The hardware scheduler on NVIDIA GPUs is highly inefficient in utilizing micro-architectural hardware resources. It places blocks from the same kernel within the same GPU Streaming Multiprocessor (SM) core, resulting in a stacking colocating problem, where identical blocks are placed within the same SM core, saturating only a subset of intra-SM hardware resources while leaving others underutilized. The primary challenge in addressing this issue is that the NVIDIA hardware is closed-source, preventing us from directly modifying the hardware scheduler. To bridge the semantic gap between the resource demands of kernels and the scheduler, we introduce µ Share, which enables intra-SM scattered colocating of kernels through a non-intrusive half-plus blocksize shaping method. It shapes the blocksize of kernels to a halfplus blocksize (i.e., slightly more than half of the SM's thread capacity), scattering identical blocks of the same kernel across different SMs. It further adopts a time-shifted launching method to reduce intra-SM resource contention. Compared to state-of-the-art systems, µ Share does not require intrusive modifications to hardware or kernel code, yet it can still improve inference throughput by 26.90%-54.09% and increases low-level hardware utilization by 38.53%-61.15%.

Abstract The increasing popularity of container technology raises significant challenges in efficiently storing millions of container images in registries to enable fast on-demand image pulling. This is further complicated by (1) registries are geographically distributed, with independent and heterogeneous storage resources; (2) container images are pulled in layers, but can be stored at different levels of granularity, i.e., layer-level or file-level, each with varying storage requirement and pulling latency. To address the above challenges, we propose MIS, a multi-granularity image storage strategy, for distributed registries to determine the storage granularity and schedule image storage collaboratively, aiming to reduce the image pulling latency while improving the storage utilization. We formulate the image storage problem into a nonlinear mixed-integer programming form with NP-hardness by incorporating both layer-level and file-level storage constraints. We propose a low computational complexity algorithm via randomized rounding with a guaranteed approximation ratio. Extensive experimental results demonstrate the effectiveness of our strategy, with image pulling latency reductions of 28.67%, 21.69%, and 28.94% respectively compared to the state-of-the-art solutions.

Low Earth Orbit (LEO) satellite mobile edge computing (SMEC) has emerged as a pivotal technology for delivering low-delay communication and computational services to under-served regions. However, ensuring sustainable operation of SMEC systems remains challenging due to limited onboard energy and battery degradation, which is critically influenced by the Depth of Discharge (DoD) of battery. This paper investigates the collaborative DoD optimization problem in heterogeneous SMEC, formulating it as a long-term stochastic optimization aimed at minimizing DoD while maintaining system stability under dynamic energy supply and stochastic task arrivals. To address this problem, we propose a Lyapunov-guided optimization framework that integrates Lyapunov optimization with a multi-agent deep reinforcement learning algorithm. Specifically, we employ Lyapunov optimization to transform the long-term objective into a sequence of per-time-slot subproblems. Subsequently, given the high complexity of minimizing the derived Lyapunov drift function directly, we utilize a Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm to solve these subproblems. This approach leverages MADDPG’s proven capability in handling complex sequential decision-making processes while enabling real-time adaptive optimization in dynamic environments. Extensive simulations demonstrate that the proposed framework achieves significant reduction in DoD, bounded delay, and stable queue dynamics in diverse scenarios.

Microservices are widely shared in production user-facing applications. These shared microservices have various resource usage patterns when queries from different call graphs of different services access them. However, existing microservice management works fail to efficiently scale resources for them, mainly due to the lack of fine-grained scheduling of diverse queries. We therefore propose Delphinus , a runtime system that efficiently manages resources for shared microservices while ensuring the Quality-of-Service (QoS). Delphinus comprises a group-oriented query scheduler and a borrowing-based load adapter . The query scheduler identifies diverse queries, groups the containers of shared microservices, and schedules the queries into separate groups. The load adapter efficiently scales resources for shared microservices, and fully utilizes the idle containers among groups when the loads of diverse queries change. Results show that Delphinus reduces CPU and memory usage by 40.1% and 36.4% for shared microservices, respectively, compared to state-of-the-art works.

Modern network-centric operations increasingly rely on multi-layer Kill Webs (KWs), enabling redundant and non-linear sensing-to-strike pathways while introducing a combinatorial equipment selection problem under uncertainty and resource constraints. This paper formulates a multi-layer KW equipment combination selection as a sequential decision-making problem by explicitly modeling heterogeneous equipment capabilities, resource constraints, and the network topology. To address this problem, we developed an RL learning-based optimization framework, where a multi-objective reward function integrates normalized relevance, operational risk, and timeliness, with a penalty mechanism for infeasible or incomplete kill-chain closure. Based on the jointly captured state information (e.g., network structure, equipment attributes, target characteristics, and resource availability), an Actor-Critic (AC) algorithm is developed to learn adaptive equipment combination selection across different operational stages using temporal-difference advantage estimation and entropy regularization. Simulation results under diverse battlefield scenarios demonstrate that the proposed framework consistently outperforms Deep Q-Network (DQN), Proximal Policy Optimization (PPO), and Particle Swarm Optimization (PSO), achieving at least a 19.6% improvement in overall operational effectiveness while maintaining low decision latency.

Federated learning, as a compelling machine learning framework, enables collaborative model training without exposing private data. However, the excessive communication overhead remains a major challenge. To tackle this challenge, hierarchical federated edge learning (HFEL) framework has been proposed for reducing the communication load via migrating the model aggregation partially from cloud to edge servers. Although HFEL has significant potential, it is still constrained by end-devices with limited computational capabilities and unfavorable network conditions. A common approach to reduce this effect is to involve only the fastest end-devices in the training process. But because only parts of end-devices' data samples can be selected by such means, it damages the diversity of training data, and hence greatly affects the model's quality. In addition, for further improving the training performance, a proper bandwidth allocation strategy is also needed to make full use of the shared network resource of edge servers. To this end, we proposeDAHFF, aDiversity-AwareHierarchicalFastFederated learning framework consisting ofVirtual Queue based Device Selectionphase andBinary Search based Bandwidth Allocation, which are responsible for selecting participated end-devices and allocating bandwidth for selected devices, respectively. Extensive experiments on different deep learning models show that our proposed framework can averagely speed up the training performance by$2.07\times$in comparison with state-of-the-art approaches.

Vehicular named data networking (VNDN), which integrates the principles of named data networks with vehicular ad hoc networks, represents a promising paradigm for future intelligent transportation systems. Nevertheless, VNDN faces significant hurdles, including broadcast storms from excessive interest packet flooding and reverse-path disruptions due to high vehicular mobility. To address these challenges, we introduce an adaptive forwarding with path optimization method. First, a dynamic caching algorithm is designed to optimize roadside unit storage efficiency and maximize cache hit rates. Second, a gated recurrent unit-based adaptive data forwarding mechanism is introduced to dynamically select optimal forwarders and preserve reverse paths via decentralized heartbeat detection and interface remapping, improving link reliability. Simulation outcomes demonstrate that the proposed approach significantly lowers data retrieval delays while curbing overall communication overhead.

Graph neural networks (GNNs) are gaining popularity in diverse application domains and growing in complexity.As a result, it is crucial to achieve high-performance GNN execution.Among various techniques, algebraic transformations, including operator reordering and operator fusion, have been successfully applied to improve the computation and memory access efficiencies of DNN models.However,

While GNN-based services often experience load fluctuation, applying serverless computing to serve GNN inference reduces the cost and allows elastic resource scaling. However, GNN serverless shows poor performance due to heavy data fetching latency and long cold startup overhead, and our observation indicates opportunities for reducing data redundancy and mitigating cold startup latency. In this paper, we present FaaSGNN, a serverless GNN inference framework that enables low latency and memory efficient GNN serving through three key designs: (i) serverless-native on-demand graph fetching strategy that enables lightweight in-container graph sampling with full dataset resides in remote; (ii) memory-aware adaptive feature caching policy, which facilitates data reuse between requests to reduce redundant fetching; and (iii) load-aware request scheduler, which reschedules requests to bypass cold start and achieve load balance between containers. Experimental results show that FaaSGNN achieves a 5.6x lower end-to-end latency and 57.1% less memory usage on average compared to state-of-the-art works.

Serverless edge computing, characterized by fine-grained resource allocation and rapid task scheduling, is effectively implemented in edge clouds to support a diverse array of Deep Neural Network (DNN) based inference tasks. However, before executing inference tasks, the system needs to load DNN models into a container, a process known as cold start. The cold start introduces significant latency, thereby prolonging the task completion time. In particular, model loading time constitutes approximately 50%–70% of the overall task lifecycle, making it comparable to the duration of inference execution. Fortunately, we notice that some layers are required by multiple models and only need to be loaded once if these tasks are scheduled onto one server, i.e., DNN layer sharing. However, the computational resource differences among edge servers and their limited memory capacity make task scheduling and layer loading decisions particularly challenging. Therefore, to fully leverage the potential of DNN layer sharing, we investigate the Layer Sharing aware Task Scheduling (LSTS) problem with the goal of minimizing the task completion time. We formulate it into a Quadratic Integer Programming (QIP) problem and linearize it into an Integer Linear Programming (ILP) form, which is then proved as NP-hard. To tackle the computation complexity, we propose a Randomized Rounding-based Layer-Sharing-Aware Task Scheduling algorithm (LSTS-RR). Through comprehensive experimental evaluations, we confirm the effectiveness of our algorithm, as it reduces task completion time by more than 36% compared to other state-of-the-art approaches across a range of widely recognized DNN models.

Serverless computing is increasingly being adopted to provision various on-demand services at the edge cloud, including inference tasks based on deep neural networks (DNNs) for the Internet of Things (IoT). This approach leverages the advantages of flexible resource allocation and fine-grained resource management. However, the provisioning of on-demand inference typically requires downloading the DNN model at runtime, which can introduce significant delays. In the edge cloud with heterogeneous network connections, the inevitable model downloading time and the inter-model data transmission impose high challenges to the QoS of inference tasks. In this paper, we investigate how to jointly consider both model downloading time and communication time to minimize inference time. We first formulate this problem into a nonlinear optimization form and proved it as NP-hard. We further propose a Priority-based Model Assignment (PASS) algorithm in polynomial time and trace-driven experimental results show that it reduces the average inference time by 23.6% compared to existing state-of-the-art solutions.

Low Earth Orbit (LEO) satellites face critical energy constraints when performing onboard computation during Earth's umbra periods. While task migration to illuminated satellites mitigates battery stress, it introduces a trade-off between migration frequency and energy cost. This paper investigates the energy-efficient task migration problem for LEO satellite constellations subject to umbra constraints. We formally formulate this problem as a mixed-integer nonlinear programming (MINLP) model, which captures the complex trade-offs among energy consumption, computational load, and inter-satellite communication constraints under umbra conditions. To address this challenge, we propose a Score-based Priority Satellite Task Migration with Dynamic Programming (SPTD) algorithm that optimizes total energy consumption by assigning priority to candidate satellites based on their remaining energy capacity, computational capacity, and inter-satellite link bandwidth. Extensive simulations demonstrate that SPTD achieves significantly lower total energy consumption compared to state-of-the-art migration strategies, with performance approaching the theoretical optimum. The results validate the algorithm's effectiveness across varying constellation scales, providing a foundation for sustainable operation of future LEO satellite computing infrastructure.

The rapid deployment of intelligent applications on edge cloud calls for efficient and responsive DNN inference, especially under the burst scenarios of inference request. Serverless inference offers a promising solution by enabling rapid and flexible activation of inference tasks to cope with peak request, but its achievable performance is highly influenced by the initialization overhead. Existing studies on inference acceleration mainly focuses on execution optimization, they usually overlook the fact that inference performance also heavily depends on the initialization. In this paper, we propose EdgePrios, a novel priority-based scheduling mechanism that jointly optimizes initialization and execution phases for serverless inference acceleration. EdgePrios dynamically prioritizes tasks by considering workloads, dependency relationships, and the current status of available resources. It enables precise assignment of tasks to computing resources while minimizing overall inference time in edge cloud. Extensive trace-driven evaluations demonstrate the efficiency of EdgePrios as it outperforms state-of-the-art methods, achieving 15.6%-38.8% reduction in inference time under varying resource configurations, network bandwidths, and application topologies.

Thanks to the fine-grained resource management capabilities, serverless computing has been extended to edge cloud environments to support diverse Artificial Intelligence of Things (AIoT) applications, particularly those involving complex workflows of interdependent deep neural network (DNN) inference tasks. However, the inherent on-demand provisioning nature of serverless computing imposes the fact that, in serverless inference processes, the DNN models are typically maintained in the remote storage cluster and retrieved as needed. This inevitably incurs substantial latency overhead, particularly in resource-constrained edge cloud. In this paper, we investigate how to accelerate the AI application with joint consideration of both the model downloading time and intermediate data transmission time. We first formulate this problem into a nonlinear optimization form and prove it as NP-hard. We further propose a Priority-Based Model Assignment (PASS) algorithm and theoretically analyze its upper bound. The trace-driven experimental results demonstrate that our proposed algorithm outperforms other sate-of-art solutions and reduces the average application completion time by 23.6%.

Over 50 million people globally suffer from Alzheimer's disease (AD), emphasizing the need for efficient, early diagnostic tools. Traditional methods like Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scans are expensive, bulky, and slow. Microwave-based techniques offer a cost-effective, non-invasive, and portable solution, diverging from conventional neuroimaging practices. This article introduces a deep learning approach for monitoring AD , using realistic numerical brain phantoms to simulate scattered signals via the CST Studio Suite. The obtained data is preprocessed using normalization, standardization, and outlier removal to ensure data integrity. Furthermore, we propose a novel data augmentation technique to enrich the dataset across various AD stages. Our deep learning approach combines Recursive Feature Elimination (RFE) with Principal Component Analysis (PCA) and Autoencoders (AE) for optimal feature selection. Convolution Neural Network (CNN) is combined with Gated Recurrent Unit (GRU), Bidirectional Long Short Term Memory (Bidirectional-LSTM), and Long Short-Term Memory (LSTM) to improve classification performance. The integration of RFE-PCA-AE significantly elevates performance, with the CNN+GRU model achieving an 87% accuracy rate, thus outperforming existing studies.

