal path. In this paper, we argue that a packet can be delivered to its destination through a path other than the one optimal path. Based on this observation, we for the first time propose Nexthop-Selectable FIB Aggregation that is fundamentally different from all previous aggregation schemes. IP prefixes are aggregated if they have numerical prefix matching and share one common next hop. Consequently, IP prefixes that cannot be aggregated, due to lack of the same next hop, are aggregated; and we achieve a substantially higher aggregation ratio. In this paper, we provide a systematic study on this Nexthop-Selectable FIB Aggregation problem. We present several practical choices to build the sets of selectable next hops for the IP prefixes. To maximize the aggregation, we formulate the problem as an optimization problem. We show that the problem can be solved by dynamic programming. While the straightforward application of dynamic programming has exponential complexity, we propose a novel algorithm that is O(N). We then develop an optimal online algorithm with constant running time. We evaluate our algorithms through a comprehensive set of simulations with BRITE with RIBs collected from RouteViews. Our evaluation shows that we can reduce more than an order of the FIB size.

Recently wireless sensor networks have been widely suggested for Structural Health Monitoring. In such applications, diverse sensor nodes are deployed in a building structure, collecting ambient data such as temperature and strain from various locations and reporting them to a central base station for processing and diagnosing. For today's high-rise structures (e.g., the Guangzhou New TV Tower, a project that we have participated in, peaks at 600m above ground), the extensive vertical dimension creates enormous challenges toward sensor data collection, beyond those addressed in state-of-the-art motelike systems. For example, with a straightforward base station placement, a huge amount of data will accumulate as being relayed to the base station. As such, the sensor nodes close to the base station would quickly run out of energy for relaying the traffic. The accumulated traffic would also saturate the wireless medium, introducing significant interferences and collisions. The extensive height of these building structures, however, make elevators an indispensable component. This motivates us to develop EleSense, a novel high-rise structure monitoring framework that explores using elevators. In EleSense, an elevator is attached with the base station and collects data when it moves across different floors to serve passengers, which can effectively reduce the traffic accumulation and the collection delay. To maximally exploit the benefit, we take a unique angle with the cross-layer design. We present an abstraction of the high-rise structure monitoring problem that exploits elevators, and model it as a joint optimization across link scheduling, packet routing and end-to-end delivery. We propose a centralized algorithm to solve it optimally. We further propose a distributed implementation to accommodate the hardware capability of a sensor node and address other practical issues. We evaluate EleSense through ns-2 simulations and with real configurations from the Guangzhou New TV Tower. The results show that EleSense has a throughput gain of 30.4% to 200.6% over the case without elevators. We also observe a gain of 40.5% to 127.5% over a straightforward 802.11 MAC scheme without the cross-layer optimization. Moreover, EleSense can significantly reduce the communication costs while maintaining excellent fairness with reliable data delivering.

Wireless sensor networks have been widely used for ambient data collection in diverse environments. While in many such networks the nodes are randomly deployed in massive quantity, there is a broad range of applications advocating manual deployment. A typical example is structure health monitoring, where the sensors have to be placed at critical locations to fulfill civil engineering requirements. The raw data collected by the sensors can then be forwarded to a remote base station (the sink) through a series of relay nodes. In the wireless communication context, the operation time of a battery-limited relay node depends on its traffic volume and communication range. Hence, although not bounded by the civil-engineering-like requirements, the locations of the relay nodes have to be carefully planned to achieve the maximum network lifetime. The deployment has to not only ensure connectivity between the data sources and the sink, but also accommodate the heterogeneous traffic flows from different sources and the dominating many-to-one traffic pattern. Inspired by the uniqueness of such application scenarios, in this paper, we present an in-depth study on the traffic-aware relay node deployment problem. We develop optimal solutions for the simple case of one source node, both with single and multiple traffic flows. We show however that the general form of the deployment problem is difficult, and the existing only connectivity-guaranteed solutions cannot be directly applied here. We then transform our problem into a generalized version of the Euclidean Steiner Minimum Tree problem (ESMT). Nevertheless, we face further challenges as its solution is in continuous space and may yield fractional numbers of relay nodes, where simple rounding of the solution can lead to poor performance. We thus develop algorithms for discrete relay node assignment, together with local adjustments that yield high-quality practical solutions. Our solution has been evaluated through both numerical analysis and ns-2 simulations and compared with state-of-the-art approaches. The results show that for all test cases where the continuous space optimal solution can be computed within acceptable time frames, the network lifetime achieved by our solution is very close to the upper bound of the optimal solution (the difference is less than 13.5 percent). Moreover, it achieves up to 6-14 times improvement over the existing traffic-oblivious strategies.

Network survivability is an important topic for the Internet. To improve the performance of the Internet during failure, IP Fast Reroute (IPFRR) mechanisms are proposed to establish backup routes for failure-affected packets. NotVia, a most prominent one, provides 100% protection coverage for single-node failures. However, it brings in nontrivial computing and memory pressure to routers with special NotVia addresses, in which only some are necessary for a specific router. Besides, the protection path of NotVia is 20% longer than the optimal path on average. In this paper, we propose early decapsulated NotVia (ED-NotVia) handling the aforementioned problems and thus making NotVia more practical. We first analyze the properties of necessary NotVia addresses to any specific node. Then we develop a heuristic Nec-NotVia Algorithm for a node to find the necessary NotVia addresses and compute routes for them, where unnecessary addresses are eliminated. Based on this elimination, early decapsulation is imported to optimize the protection path with marginal overhead. We evaluate our algorithm and demonstrate the effectiveness of ED-NotVia using topologies from Rocketfuel and Brite. The results show that 1) only 5% to 20% of SPT(Shortest Path Tree)-related NotVia addresses (1.23% to 6.41% of all the NotVia addresses) in an AS are necessary for a node; 2) by computing the routes for 15% to 40% SPT-related NotVia addresses, ED-NotVia provides 98% protection coverage; and 3) the protection path stretch ratio of ED-NotVia is only 1.03 on average as compared to 1.20 for NotVia.

There are heavy studies recently on applying wireless sensor networks for structural health monitoring. These works usually focus on the computer science aspect, and the considerations include energy consumption, network connectivity, etc. It is commonly believed that for the current resource limited wireless sensors, system design could be more efficient if the application requirements are incorporated. Nevertheless, we often find that, rather than integration, assumptions have to be made due to lack of knowledge of civil engineering; for example, to evaluate routing algorithms, the sensor placement is assumed to be random or on grids/trees. These may not be practically meaningful to the respective application demands, and make the great efforts by the computer science community on developing efficient methods from the sensor network aspect less useful. In this paper, we study the very first problem of the SHM systems: the sensor placement and focus on the civil requirements. We first study the current general framework of structure health monitoring. We redevelop the framework that includes a new sensor placement module. This module implements the most widely accepted sensor placement scheme from civil engineering but focusing on its usefulness for computer science. It provides such interfaces that can rank the placement quality of the candidate locations in a step by step manner. We then optimize system performance by considering network connectivity and data routing issues; with the objective on energy efficiency. We evaluate our scheme using the data from the structural health monitoring system on the Ting Kau Bridge, Hong Kong. We show that a uniform and a state-of-the-art placement are not very meaningful in placement quality. Our scheme achieves almost the same sensor placement quality with that of the civil engineering with five-fold improvement in system lifetime. We conduct an experiment on the in-built Guangzhou New TV Tower, China; and the results validate the effectiveness of our scheme.

In this paper, we develop a novel framework that attempts to reduce network traffic for error-bounded data collection in wireless sensor networks. In many sensor applications, it is acceptable that the monitoring results evaluated based on collected data might deviate from the exact results; as long as the error is bounded by a certain threshold. One well-known technique for error-bounded data collection is data filtering, which explores temporal data correlation to suppress data updates. A concrete scheme was proposed in. The data collection is divided in rounds. A filter is installed on each node and the total filter size is constrained by the user-specified error budget. Intuitively, if the data change from the last update report is smaller than the filter size, the current update is suppressed, i.e., not to report to the base station. To adapt to system dynamics, the sizes of all filters are periodically shrunk and the left-over budget is re-allocated to the node with the highest load. Many follow up studies can be found.