The prosperity of machine learning applications has promoted the rapid development of GPU architecture. It continues to integrate more CUDA Cores, larger L2 cache and memory bandwidth within SM. Moreover, the GPU integrates Tensor Core dedicated to matrix multiplication. Although studies have shown that task co-location could effectively improve system throughput, existing works only focus on resource scheduling at the SM level and cannot improve resource utilization within the SM. In this paper, we propose Aker, a static kernel fusion and scheduling approach to improve resource utilization inside the SM while ensuring the QoS (Quality-of-Service) of co-located tasks. Aker consists of a static kernel fuser, a duration predictor for fused kernels, an adaptive fused kernel selector, and an enhanced QoS-aware kernel manager. The kernel fuser enables the static and flexible fusion for a kernel pair. The kernel pair could be Tensor Core kernel and CUDA Core kernel, or computing-prefer CUDA Core kernel and memory-prefer CUDA Core kernel. After the kernel fuser provides multiple fused kernel versions for a kernel pair, the duration predictor precisely predicts the duration of the fused kernels and the adaptive fused kernel selector locates the optimal fused kernel version. Finally, the kernel manager invokes the fused kernel or the original kernel based on the QoS headroom of latency-critical tasks to improve the system throughput. Our experimental results show that Aker improves the throughput of best-effort applications compared with state-of-the-art solutions by 50.1% on average, while ensuring the QoS of latency-critical tasks.

The booming development of artificial intelligence (AI) applications has greatly promoted edge intelligence technology. To support latency-sensitive Deep Neural Network (DNN) based applications, the integration of serverless inference paradigm into edge intelligence has become a widely recognized solution. However, the long DNN model downloading time from central clouds to edge servers hinders inference performance, and asks for establishing model repository within the edge cloud. This paper first identifies the inherent layer redundancy in DNN models, which is potentially beneficial to improve the storage efficiency of the model repository in the edge cloud. However, how to exploit the layer redundancy feature and allocate the DNN layers across different edge servers with capacitated storage resources to reduce the model downloading time remains challenging. To address this issue, we first formulate this problem in Quadratic Integer Programming (QIP) form, based on which a randomized rounding layer redundancy aware DNN model storage planning strategy is proposed. Our approach significantly reduces model downloading time by up to 63% compared to state-of-the-art methods, as demonstrated through extensive trace-driven experiments.

Transformers have dominated the field of natural language processing, attributed to their capability to handle sequential input data. There is a surge of work on computational and networking optimizations, aimed at improving the training efficiency of Transformers. However, transformer inference, a cornerstone of myriad AI services, remains relatively underexplored. With the challenge of variable-length inputs, conventional methods adopt padding schemes, resulting in computational waste. Moreover, works on transformer inference often overlook the integration between request scheduling and batching, which play pivotal roles in inference systems. To address these challenges, we introduce TCB, a comprehensiveTransformer inference system that integrates aConcatBatching scheme to reduce computational redundancy by concatenating requests. In addition, we present an online request batching algorithm, designed to augment the throughput of scheduled requests. Consider a muiti-server case, we further introduce a joint request assignment and batching scheduling policy to fully utilize resources on servers while ensuring quality-of-service of inference. Extensive experiments demonstrate that our proposed methods can significantly outperform existing works.

Cloud computing has emerged as a dominant force in tackling dynamic, unpredictable, and adaptable computing demands. However, the vast scale of cloud computing, the complexities of diverse scenarios, and the unpredictable nature of user requests pose significant challenges to achieving efficient and effective cloud computing scheduling. To address these issues and enhance computational efficiency while reducing energy consumption in cloud computing systems, we introduce a workload-adaptive deep reinforcement learning algorithm that dynamically adjusts the discount factor in accordance with real-time workload changes. To demonstrate the effectiveness of our proposed method, we utilize a real dataset from the Google Cloud computing center, which encompasses workloads from 12,000 machines running over 670,000 applications and 40 million jobs. Simulation results reveal that our method achieves a 9.93% reduction in task rejection rate and a 5.01% decrease in energy consumption cost compared to state-of-the-art solutions.

In serverless computing, an idle container is not recycled directly, in order to mitigate time-consuming cold container startup. These idle containers still occupy the memory, exasperating the memory shortage of today's data centers. By offloading their cold memory to remote memory pool could potentially resolve this problem. However, existing offloading policies either hurt the Quality of Service (QoS) or are too coarse-grained in serverless computing scenarios.

Distributed Deep Neural Network (DNN) inference is a promising technology to explore the distributed resources in edge cloud to realize edge intelligence. Meanwhile the inherent resource sharing nature of edge cloud infrastructure also raises serious concerns on security and privacy. Software Guard Ex-tensions (SGX) emerges as a potential hardware-level solution but its limited secure memory (i.e., enclave page cache) imposes new challenges, especially in contrast to memory-hungry DNN models. A task's performance will be severely affected when its memory footprint is beyond the enclave page cache size, due to expensive secure page swapping. In this case, how to appropriately partition a DNN model and assign the partitions to distributed edge servers to efficiently utilize edge resources for fast secure inference becomes a challenging problem. In this paper, we first show that this problem is NP-hard. We further propose a MEmory -aware Distributed Inference Acceleration (MEDIA) algorithm, whose guaranteed approximation ratio is also formally analyzed. We have implemented a prototype system and applied some well-known representative DNN models to evaluate MEDIA's performance. Through extensive experiments, we verify the efficiency of MEDIA by the fact that it reduces the inference time by 19.5%-38.1 % in comparison with state-of-the-art approaches.

Transformer model empowered architectures have become a pillar of cloud services that keeps reshaping our society. However, the dynamic query loads and heterogeneous user requirements severely challenge current transformer serving systems, which rely on pre-training multiple variants of a foundation model, i.e., with different sizes, to accommodate varying service demands. Unfortunately, such a mechanism is unsuitable for large transformer models due to the additional training costs and excessive I/O delay. In this paper, we introduce OTAS, the first elastic serving system specially tailored for transformer models by exploring lightweight token management. We develop a novel idea called token adaptation that adds prompting tokens to improve accuracy and removes redundant tokens to accelerate inference. To cope with fluctuating query loads and diverse user requests, we enhance OTAS with application-aware selective batching and online token adaptation. OTAS first batches incoming queries with similar service-level objectives to improve the ingress throughput. Then, to strike a tradeoff between the overhead of token increment and the potentials for accuracy improvement, OTAS adaptively adjusts the token execution strategy by solving an optimization problem. We implement and evaluate a prototype of OTAS with multiple datasets, which show that OTAS improves the system utility by at least 18.2%.

Edge native applications usually consist of several dependent tasks encapsulated in containers and started on-demand in the edge cloud. Unfortunately, the application performance is deeply affected by the notorious cold startup problem of containers. Pre-warming Zygote container pre-imported certain common packages has been proven as an effective startup acceleration solution. Since a Zygote can be shared among colocated tasks that require identical common packages, not only the Zygote planning but also the task scheduling decisions shall be carefully made to maximize the benefit of the Zygotes pre-warmed in limited memory. Additionally, task dependency necessitates co-locating highly dependent tasks on the same server, naturally raising a dilemma in task scheduling. To this end, in this paper, we investigate the problem of how to plan Zygote and schedule tasks for application completion time minimization, which is proved to be NP-hard. We further propose a Priority and Popularity (P&P) based edge native application acceleration algorithm. Both theoretical analysis and extensive experiments demonstrate the effectiveness of our proposed algorithm. The experiment results show that P&P can reduce the application completion time by 11.7%.

Edge-End Consortium, with the integration of edge computing servers and end devices like IoT devices, has emerged as a promising infrastructure for on-site computing power provision. But the high heterogeneity has raised unprecedented challenges in its resource management and task scheduling. Both WebAssembly and Container provide lightweight and portable way to consolidate and deploy microservices to combat the heterogeneity problem. However, the inherent advantages and disadvantages of WebAssembly and Container make it nontrivial to decide the microservice consolidation way and the deployment site, especially in the consideration of dependency between microservices. In this paper, we investigate the problem of mixed deployment of WebAssembly and Container based dependent microservices to strike a balance between communication and deployment cost toward high overall cost efficiency. We first cast this problem into a Quadratic Integer Programming (QIP) formulation and prove it as NP-hard. We then introduce a Randomized Rounding WebAssembly and Containerized Microservice Deployment (RR-WCMD) algorithm with polynomial computation complexity and guaranteed performance efficiency. Experiment results show that RR-WCMD can significantly lower the cost by an average of 34% in comparison with state-of-theart algorithms, thanks to the joint consideration of consolidation way and deployment site.

Graph Neural Networks (GNNs) have emerged as a powerful technique across various applications, due to their effective message-passing mechanism. However, their deployment is constrained by limited computational resources, energy concerns, and low-latency processing requirements. While existing works employ logit-based knowledge distillation from GNNs to guide Multilayer Perceptrons (MLPs) training, these methods may lead to reduced accuracy and compromised robustness. These drawbacks arise from two primary factors: the insufficient exploitation of the rich information embedded within the graph structures and the inherent susceptibility of MLPs to noisy data. To tackle these issues, we propose a Mixed Multi-order Knowledge Distillation (MMKD) method, which combines the GNN's logits with hidden layer information through the multi-order distillation to improve the accuracy of the MLP. Moreover, we employ both raw data and perturbed data as input, enhancing the density of knowledge extraction as well as the MLPs' generalization. Extensive experiments across seven benchmark datasets verify the superior performance of our approach in terms of effectiveness and robustness. In comparison with the baseline, our approach achieves an accuracy improvement of up to 8.68% in typical GNN tasks.

Thanks to the capability of fine-grained resource allocation and fast task scheduling, serverless computing has been adopted into edge cloud to accommodate various applications, e.g., deep neural network (DNN) inference for Artificial Intelligence of Things (AIoT). In serverless edge cloud, the servers are started up on-demand. However, as a container-based architecture, the inherent sequential startup feature of container imposes high affection on the DNN inference performance in serverless edge clouds. In this article, we investigate the distributed DNN inference problem in serverless edge cloud with the consideration of such characteristics, aiming to eliminate the extra container startup time cost to minimize the DNN inference latency. We formulate this problem into a nonlinear optimization form and then linearize it into an integer programming problem, which is proved as NP-hard. To tackle the computation complexity, we propose a priority-based layer scheduling (PLAYS) algorithm. Extensive experiment results verify the effectiveness and the adaptability of our PLAYS algorithm in comparison with other state-of-art algorithms under several well known DNN models.

User-facing applications often experience excessive loads and are shifting towards the microservice architecture. To fully utilize heterogeneous resources, current datacenters have adopted the disaggregated storage and compute architecture, where the storage and compute clusters are suitable to deploy the stateful and stateless microservices, respectively. Moreover, when the local datacenter has insufficient resources to host excessive loads, a reasonable solution is moving some microservices to remote datacenters. However, it is nontrivial to decide the appropriate microservice deployment inside the local datacenter and identify the appropriate migration decision to remote datacenters, as microservices show different characteristics, and the local datacenter shows different resource contention situations. We therefore propose ELIS, an intra- and inter-datacenter scheduling system that ensures the Quality-of-Service (QoS) of the microservice application, while minimizing the network bandwidth usage and computational resource usage. ELIS comprises aresource manager, across-cluster microservice deployer, and areward-based microservice migrator. The resource manager allocates near-optimal resources for microservices while ensuring QoS. The microservice deployer deploys the microservices between the storage and compute clusters in the local datacenter, to minimize the network bandwidth usage while satisfying the microservice resource demand. The microservice migrator migrates some microservices to remote datacenters when local resources cannot afford the excessive loads. Experimental results show that ELIS ensures the QoS of user-facing applications. Meanwhile, it reduces the public network bandwidth usage, the remote computational resource usage, and the local network bandwidth usage by 49.6%, 48.5%, and 60.7% on average, respectively.

To satisfy differentiated service requirements of delay-sensitive and computing-intensive tasks in satellite communication networks, we propose a cache-assisted low-latency content distribution scheme in multi-cell multi-carrier nonorthogonal multiple access (MCMC-NOMA)-based satellite-edge-terminal cooperation environments. In the paper, a delay minimization problem is formulated to achieve optimal content distribution by jointly optimizing the allocation of subchannels and transmit power, which is decomposed and solved in each slot by using many-to-one matching and successive convex approximation, respectively. Simulation results show that the proposed model significantly improve network latency and content distribution in comparison to its existing counterparts in cloud-edge-terminal cooperation networks.

Using docker to encapsulate the task has been regarded as a potential way to achieve efficient task orchestration and management in edge computing. Despite the lightweight nature of containers, downloading a larger container image can still be resource-intensive, particularly in resource-constrained edge environments. Fortunately, the unique layered architecture of the container allows multiple containerized tasks to share the same layer, thereby offering an opportunity for reducing the image downloading overhead via sharing the common layers. To explore the potential of layer sharing on image downloading over-head reduction, we investigate a joint task scheduling and image routing problem in edge environment, aiming at minimizing the image downloading overhead. We first formulate the problem into an integer linear programming form, and then propose a layer-aware scheduling and routing (LSR) algorithm to tackle this problem. Finally, to evaluate the effectiveness of our proposed algorithm, we conduct a group of simulation experiments. The experimental results show that our proposed algorithm can reduce the download time by about 20% in comparison with other approaches.

Edge computing is considered a promising computing paradigm that can mitigate energy consumption and workload of end devices through task offloading to edge servers. Albeit with high potential, edge computing is still challenged by various forms of “system noise”, e.g., node failures, system failures, and poor network conditions. To this end, distributed coded computing has been proposed for alleviating such effects by introducing redundancy into the computation. However, traditional distributed coded computing only focuses on leveraging the unreliable computing resource, and this indeed increases the risk of task non-completion within the acceptable timeframe. To address this problem, in this paper, we propose a partial distributed coded computing framework that can leverage the reliable and unreliable resources in the edge environment. We further investigate the problem of how to offload the coded subtasks for energy reduction while meeting the task tolerant latency. To tackle the computation complexity, we then propose an Iterated Greedy Algorithm. The experimental results verify the efficiency of our proposed algorithm, and it can reduce the energy consumption by 20% compared with other algorithms.

Many emerging user-facing services adopt Graph Neural Networks (GNNs) to improve serving accuracy. When the graph used by a GNN model changes, representations (embedding) of nodes in the graph should be updated accordingly. However, the node representation update is too slow, resulting in either long response latency of user queries (inference is performed after update completes) or high staleness problem (inference is performed based on stale data).Our in-depth analysis shows that the slow update is mainly due to neighbor explosion problem in graphs and duplicated computation. Based on such findings, we propose STAG, a GNN serving framework that enables low latency and low staleness of GNN-based services. It comprises a collaborative serving mechanism and an additivity-based incremental propagation strategy. With collaborative serving mechanism, only part of node representations are updated during the update phase, and the final representations are calculated in the inference phase. It alleviates the neighbor explosion problem. The additivity-based incremental propagation strategy reuses intermediate data during update phase, eliminating duplicated computation. Experimental results show that STAG greatly reduces staleness time with a slight increase in response latency, and support 2.7~27x workload compared to existing approaches.

Latency-critical applications directly interact with end users and often experience the diurnal load pattern. In production, best-effort applications are often co-located with them to utilize the idle cores at the low load. Meanwhile, modern computers are evolving towards heterogeneous NUMA architecture, where the cores have different computation abilities, memory access latencies and network communication delays. Prior co-location scheduling work did not consider the NUMA architecture, and failed to maximize the throughput of best-effort applications while ensuring the required QoS of latency-critical applications. Our investigation shows that NUMA effect has complex impacts on the latency of latency-critical applications and the throughput of best-effort applications. We therefore propose PAC, a preference-aware co-location scheduling scheme that considers the NUMA effect for heterogeneous NUMA architectures. PAC has a performance monitor and a core scheduler. Specifically, the performance monitor identifies the "dangerous" latency-critical applications that require upgrading core allocations. We propose two low-overhead scheduling strategies for the scheduler. The strategies identify the bottlenecks of applications and adjust core allocations accordingly. Experimental result shows that PAC improves the throughput of best-effort applications by 3.87× while ensuring the required QoS of latency-critical applications.

Container based microservices have been widely applied to promote the cloud elasticity. The mainstream Docker containers are structured in layers, which are organized in stack with bottom-up dependency. To start a microservice, the required layers are pulled from a remote registry and stored on its host server, following the layer dependency order. This incurs long microservice startup time and hinders the performance efficiency. In this paper, we discover that, for the first time, the layer pulling order can be adjusted to accelerate the microservice startup. Specifically, we address the problem on microservice layer pulling orchestration for startup time minimization and prove it as NP-hard. We propose a Longest-chain based Out-of-order layer Pulling Orchestration (LOPO) strategy with low computational complexity and guaranteed approximation ratio. Through extensive real-world trace driven experiments, we verify the efficiency of our LOPO and demonstrate that it reduces the microservice startup time by 22.71% on average in comparison with state-of-the-art solutions.

The cold startup of the container is regarded as a crucial problem to the performance of serverless computing, especially to the resource-capacitated edge clouds. Pre-warming hot containers has been proved as an efficient solution but is at the expense of high memory consumption. Instead of pre-warming a complete container for a function, recent studies advocate Zygote container, which pre-imports some packages and is able to import the other dependent packages at runtime, so as to avoid the cold startup problem. However, as different functions have different package dependencies, how to plan the Zygote generation and pre-warming in a resource-capacitated edge cloud becomes a critical challenge. In this paper, aiming to minimize the overall function startup time and subjective to the resource capacity constraints, we formulate this problem into a Quadratic Integer Programming (QIP) form. We further propose a Randomized Rounding based Zygote Planning (RRZP) algorithm. The performance efficiency of our algorithm is proved via both theoretical analysis and trace-driven simulations. The results show that our algorithm can significantly reduce the startup time by 25.6%.

Although the containers are featured by light-weightness, it is still resource-consuming to pull and startup a large container image, especially in relatively resource-constrained edge cloud. Fortunately, Docker, as the most widely used container, provides a unique layered architecture that allows the same layer to be shared between microservices so as to lower the deployment cost. Meanwhile, it is highly desirable to deploy dependent microservices of an application together to lower the operation cost. Therefore, the balancing of microservice deployment cost and the operation cost should be considered comprehensively to achieve minimal overall cost of an on-demand application. In this paper, we first formulate this problem into a Quadratic Integer Programming form (QIP) and prove it as a NP-hard problem. We further propose a Randomized Rounding-based Microservice Deployment and Layer Pulling (RR-MDLP) algorithm with low computation complexity and guaranteed approximation ratio. Through extensive experiments, we verify the high efficiency of our algorithm by the fact that it significantly outperforms existing state-of-the-art microservice deployment strategies.

To enable inter-container communication of services on different hosts, container overlay network, the most widely used container network mode, provides a layer of virtual network between containers to transparent the physical device heterogeneous. However, overlay network produces two-layered packet encapsulation, and current container network management system cannot identify the source containers from the two-layered encapsulated packets or control the network traffics of different services. To tackle this issue, a traffic control system for container overlay networks (CONTC) is proposed and implemented by redesigning the packets processing procedure in overlay network model, enabling accurate network packet identification, multi-level network resource management and user-centric customized control. Extensive experiment results are conducted on different system settings and practical service cases to show that CONTC can provide accurate network traffic control for data flows with different network protocols and different packet sizes at both the service level and the container level. The results based on open-source microservice benchmarks also validate the correctness and effectiveness of CONTC by reducing the tail latency of single-service and multi-service environments by 37.53% and 22.33%, respectively.

Network Function Virtualization (NFV) has already become an essential technology for improving the scalability and flexibility of modern computer networks. The performance gap has become the main issue that impedes the development of NFV. GPUs, with massive parallel processors, are advocated to accelerate the Virtualized Network Functions (VNFs). However, the special architecture and workflow of GPUs introduce new challenges, especially on the batched processing, and resource allocation. In this paper, we propose GPU-based NFV Acceleration framework (GNFA) with an efficient packet batching and resource allocation solution. Considering the increased latency caused by the accumulation of the GPU kernel invoking overhead, we first invent a latency reduction mechanism called SM Performance Compensation (SPC). A Partition and Adjustment based Batching and Resource Allocation (PABARA) algorithm that jointly considers batch size tuning and GPU thread allocation is also proposed. We have practically implemented GNFA and extensively evaluated its performance on some well-known VNFs. The experiment results show that GNFA can effectively promote the GPU resource utilization and improve the NFV performance in terms of per-packet latency.

By leveraging standard IT virtualization technology and Commercial-Off-The-Shelf (COTS) servers, Network Function Virtualization (NFV) decouples network functions from proprietary hardware devices for flexible service provisioning. But the potential of NFV is significantly limited by its performance inefficiency. With the unparalleled advantages of multi-core parallelism and high memory bandwidth, Graphics Processing Units (GPUs) are regarded as a promising way to accelerate Virtualized Network Functions (VNF). However, the special architecture of GPU brings new challenges to task scheduling and resource allocation. To this end, we propose aGPUorientedspatio-temporal sharing framework for NFV calledGost, aiming for GPU based VNF performance promotion. The execution order and GPU resource allocation (i.e., the number of threads) are considered in task scheduling to minimize the end-to-end latency for VNF flows. First, we formulate the task scheduling problem into a nonlinear programming form, and then transform it into an equivalent Integer Linear Programming (ILP) form. The problem is proved as NP-hard. We customize the classical list scheduling algorithm and propose a List Scheduling based Spatio-Temporal GPU sharing strategy (LSSTG), whose achievable worst-case performance is also formally analyzed. We practically implementGostprototype, based on which extensive experiments verify the high performance efficiency of LSSTG compared to state-of-the-art in terms of latency and throughput.

Federated Distillation (FD) extends classic Federated Learning (FL) to a more general training framework that enables model-heterogeneous collaborative learning by Knowledge Distillation (KD) across multiple clients and the server. However, existing KD-based algorithms usually require a set of shared input samples for each client to produce soft-prediction for distillation. Worse still, such a manual selection is accompanied by careful deliberations or prior information on clients’ private data distribution, which is not in line with the privacy-preserving characteristic of classic FL. In this paper, we propose a novel training framework to achieve data-independent knowledge transfer by properly designing a distributed generative adversarial network (GAN) between the server and clients that can synthesize shared feature representations to facilitate the FD training. Specifically, we deploy a generator on the server and reuse each local model as a federated discriminator to form a lightweight efficient distributed GAN that can automatically synthesize simulated global feature representations for distillation. Moreover, since the synthesized feature representations are usually more faithful and homologous with global data distribution, faster and better training convergence can be obtained. Extensive experiments on different tasks and heterogeneous models demonstrate the effectiveness of the proposed framework on model accuracy and communication overhead.

While deep neural network (DNN) models are mainly trained using GPUs, many companies and research institutions build shared GPU clusters. These clusters host DNN training jobs, DNN inference jobs, and CPU jobs (jobs in traditional areas). DNN training jobs require GPU for main computation and CPU for auxiliary computation. Some DNN inference jobs could rely solely on CPU, while others must utilize both CPU and GPU. Our investigation demonstrates that the number of cores allocated to a training job significantly impacts its performance, and that DNN inference jobs can make use of the limited CPU cores on the GPU nodes. To accomplish this, we characterize representative deep learning models in terms of their CPU core requirements for their training jobs and inference jobs, and investigate their sensitivity to other CPU-side resource contention. Based on the characterization, we propose SODA, a scheduling system comprised of an adaptive CPU allocator, a multi-array job scheduler, a hardware-aware inference job placer, and a real-time contention eliminator. The experimental results indicate that SODA increases GPU utilization by an average of 19.9%, while maintaining the quality of service target for all DNN inference jobs and the queuing performance of CPU jobs.

As a rising star of social apps, short video apps, e.g., TikTok, have attracted a large number of mobile users by providing fresh and short video contents that highly match their watching preferences. Meanwhile, the booming growth of short video apps imposes new technical challenges on the existing computation and communication infrastructure. Traditional solutions maintain all videos on the cloud and stream them to users via contend delivery networks or the Internet. However, they incur huge network traffic and long delay that seriously affects users’ watching experiences. In this article, we propose an edge-assisted short video sharing framework to address these challenges by caching some highly preferred videos at edge servers that can be accessed by users via high-speed network connections. Since edge servers have limited computation and storage resources, we design an online algorithm with provable approximation ratio to decide which videos should be cached at edge servers, without the knowledge of future network quality and watching preferences changes. Furthermore, we improve the performance by jointly considering video fetching and user-edge association. Extensive simulations are conducted to evaluate the proposed algorithms under various system settings, and the results show that our proposals outperform existing schemes.

Edge computing has become an alternative low-latency provision of cloud computing thanks to its close-proximity to the users, and the geo-distribution nature of edge servers enables the utilization green energy from the environment on-site. To pursue the goal of low-carbon edge computing, it is desirable to minimize the operational expenditure by scheduling the computing resource and green energy according to the spatially and temporally varying user demands. In this article, inspired by the successful application of deep reinforcement learning (DRL) in diverse domains, we propose a DRL-based edge computing management strategy which continuously explores the states and adaptively makes decisions on service management and energy scheduling, towards long-term cost minimization. Different from model-based solutions, our proposal is a model-free method, without any assumption on statistical knowledge as a priori, and therefore is practical in implementation. To speedup the agent training procedure, we further design a prioritized replay memory by utilizing the model-based solution as a guideline to set the transition priority. Extensive experiment results based on real-world traces validate that our proposed DRL-based strategy can make considerably progress compared to the one-shot greedy strategy, and it can learn the system dynamically to manage the edge computing services at runtime.

Task offloading to edge servers has become a promising solution to tackle the computation resource poverty of the end devices. However, the zero-trust edge computing platform is highly challenged by the growing concern on security and privacy. Thus, Trust Execution Environment (TEE), like TrustZone, is advocated to empower edge clouds to enable secure task offloading. To explore TrustZone, the inevitable involvement of data encryption and decryption operations makes existing offloading strategies not applicable any more, especially when the task dependency is considered. In addition, TrustZone has distinguishable task scheduling paradigm as one CPU core does not allow multitask coexist at the same time. Taking the above issues into consideration, we investigate a dependency-aware task offloading problem for makespan minimization in TrustZone empowered edge clouds. By inventing an extended graph to describe the task execution process, we provide a formal statement to the problem and prove its NP-hardness. We then propose a Customized List Scheduling (CLS) based approximate algorithm and theoretically analyze its achievable performance. Extensive testbed based experiment results show that our approximation algorithm can effectively reduce the makespan and significantly outperforms existing state-of-the-art offloading approaches in TrustZone empowered edge clouds.

GPU has been widely applied to accelerate the DNN based applications. However, single GPU is overwhelmed by the increasing computation requirement of large-scale DNN inference task. Although GPUs cluster alleviates the pressure of massive inference task, it still traps into low efficiency owing to the limited computing power and network bandwidth. Besides, the exclusiveness of GPU device may cause the straggler problem and hence long inference time. Reinforcement Learning (RL) has been widely used in such task scheduling problems. But the delayed reward during the training may slow the convergence speed or even result in non-convergence. To this end, in this paper, we design an improved reinforcement learning based algorithm, called DRM-DQL, to achieve a layer-aware DNN inference task scheduling. We first analyze and model the layer-wise inference task scheduling problem by deep Q-learning. Then, a delayed reward matching strategy is proposed for matching the global reward value to the immediate reward value, which help the algorithm to get the right experience in DNN layer scheduling. The experiment results demonstrate that our algorithm performs better than both heuristic algorithm and the vanilla DQL algorithm, and show the robustness in various network bandwidths, computing power, and DNN model structures.