Recently, many video sharing sites provide external links so that their video or audio contents can be embedded into external web sites. For example, users can copy the embedded URLs of the videos of YouTube and post on their own blogs. Clearly, the purpose of such functionality is to increase the distribution of the videos and the associated advertisement. In this paper, we provide a comprehensive measurement study and analysis on these external links. With the traces collected from two major VOD sites, YouTube and Youku of China, we show that the external links have various impact on the popularity of the VOD sites. More specifically, for videos that have been uploaded for eight months in Youku, around 15% of views can come from external links. Some contents are densely linked, for example, the comedy videos can attract more than 800 external links on average. We also study the relationship between the external links and the internal links. We show that there are correlations; for example, if a video is popular itself, it will likely have a larger number of external links. Another observation is that we always find that the external links of Youku usually have a higher impact than that of YouTube. We conjecture that a more regional site may enjoy a relatively higher impact from the external links.

Many sensor network applications are tightly coupled with the geometric environment where the sensor nodes are deployed. The topological skeleton extraction for the topology has shown great impact on the performance of such services as location, routing, and path planning in wireless sensor networks. Nonetheless, current studies focus on using skeleton extraction for various applications in wireless sensor networks. How to achieve a better skeleton extraction has not been thoroughly investigated. There are studies on skeleton extraction from the computer vision community; their centralized algorithms for continuous space, however, are not immediately applicable for the discrete and distributed wireless sensor networks. In this paper, we present a novel Connectivity-bAsed Skeleton Extraction (CASE) algorithm to compute skeleton graph that is robust to noise, and accurate in preservation of the original topology. In addition, CASE is distributed as no centralized operation is required, and is scalable as both its time complexity and its message complexity are linearly proportional to the network size. The skeleton graph is extracted by partitioning the boundary of the sensor network to identify the skeleton points, then generating the skeleton arcs, connecting these arcs, and finally refining the coarse skeleton graph. We believe that CASE has broad applications and present a skeleton-assisted segmentation algorithm as an example. Our evaluation shows that CASE is able to extract a well-connected skeleton graph in the presence of significant noise and shape variations, and outperforms the state-of-the-art algorithms.

For the recent two years, YouTube has already become one of the most successful VOD systems. Since it is established in early 2005, the number of the visits to YouTube not only maintain a high level, but also steadily increases even these days. It is widely believed that the online social networks contribute to this success. To study one new aspect of how much social networks contribute to the popularity of videos in YouTube, this paper provides an in-depth study into the effects of the external links in YouTube. Our study shows interesting characteristics of external links in YouTube. The existed studies reveal that `Music' and `Comedy' videos are the most popular ones. Compared with this, our study shows that the `Sports' and `Science & Technology' videos gains the most largest proportion of views from external links. Also, the percentage of views from external links reduces with power law. More specifically, we find that the top-5 external links contribute the popularity of the video most in the first four days and percentage can be up to 25%. For the video over than one years old, the top-5 external links are also able to contribute a stable part at about 4.5%. To end this, we hope our work can serve as a initial step for the study of the external environment.