Unmanned Aerial Vehicle (UAV) has been widely applied in many domains. But the computation and energy resource limitation severely hinders its development and application. Mobile Edge Computing (MEC) emerges as a promising platform to process the tasks offloaded from the UAVs to effectively improve the Quality-of-Service (QoS). To this vision, it is first required that the edge servers must be deployed with the needed service to handle the offloaded task. Fortunately, by exploring cloud native computing technology, it is possible to deploy container-based microservice to MEC in a prompt way. In this case, it raises the task scheduling problem on whether to deploy a new service or to utilize an existing service to balance the overhead between data transmission and the microservice deployment (i.e., container image pulling) for overall task completion time minimization. In this paper, the problem is first formulated in Integer Linear Programming (ILP) form, and proved to be NP-hard. We further propose an incentive-based request scheduling algorithm. Experiments based on track-driven simulations show that the total completion time for all tasks is reduced by 21.37% compared to the state-of-the-art solution.

Cloud native computing featured by microservice has been regarded as a compelling trend in cloud application development. Edge computing, as an alternative or complemen-tary to cloud computing, is potential to expand the microservice to edge computing, simplifying the development and deployment of edge applications. Despite that, there is still a challenge on how to manage the microservices efficiently in the open and heterogeneous distributed environment. To this end, service mesh provides a potential solution in efficient microservices management. However, as a traditional cloud-oriented architecture, it can not be applied into edge computing directly since the centralized controller policy. To address this problem, in this paper, we propose an edge service mesh architecture with distributively deployed controllers for edge native computing. We further inves-tigate the problem on how to deploy these distributive controllers in a cost efficient manner with the consideration of control cost and the synchronization cost. The problem is formulated into a non-linear optimization form and then linearized into an integer linear programming (ILP) problem. To tackle the computation complexity, we then come up with a customized k-means based algorithm (i.e., ck-means) in polynomial computation complexity. The experimental results verify the efficiency of our ck-means algorithm in comparison with the traditional k-means algorithm.

Transformer has dominated the field of natural language processing because of its strong capability in learning from sequential input data. In recent years, various computing and networking optimizations have been proposed for improving transformer training efficiency. However, transformer inference, as the core of many AI services, has been seldom studied. A key challenge of transformer inference is variable-length input. In order to align these input, existing work has proposed batching schemes by padding zeros, which unfortunately introduces significant computational redundancy. Moreover, existing transformer inference studies are separated from the whole serving system, where both request batching and request scheduling are critical and they have complex interaction. To fill the research gap, we propose TCB, a Transformer inference system with a novel ConcatBatching scheme as well as a jointly designed online scheduling algorithm. ConcatBatching minimizes computational redundancy by concatenating multiple requests, so that batch rows can be aligned with reduced padded zeros. Moreover, we conduct a systemic study by designing an online request scheduling algorithm aware of ConcatBatching. This scheduling algorithm needs no future request information and has provable theoretical guarantee. Experimental results show that TCB can significantly outperform state-of-the-art.

Lightweight container-based microservice has been widely advocated to promote the elasticity of edge cloud. The inherent layered structure of containers offers a compelling way to cope with the resource scarcity of edge servers through layer sharing, which can significantly increase storage utilization and improve the edge throughput. Recent studies show that it is possible to share layers not only within the same server but also between servers, which microservice deployment can take full advantage of. In this paper, we investigate the problem of how to collaboratively deploy microservices by incorporating both intra-server and inter-server layer sharing to maximize the edge throughput. We formulate this problem into an integer linear programming form and prove it as NP-hard. We propose a randomized rounding based heuristic algorithm, and conduct formal analysis on the guaranteed approximation ratio. Through extensive experiments, we verify the efficiency of our proposed algorithm, and the results demonstrate that it can deploy 6× and 12× more microservice instances, and improve the edge throughput by 27.74% and 38.46% in comparison with state-of-the-art strategies.

User-facing applications often experience excessive loads and are shifting towards microservice software architecture. While the local datacenter may not have enough resources to host the excessive loads, a reasonable solution is moving some microservices of the applications to remote datacenters. However, it is nontrivial to identify the appropriate migration decision, as the microservices show different characteristics, and the local datacenter also shows different resource contention situations. We therefore propose ELIS, an inter-datacenter scheduling system that ensures the required Quality-of-Service (QoS) of the microservice application with excessive loads, while minimizing the resource usage of the remote datacenter. ELIS comprises a resource manager and a reward-based microservice migrator. The resource manager finds the near-optimal resource configurations for different microservices to minimize resource usage while ensuring QoS. The microservice migrator migrates some microservices to remote datacenters when local resources cannot afford the excessive loads. Our experimental results show that ELIS ensures the required QoS of user-facing applications at excessive loads. Meanwhile, it reduces overall/remote resource usage by 13.1% and 58.1% on average, respectively.

Serverless edge computing is booming as an efficient carrier of deploying complex applications composed of dependent functions, whose assignment decisions highly influence the application performance. Although similar problem has been widely studied, none of existing approaches considers the diversity of communication styles, which is specially introduced in serverless computing and also imposes high influence to the performance efficiency. We compare two communication styles, called direct-passing and remote-storage, to transmit intermediate data between functions. We find that there is no single communication style that can prevail under all scenarios and the optimal selection depends on several factors, such as fanout degree, data size, and network bandwidth. Hence, how to select the appropriate communication style for each inter-function communication link, together with the function assignment decision, is essential to the application performance. To this end, we propose a Priority-based ASsignment and Selection (PASS) algorithm with joint consideration of function assignment and communication style selection. We theoretically analyze the approximation ratio of PASS algorithm and extensive experiments on real-world applications show that PASS can averagely reduce the completion time by 24.1% in comparison with state-of-the-art approaches.

With collaborative DNN inference, part of queries run on their source edge device to reduce latencies. Because edges show diverse performance and network conditions, different layers should run on different devices, and queries on the datacenter show irregular structures. However, emerging schemes are not able to process such irregular queries. We propose ICE, a collaborative inference service scheme that effectively supports irregular queries. ICE comprises a query slicer, a query manager, and a lag enhancer. The query slicer maps the execution of queries based on the edges' performance and network conditions. The query manager batches irregular queries adaptively and schedules the irregular queries based on their progress. The lag enhancer reduces the QoS violation when queries run slower due to interference on the edge. Experiments show that ICE improves the supported peak load of the datacenter by 43.2% on average while guaranteeing the required 99%-ile latencies compared with state-of-the-art techniques.

Underwater acoustic sensor networks (UASNs) currently provide an important technical means of underwater communication, but there are difficulties in power updating or power replenishment because the sensor nodes work in an underwater environment. Therefore, energy consumption optimization has become the focus of research on UASNs. Node clustering is widely considered to be able to optimize network energy consumption. Although the current clustering-based routing method prolongs the network life cycle to a certain extent, some nodes may fail due to excessive energy consumption caused by excessive data transmission tasks, and the problems of high and uneven energy consumption still exist. To extend the life cycle of UASNs, this article proposes an unequal clustering method based on particle swarm optimization. Our method uses iterative updates to dynamically adjust the cluster size based on the remaining energy of the cluster head, the distance from the cluster head to the Sink node, and the number of times the cluster head forwards data between clusters to balance the cluster head load. The energy consumption of the whole path from the cluster head to the Sink node and the number of hops required are considered in the intercluster transmission phase. Simulation results show that this method can effectively reduce the network energy consumption.

Offloading computation tasks through cloud–edge collaboration has been a promising way to improve the Quality of Service (QoS) of applications. Usually, cloud server (CS) and edge server (ES) are selfish and rational and, therefore, it is imperative to develop incentive mechanisms, which can encourage idle ESs or the CS to participate in the task offloading process. In this article, we propose a computation offloading method based on the game theory, which is suitable for cloud–edge computing networks. It is considered that the CS has a lot of computation tasks to conduct, and ESs usually have idle computational resources. The CS can offload computation tasks to ESs with idle computational resources to reduce its own cost and pressure, and ESs can profit by selling their computational resources. The interaction between the CS and ESs is modeled as a Stackelberg game, and the proposed game is analyzed by using the backward induction method. It is proved that the game can achieve a unique Nash equilibrium. Then, a gradient-based iterative search algorithm (GISA) is proposed to obtain the optimal solution in order to maximize the utility of the CS and ESs. Finally, numerical simulation results show that our proposed method greatly outperforms other benchmark schemes under different scenarios, and can encourage ESs to trade their computational resources with the CS effectively.

Federated learning (FL) has been widely recognized as a promising approach by enabling individual end-devices to cooperatively train a global model without exposing their own data. One of the key challenges in FL is the non-independent and identically distributed (Non-IID) data across the clients, which decreases the efficiency of stochastic gradient descent (SGD) based training process. Moreover, clients with different data distributions may cause bias to the global model update, resulting in a degraded model accuracy. To tackle the Non-IID problem in FL, we aim to optimize the local training process and global aggregation simultaneously. For local training, we analyze the effect of hyperparameters (e.g., the batch size, the number of local updates) on the training performance of FL. Guided by the toy example and theoretical analysis, we are motivated to mitigate the negative impacts incurred by Non-IID data via selecting a subset of participants and adaptively adjust their batch size. A deep reinforcement learning based approach has been proposed to adaptively control the training of local models and the phase of global aggregation. Extensive experiments on different datasets show that our method can improve the model accuracy by up to 30 percent, as compared to the state-of-the-art approaches.

Internet of Things (IoT) anomaly detection is significant due to its fundamental roles of securing modern critical infrastructures, such as falsified data injection detection and transmission line faults diagnostic in smart grids. Researchers have proposed various detection methods fostered by machine learning (ML) techniques. Federated learning (FL), as a promising distributed ML paradigm, has been employed recently to improve detection performance due to its advantages of privacy-preserving and lower latency. However, existing FL-based methods still suffer from efficiency, robustness, and security challenges. To address these problems, in this article, we initially introduce a blockchain-empowered decentralized and asynchronous FL framework for anomaly detection in IoT systems, which ensures data integrity and prevents single-point failure while improving the efficiency. Further, we design an improved differentially private FL based on generative adversarial nets, aiming to optimize data utility throughout the training process. To the best of our knowledge, it is the first system to employ a decentralized FL approach with privacy-preserving for IoT anomaly detection. Simulation results on the real-world dataset demonstrate the superior performance from aspects of robustness, accuracy, and fast convergence while maintaining high level of privacy and security protection.

User-facing services are now evolving towards the microservice architecture where a service is built by connecting multiple microservice stages. Since the entire service is heavy, the microservice architecture shows the opportunity to only offload some microservice stages to the edge devices that are close to the end users. However, emerging techniques often result in the violation of Quality-of-Service (QoS) of microservice-based services in cloud-edge continuum, as they do not consider the communication overhead or the resource contention between microservices and external co-located tasks. We propose Nautilus, a runtime system that effectively deploys microservice-based user-facing services in cloud-edge continuum. Nautilus ensures the QoS of microservice-based user-facing services while minimizing the required computational resources, which is comprised of a communication-aware microservice mapper, a contention-aware resource manager and an IO-sensitive and load-aware microservice migration scheduler. The mapper divides the microservice graph into multiple partitions based on the communication overhead and maps the partitions to appropriate nodes. On each node, the resource manager determines the optimal resource allocation for its microservices based on reinforcement learning that may capture the complex contention behaviors. Once the microservices are suffered from external IO pressure, the IO-sensitive microservice scheduler migrates the critical one to idle nodes. Furthermore, when the load of microservices changes dynamically, the load-aware microservice scheduler migrates microservices from busy nodes to idle ones to ensure the QoS goal of the entire service. Our experimental results show that Nautilus can guarantee the required QoS target under external shared resources contention while the state-of-the-art suffers from QoS violations. Meanwhile, Nautilus reduces the computational resource usage by 23.9% and the network bandwidth usage by 53.4%, while achieving the required 99%-ile latency.

Edge intelligence has emerged as a prevalent enabling technology to support various intelligent applications. Along with the prosperity, it also raises great concern on the security and privacy since the edge servers are usually shared and untrusted. The security-sensitive code (i.e., the pre-trained model) and data may be easily stolen by malicious tenants, and even untrusted infrastructure providers. To this end, Software Guard Extensions (SGX) is proposed to provide an isolated Trust Execution Environment (TEE) for security and privacy guarantee. However, we find that running tasks in SGX suffer certain performance degradation due to the limited Enclave Page Cache (EPC) size. This further leads to frequent page swapping operations and the high enclave call overhead, which are also influenced by the task (i.e., DNN layer) dispatching and scheduling. To this end, in this paper, we designLasagna, as an SGX based secure DNN inference acceleration framework, which explores the layered-structure of DNN models to well balance the usage of the scarce EPC resources and the computation resources. Lasagna mainly consists of a global task balancer and a local task scheduler, responding for task dispatching across distributed edge servers and task scheduling in local server, respectively. We evaluate Lasagna over different well-known DNN models, and the results show that Lasagna effectively speeds up the inference performance by$1.11\times -1.51\times$.

Most of the existing FL systems focus on a data-parallel architecture where training data are partitioned by samples among several parties. In some real-life applications, however, partitioning by features is also of practical relevance and the number of features is usually unbalanced among parties. The corresponding learning framework is referred to as Vertical Federated Learning (VFL). Though some pioneering work focused on VFL, the convergence properties of VFL on unbalanced features, especially when parties conduct different numbers of local updates concerning heterogeneous computational capabilities are still unknown. In this article, we propose a new learning framework to improve the training efficiency of VFL on unbalanced features. Given the number of features and the computational capability owned by each party, our thorough theoretical analysis exhibits that the number of local updates conducted by each party has a great effect on the convergence rate and the computational complexity, both of which jointly determine the overall training efficiency in an interrelated and sophisticated way. Based on our theoretical findings, we formulate an optimization problem and derive the optimal solution by selecting an adaptive number of local training rounds for each party. Extensive experiments on various datasets and models demonstrate that our approach significantly improves the training efficiency of VFL.

Edge intelligence has already been widely regarded as a key enabling technology in a variety of domains. Along with the prosperity, increasing concern is raised on the security and privacy of intelligent applications. As these applications are usually deployed on shared and untrusted edge servers, malicious co-located attackers, or even untrustworthy infrastructure providers, may acquire highly security-sensitive data and code (i.e., the pre-trained model). Software Guard Extensions (SGX) provides an isolated Trust Execution Environment (TEE) for task security guarantee. However, we notice that DNN inference performance in SGX is severely affected by the limited enclave memory space due to the resultant frequent page swapping operations and the high enclave call overhead. To tackle this problem, we propose Lasagna, an SGX oriented DNN inference performance acceleration framework without compromising the task security. Lasagna consists of a local task scheduler and a global task balancer to optimize the system performance by exploring the layered-structure of DNN models. Our experiment results show that our layer-aware Lasagna effectively speeds up the well-known DNN inference in SGX by 1.31x-1.97x.

The basic I/O operations of a system can be categorized as two distinct modes: synchronous (sync) I/O and asynchronous (async) I/O, whose performance varies on the system statues, workloads and storage devices. Appropriately applying I/O modes is critical to the system performance. However, the I/O access of diverse applications in a server, especially in a cloud, is volatile and irregular. As a result, this can lack a flexible and adaptive I/O modes, leading to the sub-optimal I/O performance. To tackle this problem, in this paper, we propose IObrain, an intelligent I/O mode recommendation system, which can adopt the appropriate I/O mode in a dynamic and self-adaptive manner according to both application needs and system statuses. IObrain first trains a lightweight recommendation model with decision tree. Then, a query hook is interposed into the storage engine to intercept the read/write operations from upper application. In this way, IObrain queries the recommendation model first before executing a read/write operation to find the right I/O mode. In addition, two techniques, called inference cache and gRPC bridge, are proposed to reduce the inherent query latency. We practically implement IObrain and verify the advantage of IObrain based on the prototype system. The experimental results show that, compared to existing approach, IObrain improves the I/O performance by up to 1.33× with mild running costs.

Container-based microservice has emerged as a promising technique in promoting edge computing elasticity. At the runtime, microservices, encapsulated in form of container images, need to be frequently downloaded from remote registries to local edge servers, which may incur significant overhead in terms of excessive download traffic and large local storage. Given the limited resources at the edge, it is of critical importance to minimize such overhead in order to enhance microservice offerings. A distinctive feature in container-based microservice, which has not been exploited, is that microservice images are in layered structure and common layers can be shared by co-located microservices. In this paper, we study a layer aware micro-service placement and request scheduling at the edge. Intuitively, throughput and number of hosted microservices can be significantly increased by layer sharing between co-located images. We formulate this into an optimization problem with approximate submodularity, and prove this to be NP-hard. We design an iterative greedy algorithm with guaranteed approximation ratio. Extensive experiments validate the efficiency of our method, and the results demonstrate that the number of placed microservices can be increased by 27.61% and the microservice throughput can be improved by 73.13%, respectively, in comparison with the state-of-the-art microservice placement strategy.

Container, as a light-weight virtualization technology with the advantages of continuous integration and easy deployment, has been widely adopted to support diverse microservices. At runtime, non-local container images need to be frequently pulled from remote registries to local servers, resulting in large pulling traffic and hence long startup time. A distinctive feature in container-based microservice, which has not been exploited, is that container images are in layered structure and some common base layers can be shared between co-located microservices. In this paper, we propose a layer sharing microservice deployment and image pulling strategy which explores the advantage of layer sharing to speedup microservice startup and lower image storage consumption. The problem is formulated into an Integer Linear Programming (ILP) form. An Accelerated Distributed Augmented Lagrangian (ADAL) based distributed algorithm executed cooperatively by registries and servers is proposed. Through extensive trace driven experiments, we validate the high efficiency of our ADAL based algorithm as it accelerates the microservice startup by 2.30 times in average and reduces the storage consumption by 55.33%.

User-facing services are now evolving towards the microservice architecture where a service is built by connecting multiple microservice stages. While an entire service is heavy, the microservice architecture shows the opportunity to only offload some microservice stages to the edge devices that are close to the end users. However, emerging techniques often result in the violation of Quality-of-Service (QoS) of microservice-based services in cloud-edge continuum, as they do not consider the communication overhead or the resource contention between microservices.We propose Nautilus, a runtime system that effectively deploys microservice-based user-facing services in cloud-edge continuum. It ensures the QoS of microservice-based user-facing services while minimizing the required computational resources. Nautilus is comprised of a communication-aware microservice mapper, a contention-aware resource manager and a load-aware microservice scheduler. The mapper divides the microservice graph into multiple partitions based on the communication overhead and maps the partitions to the nodes. On each node, the resource manager determines the optimal resource allocation for its microservices based on reinforcement learning that may capture the complex contention behaviors. The microservice scheduler monitors the QoS of the entire service, and migrates microservices from busy nodes to idle ones at runtime. Our experimental results show that Nautilus reduces the computational resource usage by 23.9% and the network bandwidth usage by 53.4%, while achieving the required 99%-ile latency.

Network Function Virtualization (NFV) enables network functions to run on general-purpose servers, thus alleviates the reliance on dedicated hardware and significantly improves the scalability and flexibility in networking service provisioning. Meanwhile, it is recognized that Virtualized Network Functions (VNFs) suffer from serious performance problem. Graphics Processing Unit (GPU), with massive processing cores, has been advocated as a potential accelerator for improving the performance efficiency of VNFs. However, the special architecture of GPU makes existing CPU-oriented task scheduling strategies fail to be applied, limiting the acceleration potential of GPUs. To this end, we propose a GPU-oriented spatio-temporal sharing framework as Gost to improve the performance of GPU-accelerated VNFs. We also study how to minimize the end-to-end latency of VNF flows via careful scheduling on the execution order and the GPU resource allocation (i.e., the number of threads). We first formally describe the problem as a non-linear integer programming problem, which is then equivalently transformed into an integer linear programming (ILP) form. Considering the high computation complexity of solving ILP, we further propose a customized list scheduling based spatio-temporal GPU sharing strategy (LSSTG). We have practically implemented a prototype of Gost, based on which we also verify the high efficiency of LSSTG by extensive experiments.

Summary Dataflow computing has become a promising computing paradigm as an alternative to traditional control‐centric computing paradigm to facilitate big data processing. Big data process often happens in cloud computing environment as the datacenter provisions a large amount of resource. Dataflow computing, as a data‐centric computing paradigm, requires the dataflows to be shuffled among different codelets (ie, data processing units) deployed in the datacenter servers. It is significant to well schedule the dataflow transferring for communication efficiency. It is highly regarded that the datacenter network shall be managed by software defined networking (SDN) technology for flexibility consideration. In SDN managed datacenter, a dataflow requires a forwarding rule in the forwarding table of each switch on its routing path. However, the SDN switches are limited in the forwarding table size. This introduces an unignorable issue in the codelet deployment problem. Therefore, we are motivated to take such forwarding table size constraints into the problem of dataflow codelet deployment in the datacenters managed by SDN. In particular, we aim at minimizing the communication cost efficiency while guarantee the dataflow computing performance at the same time. The communication cost minimization problem is formulated into an integer linear programming form, which is relaxed to design a heuristic algorithm. The experiment results show that our relaxation algorithm can significantly improve the communication cost efficiency via ingenious codelet placement.

The underwater acoustic sensor network (UASN) is a specific deployment of Internet-of-Things (IoT) technology in the underwater environment, since energy constraints limit the lifetime of UASNs, effectively balancing the energy consumption of acoustic sensor nodes in UASNs is important to maximize the amount of information collected and to prolong the network lifetime. Node clustering is widely regarded as one of the most important energy-efficient schemes for UASNs. However, most existing clustering schemes focus on the cooperation-based election of cluster headers (CHs) in a centralized manner. Due to the limited energy capacity, acoustic sensor nodes are designed to save their own energy, hindering the realization of such cooperation. To address this issue in this article, game theory is applied to UASNs to balance network energy consumption and model acoustic sensor nodes as rational and selfish players. Specifically, a game-theory-based clustering (GTC) scheme for UASNs is developed. In the CH election phase, each node makes a decision in pursuit of a greater payoff based on the Nash equilibrium. An incentive mechanism is invented to induce nodes to make more beneficial collective decisions and plays a role in the CH rotation to effectively balance the energy consumption. Meanwhile, the network area is divided into nonuniform sectors to ensure the energy consumption of the CH is more evenly distributed. Simulation results show that the proposed GTC scheme can effectively balance network energy consumption and extend the network lifetime.

Efficient urban traffic monitoring is a key enabler for intelligent planning and management of modern cities. Network tomography can monitor the urban traffic with a comparably small number of traffic detectors like cameras, and has become an appealing technique for urban traffic management. However, previous work on network tomography based traffic monitoring focuses primarily on developing estimators using the given end-to-end travel time measurements, while the design of data collection for efficiently distributed collecting and processing the raw monitoring videos to such measurements is often neglected. We fill this gap by exploring the vision of edge intelligence for optimal urban traffic monitoring, and tackle the following two problems in regard of limited telecommunications resources: 1) when the total number of monitoring videos that are successfully processed into the end-to-end travel time measurements is pre-bounded, we employ a Fisher Information Matrix (FIM) to help determine the best quota scheme for the monitoring videos that each traffic detector need to generate and 2) when the centralised processing of monitoring videos alone is insufficient, we make use of the computation capabilities from these edge devices, i.e., traffic detectors, and employ a multi-agent reinforcement learning approach to help them conduct intelligent computation offloading individually. Extensive simulations demonstrate that our proposed scheme effectively reduces the estimation error of network tomography compared to common approaches with either uniform or random strategy.

Large-scale datacenters often host latency-sensitive services that have stringent Quality-of-Service requirement and experience diurnal load pattern. Co-locating best-effort applications that have no QoS requirement with the latency-sensitive services has been widely used to improve the resource utilization of datacenters with careful shared resource management. However, existing co-location techniques tend to result in the power overload problem on power constrained servers due to the ignorance of the power consumption. To this end, we propose Sturgeon, a runtime system proactively manages resources between co-located applications in a power constrained environment, to ensure the QoS of latency-sensitive services while maximizing the throughput of best-effort applications. Our investigation shows that, at a given load, there are multiple feasible resource configurations to meet both QoS requirement and power budget, while one of them yields the maximum throughput of best-effort applications. To find such a configuration, we establish models to accurately predict the performance and power consumption of the co-located applications. Sturgeon monitors the QoS of the services periodically, in order to eliminate the potential QoS violation caused by the unpredictable interference. Besides, when the datacenter hosts different types of applications to perform co-location, Sturgeon places applications with their preferable candidates to improve the overall throughput. The experimental results show that at server level Sturgeon improves the throughput of the best-effort application by 25.43 percent compared to the state-of-the-art technique, while guaranteeing the 95%-ile latency within the QoS target; at cluster level, Sturgeon improves the overall throughput of best-effort applications by 13.74 percent compared to the baseline.

The accurate identification of congested links is crucial for network performance monitoring. Network boolean tomography uses end-to-end path measurements to identify congested links, and appears as a significant alternative when direct link monitoring is not available. However, most of existing tomographic methods assume no correlations between links, i.e., the congestion of one link is assumed to be independent from the congestion of any others, hindering their applications in practice because links could become correlated during a joint optimization procedure of many network operations like traffic routing and balancing. In this paper, we study practical network boolean tomography without such an assumption. We elaborate on the ill-posed nature of network boolean tomography to highlight the significance of integrating link correlations, and model the congested link identification from end-to-end congestion observations of paths as a problem of Maximum A-Posteriori (MAP) estimation. To avoid the explicit acquisition of any priori knowledge of link correlations, we then propose a learning-based algorithm with Long Short-term Memory (LSTM), a special recurrent neural network that is good at learning statistical dependencies of sequence elements from historical data. Numerical results over real network topologies validate our learning-based network boolean tomography.

Compute First Networking (CFN) recently is proposed as an in-network computing paradigm for well balancing between the networking and computation resource scheduling. Thanks to the proliferation of network functions virtualization, the virtualized network functions can coexist with the computing services on a shared platform like edge computing environment. Thus, one critical issue incurred by CFN is how to manage and schedule the resources among various services from different over-the-top service provider (OSP) with different resource requirements, i.e., network slicing. In this paper, we first formulate the network slicing problem as a Stackelberg game problem and prove that there exists a Nash equilibrium beneficial to both the Network Slice Broker (NSB) and OSP. Furthermore, we propose a cooperative game model on the networking and computation resource allocation within each slice and invent a Nash bargaining solution to resolve the intra-slice resource competition for slice performance promotion. Simulation results are provided to validate the effectiveness and high efficiency of the our proposed game based network slicing and resource scheduling algorithm.

With the rapid development and popularization of IoT devices in our daily lives, secure data transmission has become a critical and challenging issue in resource-constrained devices. Simple chaotic circuits can provide highly efficient data encryption and transmission simultaneously. Time delay feedback system is considered as a promising solution to meet the demand for generating highly complex chaotic signal. The time delay is an important parameter that significantly influences the security of such systems. Although there exist many methods that can extract this parameter by adopting the leaked chaotic time series and certain information of the system model, they more or less face some restrictions, especially when the system nonlinearity is quite strong. To break through these limitations, we propose a novel method that can extract the key (i.e., time delay signature, TDS) of time delay chaotic systems by machine learning strategy. The 1D time series transmitted from a time delay chaotic system is embedded in 2D planes to generate training images. A convolutional neural network (CNN) is explored to learn from these images to identify common and different features for TDS extraction. Simulation results verify the feasibility and efficiency of the proposed scheme.

While deep neural network (DNN) models are often trained on GPUs, many companies and research institutes build GPU clusters that are shared by different groups. On such GPU cluster, DNN training jobs also require CPU cores to run pre-processing, gradient synchronization. Our investigation shows that the number of cores allocated to a training job significantly impact its performance. To this end, we characterize representative deep learning models on their requirement for CPU cores under different GPU resource configurations, and study the sensitivity of these models to other CPU-side shared resources. Based on the characterization, we propose CODA, a scheduling system that is comprised of an adaptive CPU allocator, a real-time contention eliminator, and a multi-array job scheduler. Experimental results show that CODA improves GPU utilization by 20.8% on average without increasing the queuing time of CPU jobs.