In recent years, there are substantial demands to reduce packet loss in the Internet. Among the schemes proposed, finding backup paths in advance is considered to be an effective method to reduce the reaction time. Very commonly, a backup path is chosen to be a most disjoint path from the primary path, or in the network level, backup paths are computed for all links (e.g., IPRFF). The validity of this straightforward choice is based on 1) all the links may fail with equal probability; and 2) facing the high protection requirement today, having links not protected or sharing links between the primary and backup paths just simply look weird. Nevertheless, indications from many research studies have confirmed that the vulnerability of the links in the Internet is far from equality. In addition, we have seen that full protection schemes may introduce high costs. In this paper, we argue that such approaches may not be cost effective. We first analyze the failure characteristics based on real world traces from CERNET2, the China education and Research NETwork 2. We observe that the failure probabilities of the links is heavy-tail, i.e., a small set of links caused most of the failures. We thus propose a selective protection scheme. We carefully analyze the implementation details and the overhead for general backup path schemes of the Internet today. We formulate an optimization problem where the routing performance (in terms of network level availability) should be guaranteed and the backup cost should be minimized. This cost is special as it involves computation overhead. Consequently, we propose a novel Critical-Protection Algorithm which is fast itself. We evaluate our scheme systematically, using real world topologies and randomly generated topologies. We show significant gain even when the network availability requirement is 99.99\% as compared to that of the full protection scheme.

Many sensor network applications are tightly coupled with the geometric environment where the sensor nodes are deployed. The topological skeleton extraction has shown great impact on the performance of such services as location, routing, and path planning in sensor networks. Nonetheless, current studies focus on using skeleton extraction for various applications in sensor networks. How to achieve a better skeleton extraction has not been thoroughly investigated. There are studies on skeleton extraction from the computer vision community; their centralized algorithms for continuous space, however, is not immediately applicable for the discrete and distributed sensor networks. In this paper we present CASE: a novel connectivity-based skeleton extraction algorithm to compute skeleton graph that is robust to noise, and accurate in preservation of the original topology. In addition, no centralized operation is required. The skeleton graph is extracted by partitioning the boundary of the sensor network to identify the skeleton points, then generating the skeleton arcs, connecting these arcs, and finally refining the coarse skeleton graph. Our evaluation shows that CASE is able to extract a well-connected skeleton graph in the presence of significant noise and shape variations, and outperforms state-of-the-art algorithms.

We are interested in the sensor networks for scientific applications to cover and measure statistics on the sea surface. Due to flows and waves, the sensor nodes may gradually lose their positions; leaving the points of interest uncovered. Manual readjustment is costly and cannot be performed in time. We argue that a network of mobile sensor nodes which can perform self-adjustment is the best candidate to maintain the coverage of the surface area. In our application, we face a unique double mobility coverage problem. That is, there is an uncontrollable mobility, U-Mobility, by the flows which breaks the coverage of the sensor network. Moreover, there is also a controllable mobility, C-Mobility, by the mobile nodes which we can utilize to reinstall the coverage. Our objective is to build an energy efficient scheme for the sensor network coverage issue with this double mobility behavior. A key observation of our scheme is that the motion of the flow is not only a curse but should also be considered as a fortune. The sensor nodes can be pushed by free to some locations that potentially help to improve the overall coverage. With that taken into consideration, more efficient movement decision can be made. To this end, we present a dominating set maintenance scheme to maximally exploit the U-Mobility and balance the energy consumption among all the sensor nodes. We prove that the coverage is guaranteed in our scheme. We further propose a fully distributed protocol that addresses a set of practical issues. Through extensive simulation, we demonstrate that the network lifetime can be significantly extended, compared to a straight forward back-to-original reposition scheme.

There is a growing interest in using wireless sensor networks for security monitoring in the underground coal mines. In such applications, the sensor nodes are deployed to detect interested events, e.g., the density of certain gas at some locations is higher than the predefined threshold. These events are then reported to the base station outside. Using conventional multi-hop routing for data reporting, however, will result in imbalance of energy consumption among the sensors. Even worse, the unfriendly communication condition underground makes the multi-hop data transmission challenging, if not impossible. In this paper, we thus propose to leverage tramcars as mobile sinks to assist event collection and delivery. We further observe that the sensor readings have spatial and temporal correlation. More precisely, the same event may be observed by multiple neighboring sensor nodes and/or at different time. Obviously, it can be more energy-efficient if the data are selectively reported. As such, we first provide a general, yet realistic definition on the events. We then transform the event collection problem into a set coverage problem; and our objective is to maximize the system lifetime with the coverage rate of events guaranteed. We show that the problem is NP-hard even when all the events are known in advance. We present an online scheme which leverages the spatial-temporal correlation of the events to balance the communication energy of the static sensor nodes. We prove that the expected event coverage rate can be guaranteed in theory. Through extensive simulation, we demonstrate that our scheme can significantly extend system lifetime, as compared to a stochastic collection scheme.

Wireless sensor networks have been widely used for ambient data collection in diverse environments. While in many such networks the sensor nodes are randomly deployed in massive quantity, there is a broad range of applications advocating manual deployment. A typical example is structure health monitoring, where the sensors have to be placed at critical locations to fulfill civil engineering requirements. The raw data collected by the sensors can then be forwarded to a remote base station (the sink) through a series of relay nodes. In the wireless communication context, the operation time of a battery-limited relay node depends on its traffic volume and communication range. Hence, although not bounded by the civil-engineering-like requirements, the locations of the relay nodes have to be carefully planned to achieve the maximum network lifetime. The deployment has to not only ensure connectivity between the data sources and the sink, but also accommodate the heterogeneous traffic flows from different sources and the dominating many-to-one traffic pattern. Inspired by the uniqueness of such application scenarios, in this paper, we present an in-depth study on the traffic-aware relay node deployment problem. We develop optimal solutions for the simple case of one source node, both with single and multiple traffic flows. We show however that the general form of the deployment problem is difficult, and the existing connectivity-guaranteed solutions cannot be directly applied here. We then transform our problem into a generalized version of the Euclidean Steiner minimum tree problem (ESMT). Nevertheless, we face further challenges as its solution is in continuous space and may yield fractional numbers of relay nodes, where simple rounding of the solution can lead to poor performance. We thus develop algorithms for discrete relay node assignment, together with local adjustments that yield high-quality practical solutions. Our solution has been evaluated through both numerical analysis and ns-2 simulations and compared with state-of-the-art approaches. The results show that it achieves up to 6 to 14 times improvement on the network lifetime over the existing traffic-oblivious strategies.

In wireless sensor networks, filters, which suppress data update reports within predefined error bounds, effectively reduce the traffic volume for continuous data collection. All prior filter designs, however, are stationary in the sense that each filter is attached to a specific sensor node and remains stationary over its lifetime. In this paper, we propose mobile filter, a novel design that explores migration of filters to maximize overall traffic reduction. A mobile filter moves upstream along the data collection path, with its residual size being updated according to the collected data. Intuitively, this migration extracts and relays unused filters, leading to more proactive suppressing of update reports. We start by presenting an optimal filter migration algorithm for a chain topology. The algorithm is then extended to general multi-chain and tree topologies. Extensive simulations demonstrate that, for both synthetic and real data traces, the mobile filtering scheme significantly reduces data traffic and extends network lifetime against a state-of-the-art stationary filtering scheme.

In wireless sensor networks, filters, which suppress data update reports within predefined error bounds, effectively reduce the traffic volume for continuous data collection. All prior filter designs, however, are stationary in the sense that each filter is attached to a specific sensor node and remains stationary over its lifetime. In this paper, we propose mobile filter, a novel design that explores migration of filters to maximize overall traffic reduction. A mobile filter moves upstream along the data collection path, with its residual size being updated according to the collected data. Intuitively, this migration extracts and relays unused filters, leading to more proactive suppressing of update reports. While extra communications are needed to move filters, we show through probabilistic analysis that the overhead is outrun by the gain from suppressing more data updates.