To tackle the computation resource poorness on the end devices, task offloading is developed to reduce the task completion time and improve the Quality-of-Service (QoS). Edge cloud facilitates such offloading by provisioning resources at the proximity of the end devices. Modern applications are usually deployed as a chain of subtasks (e.g., microservices) where a special offloading strategy, referred as binary offloading, shall be applied. Binary offloading divides the chain into two parts, which will be executed on end device and the edge cloud, respectively. The offloading point in the chain therefore is critical to the QoS in terms of task completion time. Considering the system dynamics and algorithm sensitivity, we apply Q-learning to address this problem. In order to deal with the late feedback problem, a reward rewind match strategy is proposed to customize Q-learning. Trace-driven simulation results show that our customized Q-learning based approach is able to achieve significant reduction on the total execution time, outperforming traditional offloading strategies and non-customized Q-learning.

To tackle the computation resource poorness on the end devices, task offloading is developed to reduce the task completion time and improve the Quality-of-Service (QoS). Edge computing facilitates such offloading by provisioning resources at the proximity of the end devices. Nowadays, many tasks on end devices have an urgent demand for the security of execution environment. To address this problem, we introduce trusted execution environment (TEE) to empower edge computing for secure task offloading. To explore TEE, the offloading process should be redesigned with the introduction of data encryption and decryption. This makes traditional offloading optimization policy fail to be applied directly. To address this issue, we are motivated to take the data encryption and decryption into the offloading scheduling algorithm. In particular, we propose a Customized List Scheduling based Offloading (CLSO) algorithm, aiming at minimizing the total completion time with the consideration of energy budget limitations on the end devices. The experiment results show that our approximation algorithm can effectively reduce the total completion time and significantly outperforms existing state-of-the-art offloading strategy.

Offloading cellular traffic through Opportunistic Mobile Networks (OMNs) is an effective way to relieve the burden of cellular networks. Providing data offloading services requires a lot of resources, and nodes in OMNs are selfish and rational, they are not willing to provide data offloading services for others without any compensation. Therefore, it is urgent to design an incentive mechanism to stimulate mobile nodes to participate in data offloading process. In this paper, we propose a Reverse Auction-based Incentive Mechanism to stimulate mobile nodes in OMNs to provide data offloading services, and take the cache management into consideration. We model the incentive-driven data offloading process as a non-linear integer programming problem, then a Greedy Helper Selection Method (GHSM) and a Caching Replacement Scheme (CRS) are proposed to solve the problem. In addition, we also propose an innovative payment rule based on the Vickrey-Clarke-groves (VCG) model to ensure the individual rationality and authenticity of the proposed algorithm. Trace-driven simulation results show that the proposed algorithm can reduce the cost of Content Service Provider (CSP) significantly in different scenarios.

Large-scale datacenters often host latency-sensitive services that have stringent Quality-of-Service requirement and experience diurnal load pattern. Co-locating best-effort applications that have no QoS requirement with latency-sensitive services has been widely used to improve the resource utilization with careful shared resource management. However, existing co-location techniques tend to result in the power overload problem on power constrained computers due to the ignorance of the power consumption. To this end, we propose Sturgeon, a runtime system proactively manages resources between colocated applications in a power constrained environment, to ensure the QoS of latency-sensitive services while maximizing the resource utilization. Our investigation shows that, at a given load, there are multiple feasible resource configurations to meet both QoS requirement and power budget, while one of them yields the maximum throughput of best-effort applications. To find such a configuration, we establish models to accurately predict the performance and power consumption of the colocated applications. Sturgeon monitors the QoS periodically in order to eliminate the potential QoS violation caused by the unpredictable interference. The experimental results show that Sturgeon improves the throughput of best-effort applications by 24.96% compared to the state-of-the-art technique, while guaranteeing the 95%-ile latency within the QoS target.

Federated Learning (FL) takes advantage of distributed data to jointly train a global deep learning model on many clients, without revealing local data to the central server for privacy guarantee. However, due to the heterogeneity of the FL clients, some poor performance clients may become stragglers, impeding the global training process. It is desirable to offload these stragglers' tasks to some high performance servers, but this is at the risk of data privacy leakage. To mitigate such problem, we introduce the edge servers empowered by Trusted Execution Environment (TEE) to securely help the FL clients with poor performance. With the consideration of limited computation resource in TEE, we further investigate how to select the clients for help. Considering the time-varying processing capabilities on the FL clients, we propose an exploration-exploitation based client selection algorithm. Via evaluating our algorithm in a practical FL training task, the experiments show that the proposed algorithm indeed accelerate training process thanks to its efficient client selection.

Edge caching has been emerged as a promising solution to alleviate the redundant traffic and the content access latency in the future Internet of Vehicles (IoVs). Several Reinforcement Learning (RL) based edge caching methods have been proposed to improve the cache utilization and reduce the backhaul traffic load. However, they can only obtain the local sub-optimal solution, as they neglect the influence of environment by other agents. In this paper, we investigate the edge caching strategy with consideration of the content delivery and cache replacement by exploiting the distributed Multi-Agent Reinforcement Learning (MARL). We first propose a hierarchical edge caching architecture for IoVs and formulate the corresponding problem with the objective to minimize the long-term cost of content delivery in the system. Then, we extend the Markov Decision Process (MDP) in the single agent RL to the multi-agent system, and propose a distributed MARL based edge caching algorithm to tackle the optimization problem. Finally, extensive simulations are conducted to evaluate the performance of the proposed distributed MARL based edge caching method. The simulation results show that the proposed MARL based edge caching method significantly outperforms other benchmark methods in terms of the total content access cost, edge hit rate and average delay. Especially, our proposed method greatly reduces an average of 32% total content access cost compared with the conventional RL based edge caching methods.

Summary Graph data processing has been widely applied in a variety of domains such as industry, science, social network, and so on. It therefore has stimulated many efforts devoted to this area. To embrace the fast development trend of big graph data, graph data processing based on Pregel‐like systems has been regarded as one of the most promising ways and has widely attracted the attention of researchers. However, it still remains in its early stage and there still exist many challenges. In Pregel, the superstep synchronization is time consuming as the graph data iteration operation requires multiple synchronizations. Furthermore, the graph data partition strategy adopted by Pregel fails to support load balancing, therefore causing the increase of network I/O overhead as the scale of graph data grows. To address these issues, this paper presents an efficient computational framework for graph data processing based on the bulk synchronous parallel model. The global synchronization control mechanism is improved by determining the start time of the next round of superstep through counting the number of global message files. Furthermore, an improved graph data partition mechanism based on a balanced hash method is proposed to reduce the communication overhead between different partitions of sub‐graph computational tasks. We also re‐design the PageRank algorithm to verify the effectiveness of the proposed framework. Experimental results on different real‐world datasets verify the efficiency of our proposed framework as it outperforms Giraph (an open source Pregel‐like system) by 58%−69%, and achieves 10×−17× performance improvement over Hadoop.

Internet of Things (IoT) generates large amounts of data at the network edge. Machine learning models are often built on these data, to enable the detection, classification, and prediction of the future events. Due to network bandwidth, storage, and especially privacy concerns, it is often impossible to send all the IoT data to the data center for centralized model training. To address these issues, federated learning has been proposed to let nodes use the local data to train models, which are then aggregated to synthesize a global model. Most of the existing work has focused on designing learning algorithms with provable convergence time, but other issues, such as incentive mechanism, are unexplored. Although incentive mechanisms have been extensively studied in network and computation resource allocation, yet they cannot be applied to federated learning directly due to the unique challenges of information unsharing and difficulties of contribution evaluation. In this article, we study the incentive mechanism for federated learning to motivate edge nodes to contribute model training. Specifically, a deep reinforcement learning-based (DRL) incentive mechanism has been designed to determine the optimal pricing strategy for the parameter server and the optimal training strategies for edge nodes. Finally, numerical experiments have been implemented to evaluate the efficiency of the proposed DRL-based incentive mechanism.

Hosting virtualized network functions (VNF) has been regarded as an effective way to realize network function virtualization (NFV). Considering the cost diversity in cloud computing, from the perspective of service providers, it is significant to orchestrate the VNFs and schedule the traffic flows for network utility maximization (NUM) as it implies maximal revenue. However, traditional heuristic solutions based on optimization models usually follow some assumptions, limiting their applicability. Recent studies have shown that deep reinforcement learning (DRL) is a promising way to tackle such limitations. However, DRL agent training also suffers from slow convergence problem, especially with complex control problems. We notice that optimization models actually can be applied to accelerate the DRL training. Therefore, we are motivated to design a model-assisted DRL framework for VNF orchestration in this paper. Other than letting the agent blindly explore actions, the heuristic solutions are used to guide the training process. Based on such principle, the DRL framework is also redesigned accordingly. Experiment results validate the high efficiency of our model-assisted DRL framework as it not only converges 23× faster than traditional DRL algorithm, but also with higher performance at the same time.

Summary Burst or leakage in drinkable water distribution system has occurred frequently in recent years, causing severe damages, economic loss, and long‐lasting society impact. A viable solution is to use agile inline mobile sensors to detect and so as to mitigate the burst or leakage. Distinguishing from online fixed sensors, mobile sensors can swim freely along the piles in water distribution network, thus giving a more precise detection. To combat the low power, low computation, and low communication capability of mobile sensors, the newly emerged fog computing provides a promising means to gather and preprocess the sensing data. In practice, due to the budget limitation, we can deploy a limited number of sensors and fog nodes in the system. This introduces a challenging problem on how to deploy them in the system, ie, sensor and fog node placement. We first formulate mobile sensor placement (MSP) as a path cover problem and prove it as NP‐complete, and then we propose a customized genetic algorithm and a mixed greedy algorithm to solve MSP and fog node placement, respectively. The correctness and efficiency of the proposed algorithm are illustrated by a comprehensive experiment. Moreover, some critical factors, eg, sensor battery lifetime and movement pattern, are all extensively investigated and the results show the coverage ratio is sensitive to these factors.

Edge computing is an effective approach for resource provisioning at the network edge to host virtualized network functions (VNF). Considering the cost diversity in edge computing, from the perspective of service providers, it is significant to orchestrate the VNFs and schedule the traffic flows for network utility maximization (NUM) as it implies maximal revenue. However, traditional model-based optimization methods usually follow some assumptions and impose certain limitations. In this paper, inspired by the success of deep reinforcement learning in solving complicated control problems, we propose a deep deterministic policy gradients (DDPG) based algorithm. We first formulate the NUM problem with the consideration of end-to-end delays and various operation costs into a non-convex optimization problem and prove it to be NP-hard. We then redesign the exploration method and invent a dual replay buffer structure to customize the DDPG. Meanwhile, we also apply our formulation to guide our replay buffer update. Through extensive trace-driven experiments, we show the high efficiency of our customized DDPG based algorithm as it significantly outperforms both model-based methods and traditional non-customized DDPG based algorithm.

In Software Defined Networks (SDNs), one of the key tasks in control plane is the monitoring and measurement of the whole network. A typical SDN consists of a set of switches and a logically centralized controller responsible for network state monitoring and flow scheduling, to which low cost and efficient flow statistics collection plays an important role. However, existing flow statistics collection methods mainly focus on single-path routing and cannot accurately capture the flow statistics in the case of multi-path routing (MPR), which is widely used in modern networks. In this paper, we are motivated to propose a Multi-Path oriented Flow Statistics Collection (MFSC) strategy to minimize the total communication cost for flow statistics collection. The problem is first formulated into an integer linear programming (ILP) form. After analyzing the complexity of this problem, we present a relaxation-based algorithm with an approximation factor p, where p is the maximum number of switches passed by each flow. The experiment results demonstrate that our proposed algorithm can reduce the total communication cost by over 14% compared with the traditional solutions.

Network congestion will result in significant performance degradation or even failures of many bandwidth-hungry Internet of Things (IoT) applications. Accurate and efficient congested link identification has become a foundational issue to IoT applications like self-driving cars, digital health, smart city, and so on. However, directly monitoring the massive number of interior links often introduces high operation cost or even is infeasible in practice, giving rise to indirect monitoring techniques like network Boolean tomography. Nevertheless, in many networks, the number of their interior links is larger than their end-to-end paths, making it very challenging for network Boolean tomography to find a determined solution. To resolve this issue, most of current methods try to utilize some prerequisites, such as the link congestion probabilities. While these probabilities might be hard or even unable to be obtained accurately in dynamical networks, limiting the practical deployment. In this paper, we are motivated to design a framework of congested link identification without any prerequisite or assumption. We first novelly model the congested link identification procedures as a Markov decision processes (MDPs), and then employ a reinforcement learning technology, i.e., Q-learning, to solve this MDP. The simulation results show that our proposed scheme can autonomously and efficiently explore the unknown network environment, and is able to achieve better adaptivity and correctness, without any prior knowledge comparing to existing methods.

Network function virtualization (NFV) decouples the traditional network functions from specific or proprietary hardware, such that virtualized network functions (VNFs) can run in software form. By exploring NFV, a consecutive set of VNFs can constitute a service function chain (SFC) to provide the network service. From the perspective of network service providers, how to maximize the network utility is always one of the major concerns. To this end, there are two main issues need to be considered at runtime: 1) how to handle the unpredictable network traffic burst? and 2) how to fairly allocate resources among various flows to satisfy different traffic demands? In this paper, we investigate a fairness-aware flow scheduling problem for network utility maximization, with joint consideration of resource allocation and rate control. Based on a discrete-time queuing model, we propose a low-complexity online-distributed algorithm using the Lyapunov optimization framework, which can achieve arbitrary optimal utility with different fairness levels by tuning the fairness bias parameter. We theoretically analyze the optimality of the algorithm and evaluate its efficiency by both simulation and testbed-based experiments.

Big data analytics in datacenters often involve scheduling of data-parallel jobs. Traditional scheduling techniques based on improving network resource utilization are subject to limited bandwidth in datacenter networks. To alleviate the shortage of bandwidth, some cluster frameworks employ techniques of traffic compression to reduce transmission consumption. However, they tackle scheduling in a coarse-grained manner at task level and do not perform well in terms of flow-level metrics due to high complexity. Fortunately, the abstraction of coflow pioneers a new perspective to facilitate scheduling efficiency. In this paper, we introduce a coflow compression mechanism to minimize the completion time in data-intensive applications. Due to the NP-hardness, we propose a heuristic algorithm called Fastest-Volume-Disposal-First (FVDF) to solve this problem. For online applicability, FVDF supports stage pipelining to accelerate scheduling and exploits recurrent neural networks (RNNs) to predict compression speed. Meanwhile, we build Swallow, an efficient scheduling system that implements our proposed algorithms. It minimizes coflow completion time (CCT) while guaranteeing resource conservation and starvation freedom. The results of both trace-driven simulations and real experiments show the superiority of our algorithm, over existing one. Specifically, Swallow speeds up CCT and job completion time (JCT) by up to 1.47χ and 1.66χ on average, respectively, over the SEBF in Varys, one of the most efficient coflow scheduling algorithms so far. Moreover, with coflow compression, Swallow reduces data traffic by up to 48.41 percent on average.

In cyber-physical social systems (CPSS), a group of volunteers report data about the physical environment through their cyber devices and data aggregation is widely utilized. An important issue in data aggregation for CPSS is to protect users’ privacy. In this article, we use bitwise XOR and propose a bit-choosing algorithm to realize privacy-preserving min, k -th min, and percentile computation. By our algorithm, the aggregator can confirm whether a user’s data value is equal to certain value or within certain scale. Consequently, it is also possible to count the number of users satisfying given conditions. Our bit-choosing algorithm makes sure that the users send non-repetition replies to the aggregator to raise the aggregation accuracy. We analyze the communication cost and the achievable accuracy of our algorithm. Via performance comparison against existing protocols, the efficiency and accuracy of our algorithm are verified.

Fuel cells are promising power sources for green data centers thanks to its high energy-efficiency, low greenhouse gas emissions and high reliability. However, fuel cells have a unique feature called limited load following, i.e., they are slow in adjusting power supply due to mechanical limitation of fuel delivery. When power demand of data centers suddenly grows, fuel cells would fail to provide sufficient power supply. On the other hand, fuel cells are slow to reduce its power supply when demand decreases, leading to energy waste. In this paper, we study to mitigate the impact of limited load following by associating a set of heterogeneous batteries with fuel cells. These batteries with different characteristics (e.g., capacity, charging and discharging rate) can power data centers when the energy supply of fuel cells is insufficient. They are charged by excessive power supply when demand decreases. Given future power demand, we formulate the energy management problem as a mixed-integer nonlinear programming. An online algorithm is designed to solve the problem without future knowledge. We conduct extensive simulations using real-world traces and results show that our proposed algorithm significantly outperforms existing solutions.

This paper describes the design of a new software zero-copy network framework for NVM-assisted key-value stores, which directly stores and persists transactions from network into raw non-volatile memory used as write-ahead cache for data consistency. NVM is fast and bit-addressable which makes it the perfect choice for transient transaction log persistency than hard disks or even Flash drives, but its limited write cycle requires wear-leveling during direct access. However, popular RDMA-based zero-copy transmission normally needs to have the remote memory address beforehand and cannot cope with the address changing caused by wear-leveling easily. The software zero-copy solution proposed in this paper is designed with the awareness of NVM wear-leveling and log metadata management. Simulation results show that the new network framework improves performance by over 200× in throughput and decreases latency by more than 20× comparing to the traditional socket and hard disk based solution. When both equipped with NVM, the zero-copy network stack improves performance by 18 to 62% in throughput and 40 to 81% in latency comparing to the standard socket and with the lowest CPU consumption.

Big data analytics in datacenters often involves scheduling of data-parallel job, which are bottlenecked by limited bandwidth of datacenter networks. To alleviate the shortage of bandwidth, some existing work has proposed traffic compression to reduce the amount of data transmitted over the network. However, their proposed traffic compression works in a coarse-grained manner at job level, leaving a large optimization space unexplored for further performance improvement. In this paper, we propose a flow-level traffic compression and scheduling system, called Swallow, to accelerate data-intensive applications. Specifically, we target on coflows, which is an elegant abstraction of parallel flows generated by big data jobs. With the objective of minimizing coflow completion time (CCT), we propose a heuristic algorithm called Fastest-Volume-Disposal-First (FVDV) and implement Swallow based on Spark. The results of both trace-driven simulations and real experiments show the superiority of our system, over existing algorithms. Swallow can reduce CCT and job completion time (JCT) by up to 1.47 × and 1.66 × on average, respectively, over the SEBF in Varys, one of the most efficient coflow scheduling algorithms so far. Moreover, with coflow compression, Swallow reduces data traffic by up to 48.41% on average.

Network Function Virtualization (NFV) emerges as a promising technology to increase the network flexibility, customizability and efficiency by softwarizing traditional dedicated hardware based functions to virtualized network functions. The prosperous potential of edge cloud makes it an ideal platform to host the network functions. From the perspective of network service providers, an inevitable concern is how to reduce the overall cost for renting various resources from infrastructure providers. In this paper, unlike existing related studies assuming a preknown network traffic demand, we alternatively consider a practical case without any prior knowledge. We investigate how to dynamically minimize the overall operational cost with joint consideration of packet scheduling, network function management and resource allocation. The tradeoff between the queue backlog and overall cost is analyzed using a Lyapunov optimization framework. A backpressure based online scheduling algorithm is proposed and its efficiency is extensively evaluated by trace-driven simulations.

The recent progress in big data has inspired lots of data- parallel applications deployed in the datacenters. Although how to optimize the data flow scheduling in datacenters has been extensively studied, traditional per-flow based optimizations usually do not perform well in dealing with the transferring of a collection of parallel flows, i.e., coflow. Consequently, how to schedule the coflow towards various objectives, e.g., minimizing the coflow completion time, has attracted much attention recently. We notice that existing coflow scheduling studies usually suggest a fixed destination for each coflow. Taking the advantage of virtualization technology, we argue that the destination can be flexibly placed in the cloud. Therefore, it is essential to jointly optimize the coflow scheduling and data processor placement. In this paper, we are motivated to investigate the problem of coflow completion time minimization with joint consideration of coflow scheduling and data processor placement. We first formally describe the problem into a mixed integer non-linear programming (MINLP) problem. By linearizing the MINLP, we further propose a relaxation based heuristic algorithm. Via extensive simulation studies, the high efficiency of our heuristic algorithm is validated.

Emerging technologies such as Software-Defined Networks (SDN) and Network Function Virtualization (NFV) promise to address cost reduction and flexibility in network operation while enabling innovative network service delivery. To catch up with the time- varying traffic demands, the network changes frequently. We should come up with a sequence of instructions to manipulate the starting network into the goal network, while preserving the network semantics correctness (e.g., freedom of loops, bandwidth guaranteeing). In this case, how to migrate the virtual network functions (VNF) and update the flow forwarding rules efficiently is an important and challenging problem. In this paper, we are motivated to address the migration of VNF and flow update rule problem with joint consideration of migration cost and update delay. The problem is first formulated into a mixed integer non-linear programming (MINLP). By linearizing and relaxing the MINLP, we then present a polynomial-time two-stage heuristic algorithm. The high efficiency of our algorithm is extensively validated by simulation based studies by the fact that it performs much closer to the optimal solution.

By softwarizing traditional dedicated hardware based functions to virtualized network functions (VNFs) that can run on standard commodity servers, network function virtualization (NFV) technology promises high efficiency, flexibility and scalability. To NFV service providers, one primary concern is to maximize network throughput and reduce service time. To reach this goal, two main challenges should be tackled: 1) how to schedule the unpredictable and burst network flows; 2) how to fairly allocate resources between various flows with different resource requirements. In this paper, we are motivated to investigate a throughput maximization problem with joint consideration of fairness between multiple flows using a discrete time queuing model. By taking advantages of Lyapunov optimization techniques, we propose a low-complexity online distributed algorithm that can achieve arbitrary optimal utility with different fairness levels by tuning the fairness bias. The high efficiency of our proposal is validated by both theoretical analysis and extensive simulation studies.

Wireless networks have experienced fast development in the past decades. Various advancing wireless technologies have been proposed. To catch up with the ever-increasing diverse communication needs, cloud-radio access networks (C-RAN), which decouples the baseband processing unit (BBU) from the remote radio head (RRH), has been proposed. On the other hand, it has been widely recognized that huge energy consumption has been raised due to the massive deployment of cellular networks. Lowering the network energy consumption therefore becomes a widely concerned topic. To combat the limitations in traditional power grid, smart grid, with the emphasis on distributed energy resource (DER) and bidirectional energy sharing, is advocated to power the wireless networks. In this paper, we are motivated to investigate a joint RRH-BBU association and energy sharing problem towards brown energy usage minimization in green energy powered C-RAN. The problem is formulated into a mixed integer linear programming (MILP) form. To address the computation complexity of solving MILP, a two-phase heuristic polynomial- time algorithm is proposed and evaluated via extensive simulation based studies.

Summary Both mobile computing and cloud computing have experienced rapid development in recent years. Although centralized cloud computing exhibits abundant resources for computation‐intensive tasks, the unpredictable and unstable communication latency between the mobile users and the cloud makes it challenging to handle latency‐sensitive mobile computing tasks. To address this issue, fog computing recently was proposed by pushing the cloud computing to the network edge closer to the users. To realize such vision, we can augment existing access points in wireless networks with cloudlet servers for hosting various mobile computing tasks. In this paper, we investigate how to deploy the servers in a cost‐effective manner without violating the predetermined quality of service. In particular, we practically consider that the available cloudlet servers are heterogeneous, ie, with different cost and resource capacities. The problem is formulated into an integer linear programming form, and a low‐complexity heuristic algorithm is invented to address it. Extensive simulation studies validate the efficiency of our algorithm by it performs much close to the optimal solution.

Energy Internet (EI) is proposed as the evolution of smart grid, aiming to integrate various forms of energy into a highly flexible and efficient grid that provides energy packing and routing functions, similar to the Internet. As an essential part in EI system, a scalable and interoperable communication infrastructure is critical in system construction and operation. In this article, we survey the recent research efforts on EI communications. The motivation and key concepts of EI are first introduced, followed by the key technologies and standardizations enabling the EI communications as well as security issues. Open challenges in system complexity, efficiency, reliability are explored and recent achievements in these research topics are summarized as well.

The rapid development of cloud computing has raised big concerns over the high energy consumption of modern data centers. To satisfy the ever increasing data traffic needs, the energy consumption of data center network (DCN) also takes a significant proportion. The newly emerging technology, Software Defined Networking (SDN), which allows flexible control of network devices, brings a new opportunity towards DCN energy optimization. In this paper, we investigate how to design an energy-efficient network management strategy with guaranteed satisfaction of network traffic demands in Software Defined Data Center Networks (SD-DCNs). To this end, three issues will be tackled: 1) the subset of switches that shall be activated, i.e., switch activation, 2) multi-path routing scheduling for all flows and 3) forwarding rule placement in SDN switches. They are jointly considered and formulated as an integer linear programming (ILP) problem. A heuristic algorithm to deal with its high computational complexity is proposed. Extensive simulation-based evaluations are conducted to validate the high efficiency of our algorithm.

In past decades, we have witnessed lots of gas leakage diasters all over the world, causing serious casualties, property damage and severe negative social impact. During evacuation after gas leakage incident, the poisonous gas shall be accurately detected. Accordingly, the evacuation routine shall be carefully planned and warned to the evacuating people. Unmanned aerial vehicle (UAV) has been widely regarded as a promising tool to support crowd evacuation. In this paper, aiming at providing an efficient UAV control system for crowd evacuation, we propose a crowd simulation based UAV control system to direct crowd evacuation from the polluted area. The architecture mainly consists of a UAV fleet management module, UAV trajectory planning module, and a crowd simulation module. The architecture forms a closed control loop, emphasizing the inter-operation between the real scenario and the simulation scenario. A case study on UAV gas leakage detection scheduling is given. The results show that the proposed architecture can provide efficient way to help pedestrian in the environment to evacuate and avoid to be infected by the toxic gas.

Summary MapReduce has been widely regarded as a flexible, scalable, and easy‐to‐use distributed programming paradigm for big data processing such as social network data analysis on cloud computing platforms. To embrace the upcoming of big data era, many efforts have been devoted to accelerating the MapReduce performance from different aspects, especially intermediate result reusing like Dache. In this paper, we observe that existing intermediate result reusing mechanism is not efficient enough as many I/O operations are wasted. Efficient reusing of the intermediate results could potentially improve the MapReduce performance. Inspired by such fact, we propose a framework named MEMoMR (more efficient intermediate result reusing for MapReduce) by introducing a novel reusing mechanism that can substantially reduce the I/O overhead. To this end, we invent a new metadata description method and apply it in the reusing phase. We practically realize MEMoMR and evaluate its performance by implementing it in a real cluster. The experiment results show that MEMoMR can improve the system performance as high as 23.4%, comparing against Dache. Copyright © 2015 John Wiley & Sons, Ltd.

With the explosion of big data, processing large numbers of continuous data streams, i.e., big data stream processing (BDSP), has become a crucial requirement for many scientific and industrial applications in recent years. By offering a pool of computation, communication and storage resources, public clouds, like Amazon's EC2, are undoubtedly the most efficient platforms to meet the ever-growing needs of BDSP. Public cloud service providers usually operate a number of geo-distributed datacenters across the globe. Different datacenter pairs are with different inter-datacenter network costs charged by Internet Service Providers (ISPs). While, inter-datacenter traffic in BDSP constitutes a large portion of a cloud provider's traffic demand over the Internet and incurs substantial communication cost, which may even become the dominant operational expenditure factor. As the datacenter resources are provided in a virtualized way, the virtual machines (VMs) for stream processing tasks can be freely deployed onto any datacenters, provided that the Service Level Agreement (SLA, e.g., quality-of-information) is obeyed. This raises the opportunity, but also a challenge, to explore the inter-datacenter network cost diversities to optimize both VM placement and load balancing towards network cost minimization with guaranteed SLA. In this paper, we first propose a general modeling framework that describes all representative inter-task relationship semantics in BDSP. Based on our novel framework, we then formulate the communication cost minimization problem for BDSP into a mixed-integer linear programming (MILP) problem and prove it to be NP-hard. We then propose a computation-efficient solution based on MILP. The high efficiency of our proposal is validated by extensive simulation based studies.