Wireless sensor networks have been widely used for surveillance in harsh environments. In many such applications, the environmental data are continuously sensed, and data collection by a server is only performed occasionally. Hence, the sensor nodes have to temporarily store the data, and provide easy and on-hand access for the most updated data when the server approaches. Given the expensive server-to-sensor communications, the large amount of sensors and the limited storage space at each tiny sensor, continuous data collection becomes a challenging problem. In this article, we present partial network coding (PNC) as a generic tool for these applications. PNC generalizes the existing network coding (NC) paradigm, an elegant solution for ubiquitous data distribution and collection. Yet PNC allows efficient storage replacement for continuous data, which is a deficiency of the conventional NC. We prove that the performance of PNC is quite close to NC, except for a sub-linear overhead on storage and communications. We then address a set of practical concerns toward PNC-based continuous data collection in sensor networks. Its feasibility and superiority are further demonstrated through simulation results.

Media distribution through application-layer overlay networks has received considerable attention recently, owing to its flexibility and readily deployable nature. On-demand streaming with asynchronous requests and, in general, with VCR-like interactions nevertheless remains a challenging task in overlay networks. In this paper, we introduce the dynamic skip list (DSL), a novel randomized and distributed structure that inherently accommodates dynamic and asynchronous clients. We establish the theoretical foundations of the DSL and demonstrate a practical DSL-based streaming overlay. In this overlay, the costs for typical operations, including join, leave, fast-forward, rewind, and random seek, are all sublinear to the client population. The model also seamlessly integrates a smart data scheduling algorithm using linear network coding, yielding fast and robust downloading from multiple suppliers. Our simulation results show that the DSL-based overlay is highly scalable. It delivers reasonably smooth playback with diverse client interactivities while keeping the computation and bandwidth overheads low.

In many sensor network applications, manual or well-controlled node deployment is not practical. Random deployments, e.g., from the air, often result in unevenly distributed nodes and hence insufficient coverage. In this paper, we present a novel mobility-assisted network for field coverage, and suggest that, by introducing only a small set of mobile sensors, the field coverage can be remarkably improved. The main challenges for this mobility-assisted framework are determining the necessary coverage contribution of the mobile sensors and achieving such contribution. We develop an optimal strategy for separating the contribution from static sensors and mobile sensors. We then proposed a random walk based mobility model for the mobile sensors to achieve their necessary contribution. We demonstrate through experiment that a small set of mobile sensors can effectively solve the unbalance of the sensor distribution and significantly improve the system lifetime and coverage quality.

Providing field coverage is a key issue in many sensor network applications. For a field with unevenly distributed static sensors, a quality coverage with acceptable network lifetime is often difficult to achieve. We propose a hybrid network that consists of both static and mobile sensors, and we suggest that it can be a cost-effective solution for held coverage. The main challenges of designing such a hybrid network are, first, determining necessary coverage contributions from each type of sensors; and second, scheduling the sensors to achieve the desired coverage contributions, which includes activation scheduling for static sensors and movement scheduling for mobile sensors. In this paper, we offer an analytical study on the above problems, and the results also lead to a practical system design. Specifically, we present an optimal algorithm for calculating the contributions from different types of sensors, which fully exploits the potentials of the mobile sensors and maximizes the network lifetime. We then present a random walk model for the mobile sensors. The model is distributed with very low control overhead. Its parameters can be fine-tuned to match the moving capability of different mobile sensors and the demands from a broad spectrum of applications. A node collaboration scheme is then introduced to further enhance the system performance. We demonstrate through analysis and simulation that, in our hybrid design, a small set of mobile sensors can effectively address the uneven distribution of the static sensors and significantly improve the coverage quality.

Wireless sensor networks have recently been suggested for many surveillance applications, such as object monitoring, path protection, or area coverage. Since the sensors themselves are important and critical objects in the network, a natural question is whether they need certain level of protection, so as to resist the attacks targeting on them directly. If this is necessary, then who should provide this protection, and how it can be done? We refer to the above problem as self-protection , as we believe the sensors themselves are the best (and often the only) candidates to provide such protection. In this article, we for the first time present a formal study on the self-protection problems in wireless sensor networks. We show that, if we simply focus on enhancing the quality of field or object covering, the sensors might not necessarily be self-protected, which in turn makes the system extremely vulnerable. We then investigate different forms of self-protections, and show that the problems are generally NP-complete. We develop efficient approximation algorithms for centrally controlled sensors. We further extend the algorithms to fully distributed implementation, and introduce a smart sleep-scheduling algorithm that minimizes the energy consumption.

Wireless sensor networks have recently been suggested for many surveillance applications such as object monitoring, path protection, or area coverage. Since the sensors themselves are important and critical objects in the network, a natural question is whether they need certain level of protection, so as to resist the attacks targeting on them directly. If this is necessary, then who should provide this protection, and how it can be achieved? We refer to the above problem as self-protection, as we believe the sensors themselves are the best (and often the only) candidate to provide such protection. In this papel; we for the jirst time present a formal study on the selfprotection problem in wireless sensor networks. We show that, if we simply focus on the quality ofjield or object covering, the sensors might not necessarily be self-protected, which in turn makes the system vulnerable. We then investigate dzreerent forms of self-pmtections, and show that the problems are generally NP-complete. We develop eficient approximation algorithms for centrally-controlled sensors. We then extend the algorithms to filly distributed implementation, and introduce a smart sleep-scheduling algorithm that minimize the energy consumption.

Media distribution through application-layer overlay networks has received considerable attention recently, owing to its flexibility and readily deployable nature. On-demand streaming with asynchronous requests, and in general, with VCR-like interactions, nevertheless remains a challenging task. In this paper, we introduce the skip list, a novel randomized and distributed structure that inherently accommodates dynamic and asynchronous clients. We demonstrate a practical skip list based streaming overlay with typical VCR operations. Our simulation results show that the skip list based overlay is highly scalable, with smooth playback for diverse interactivities, and low overheads

Wireless sensor networks have been widely used for surveillance in harsh environments. In many such applications, the environmental data are continuously sensed, and data collection by a server is only performed occasionally. Hence, the sensor nodes have to temporarily store the data, and provide easy and on-hand access for the most updated data when the server approaches. Given the expensive server-to-sensor communications, the large amount of sensors and the limited storage space at each tiny sensor, continuous data collection becomes a challenging problem. In this paper, we present partial network coding (PNC) as a generic tool for the above applications. PNC generalizes the existing network coding (NC) paradigm, an elegant solution for ubiquitous data distribution and collection. Yet, PNC enables efficient storage replacement for continuous data, which is a major deficiency of the conventional NC. We prove that the performance of PNC is quite close to NC, except for a sub-linear overhead on storage and communications. We then address a set of practical concerns toward PNC-based continuous data collection in sensor networks. Its feasibility and superiority are further demonstrated through simulation results

Current approaches to providing robust network connections which are tolerant to failures involve restoration schemes which mainly focus on reserving backup paths. In this paper we propose a technique for which avoids the extra cost for reserving, by pre-identifying (but not reserving) the backup paths. We present and analyze an algorithm to solve this problem and study a practical special case in detail. Through simulations we show that our model is significantly more cost-efficient than backup path reservation. We also show how this model can fit into the MPLS architecture

Sensor networks are powerful tools for performing monitoring and surveillance tasks over large areas. A sensor is a cheap, simple device with low power and limited capabilities. In a sensor network a large number of sensors are deployed to span the whole area to be monitored. Due to the simplicity and the large quantity of the sensors involved, collecting data from a sensor network can be time and energy inefficient. In this paper, we investigate making the data gathering task from a sensor network more efficient by using a randomized, layered architecture. The layers in our architecture are constructed in a distributed fashion, with each sensor deciding locally on what layers it will exist. The key property of our technique is that the information is collected from one layer of the architecture containing a small subset of the sensors, resulting in fewer hops and thus smaller data in data aggregation. We provide provably correct results for the delay incurred and the accuracy of the results. In the context of our new techniques, we also explore ways to speed up the data gathering process even further, such as using history information. In addition, we consider how to optimize the structure of our system so that the energy consumption will be evenly distributed among each sensor, thus extending the overall lifetime of the entire network.