Traditional standalone embedded system is limited in their functionality, flexibility, and scalability. Fog computing platform, characterized by pushing the cloud services to the network edge, is a promising solution to support and strengthen traditional embedded system. Resource management is always a critical issue to the system performance. In this paper, we consider a fog computing supported software-defined embedded system, where task images lay in the storage server while computations can be conducted on either embedded device or a computation server. It is significant to design an efficient task scheduling and resource management strategy with minimized task completion time for promoting the user experience. To this end, three issues are investigated in this paper: 1) how to balance the workload on a client device and computation servers, i.e., task scheduling, 2) how to place task images on storage servers, i.e., resource management, and 3) how to balance the I/O interrupt requests among the storage servers. They are jointly considered and formulated as a mixed-integer nonlinear programming problem. To deal with its high computation complexity, a computation-efficient solution is proposed based on our formulation and validated by extensive simulation based studies.

In recent years, water pollution or contamination incidents happened frequently, causing serious disasters and negative social impact. To reduce the water contamination risk, water quality monitoring sensors should be deployed in water distribution system (WDS) to enable real-time pollution detection. It is desirable to deploy sensors everywhere so that any contamination event can be detected and reported in a timely manner. Unfortunately, this is a luxury and unrealistic vision because of high deployment cost. It is significant to lower the deployment cost provided that the quality-of-sensing, e.g., coverage and contamination detection time, can be guaranteed for effective depollution action. In this paper, we consider a water quality monitoring sensor network consisting of two kinds of sensors with different prices. The expensive one is of cellular communication capability and therefore is able to send sensing information to control center directly, while the cheaper one is of only sensor-to-sensor communication capability. We investigate a cost-efficient sensor deployment problem on how to deploy these two kinds of sensors in a given WDS to minimize the deployment cost, without violating the quality-of-sensing requirement. We first formulate the problem into a mixed integer quadratically constrained programming problem, which is then linearized into an equivalent mixed integer linear programming. We further propose a polynomial two-stage heuristic algorithm and evaluate its efficiency via extensive simulation-based studies.

Summary Vehicle Ad‐Hoc Networks (VANET) enable all components in intelligent transportation systems to be connected so as to improve transport safety, relieve traffic congestion, reduce air pollution, and enhance driving comfort. The vision of all vehicles connected poses a significant challenge to the collection, storage, and analysis of big traffic‐related data. Vehicular cloud computing, which incorporates cloud computing into vehicular networks, emerges as a promising solution. Different from conventional cloud computing platform, the vehicle mobility poses new challenges to the allocation and management of cloud resources in roadside cloudlet. In this paper, we study a virtual machine (VM) migration problem in roadside cloudlet‐based vehicular network and unfold that (1) whether a VM shall be migrated or not along with the vehicle moving and (2) where a VM shall be migrated, in order to minimize the overall network cost for both VM migration and normal data traffic. We first treat the problem as a static off‐line VM placement problem and formulate it into a mixed‐integer quadratic programming problem. A heuristic algorithm with polynomial time is then proposed to tackle the complexity of solving mixed‐integer quadratic programming. Extensive simulation results show that it produces near‐optimal performance and outperforms other related algorithms significantly. Copyright © 2015 John Wiley & Sons, Ltd.

With the rising demands on cloud services, the electricity consumption has been increasing drastically as the main operational expenditure (OPEX) to data center providers. The geographical heterogeneity of electricity prices motivates us to study the task placement problem over geo-distributed data centers. We exploit the dynamic frequency scaling technique and formulate an optimization problem that minimizes OPEX while guaranteeing the quality-of-service, i.e, the expected response time of tasks. Furthermore, an optimal solution is discovered for this formulated problem. The experimental results show that our proposal achieves much higher cost-efficiency than the traditional resizing scheme, i.e, by activating/deactivating certain servers in data centers.

In Delay Tolerant Networks (DTNs), the connections between mobile nodes are always disrupted and constant end-to-end paths rarely exist. In order to cope with these communication challenges, most existing DTN routing algorithms favour the “multi-hop forwarding” fashion where a message can be forwarded by multiple relay nodes in the hope that one of the employed relay nodes can deliver the message to the destination node. Since aggressively employing relay nodes may incur the intolerable delivery cost in DTNs, it is meaningful to design a cost-efficient routing algorithm that can achieve a high delivery performance. In this paper, we first design a novel delivery metric to measure the forwarding capability of nodes. Then, we utilize small-world properties to design the principles of relay node selection, e.g., limiting the number of relays and finding the appropriate relay nodes, and further develop a cost-efficient social-aware forwarding algorithm called TBSF. Extensive simulations on real mobility traces are conducted to evaluate the performance of TBSF, and the results demonstrate its efficiency and usefulness.

The ever-growing cellular traffic demand has laid a heavy burden on cellular networks. The recent rapid development in vehicle-to-vehicle communication techniques makes vehicular delay-tolerant network (VDTN) an attractive candidate for traffic offloading from cellular networks. In this paper, we study a bulk traffic offloading problem with the goal of minimizing the cellular communication cost under the constraint that all the subscribers receive their desired whole content before it expires. It needs to determine the initial offloading points and the dissemination scheme for offloaded traffic in a VDTN. By novelly describing the content delivery process via a contact-based flow model, we formulate the problem in a linear programming (LP) form, based on which an online offloading scheme is proposed to deal with the network dynamics (e.g., vehicle arrival/departure). Furthermore, an offline LP-based analysis is derived to obtain the optimal solution. The high efficiency of our online algorithm is extensively validated by simulation results.

Vehicular Cyber-Physical System (VCPS) provides CPS services via exploring the sensing, computing and communication capabilities on vehicles. VCPS is deeply influenced by the performance of the underlying vehicular network with intermittent connections, which make existing routing solutions hardly to be applied directly. Epidemic routing, especially the one using random linear network coding, has been studied and proved as an efficient way in the consideration of delivery performance. Much pioneering work has tried to figure out how epidemic routing using network coding (ERNC) performs in VCPS, either by simulation or by analysis. However, none of them has been able to expose the potential of ERNC accurately. In this paper, we present a stochastic analytical framework to study the performance of ERNC in VCPS with intermittent connections. By novelly modeling ERNC in VCPS using a token-bucket model, our framework can provide a much more accurate results than any existing work on the unicast delivery performance analysis of ERNC in VCPS. The correctness of our analytical results has also been confirmed by our extensive simulations.

After a decade of extensive research on application-specific wireless sensor networks (WSNs), the recent development of information and communication technologies makes it practical to realize the software-defined sensor networks (SDSNs), which are able to adapt to various application requirements and to fully explore the resources of WSNs. A sensor node in SDSN is able to conduct multiple tasks with different sensing targets simultaneously. A given sensing task usually involves multiple sensors to achieve a certain quality-of-sensing, e.g., coverage ratio. It is significant to design an energy-efficient sensor scheduling and management strategy with guaranteed quality-of-sensing for all tasks. To this end, three issues are investigated in this paper: 1) the subset of sensor nodes that shall be activated, i.e., sensor activation, 2) the task that each sensor node shall be assigned, i.e., task mapping, and 3) the sampling rate on a sensor for a target, i.e., sensing scheduling. They are jointly considered and formulated as a mixed-integer with quadratic constraints programming (MIQP) problem, which is then reformulated into a mixed-integer linear programming (MILP) formulation with low computation complexity via linearization. To deal with dynamic events such as sensor node participation and departure, during SDSN operations, an efficient online algorithm using local optimization is developed. Simulation results show that our proposed online algorithm approaches the globally optimized network energy efficiency with much lower rescheduling time and control overhead.

Service provisioning has been widely regarded as a critical issue to quality-of-service (QoS) of cloud services in datacenters. Conventional studies on service provisioning mainly focus on task mapping, i.e., how to distribute the service-oriented tasks onto the servers to achieve different goals, e.g., makespan minimization. In distributed datacenters, a task is usually routed from its generation point (i.e., control room) to the designated server within a datacenter network. Since the routing delay also has a deep influence on the task makespan, we are motivated to study how to minimize the maximum makespan of all tasks in a duty period by joint optimization of both task mapping and routing. It is formulated as an integer programming with quadratic constraints (IPQC) problem and proved as NP-hard. To tackle the computational complexity of solving IPQC, a heuristic algorithm with polynomial time is proposed. Extensive simulation results show that it performs close to the optimal one and outperforms existing algorithms significantly.

E-commerce has become one of the common commercial activities in people's daily lives. The major advantage of e-commerce over conventional commercial activities is the information transparency while people can freely share their opinions and comments. Such information has profound influence on user behaviors in e-commerce activities. Meanwhile, social network service (SNS) has also become the most popular way to get and share information on the Internet. Therefore, it is quite natural to put e-commerce and SNS together. Recently, there emerge many online social e-commerce network (SECON) services, which not only allow users to conduct e-commerce transactions but also enable users to share information as in the other SNS like Twitter. Although conventional SNS has been widely investigated, little is known about SECON. To address this problem, we conduct a trace-driven analysis on a successful SECON called Jumei, with millions of users. Our analysis is based on shared information and all activities they created, all these data are crawled from the website of Jumei. We shed light on the user activity characteristics in SECON. By analyzing the crawled data, we discover that the social ties have an important influence on commercial activities. However, to our surprise, there are many differences between SECON and SNS: (a) the network topology structure is greatly different from SNS, (b) strong ties play a more crucial role than in SNS, same to viral marketing intuition, and (c) the behavior of adoption activity is influenced weakly by peers in the social network, for example, nearly 60% users influenced by only one information propagated from social links before they decided to buy it. Furthermore, we find that it takes a long time to adopt what their followees have bought.

DTNs are prone to congestion due to limited resource on each node and unpredictable end-to-end delay. We aim to develop an effective congestion control mechanism in this paper. For this purpose, we first identify a list of major congestion factors by analyzing the causes of congestion. We then model the congestion control as a multiple attribute decision making problem (MADM), in which the weight of congestion factors is measured by an entropy method. To solve this problem, we present a MADM-based congestion control mechanism that determines a set of forwarding messages and its transmission order on each encounter event. Moreover, we design a buffer management scheme that deletes messages whose removal would incur the least impact to the network performance when the buffer overflows. Extensive real-trace driven simulation is conducted and the experimental results finally validate the efficiency of our proposed congestion control mechanism.

After a decade of extensive research on application-specific wireless sensor networks (WSNs), the recent development of information and communication technologies make it practical to realize software-defined sensor networks (SDSNs), which are able to adapt to various application requirements and to fully explore the resources of WSNs. In SDSNs, wireless sensor nodes can be dynamically reprogrammed for different sensing tasks via the over-the-air-programming technique. For a given sensing task, it is usually required to guarantee certain quality-of-sensing, e.g., coverage ratio. Intuitively, the more sensors are deployed with a program, the higher quality-of-sensing of the corresponding task can be achieved. However, this is at the expense of high reprogramming energy consumption. In this paper, we investigate how to design an energy-efficient reprogramming strategy with guaranteed quality-of-sensing for a sensing task. To this end, two issues will be tackled: 1) the subset of sensors that shall be reprogrammed, i.e., reprogramming sensor selection and 2) the program distribution routing. They are jointly considered and formulated as an integer linear programming (ILP) problem, based on which an algorithm with low computation complexity is then proposed. The high efficiency of our algorithm is validated by extensive simulation studies.

Many services and applications in vehicular ad-hoc networks (VANETs) require preserving and secure data communications. To improve driving safety and comfort, the traffic-related status information will be broadcasted regularly and shared among drivers. Without the security and privacy guarantees, attackers could track their interested vehicles by collecting and analyzing their traffic messages. Hence, anonymous message authentication is an essential requirement of VANETs. On the other hand, when a vehicle is involved in a dispute event of warning message, the certificate authority should be able to recover the real identity of this vehicle. To deal with this issue, we propose a new privacy-preserving authentication protocol with authority traceability using elliptic curve based chameleon hashing. Compared with existing schemes, our approach possesses the following features: 1) mutual and anonymous authentication for both vehicle-to-vehicle and vehicle-to-roadside communications, 2) vehicle unlinkability, 3) authority tracking capability, and 4) high computational efficiency. We also demonstrate the merits of our proposed scheme through security analysis and extensive performance evaluation.

Social-based routing protocols have shown their promising capability to improve the message delivery efficiency in Delay Tolerant Networks (DTNs). The efficiency greatly relies on the quality of the aggregated social graph that is determined by the metrics used to measure the strength of social connections. In this paper, we propose an improved metrics that leads to high-quality social graph by taking both frequency and duration of contacts into consideration. Furthermore, to improve the performance of social-based message transmission, we systematically study the community evolution problem that has been little investigated in the literation. Distributed algorithms based on the obtained social graph are developed such that the overlapping communities and bridge nodes (i.e., connecting nodes between communities) can be dynamically detected in an evolutionary social network. Finally, we take all the results above into our social-based routing design. Extensive trace-driven simulation results show that our routing algorithm outperforms existing social-based forwarding strategies significantly.

Delay/Disruption Tolerant Network (DTN) differs from the conventional networks in that it has no continuous or contemporaneous connections among wireless nodes. Its inherent characteristic of intermittent connections makes existing routing solutions hardly to be applied directly. Epidemic routing using random linear network coding has been studied and proved as an efficient way for delivering small amount of data. To our best knowledge, we are the first to study high performance reliable transmission for bulk or stream-like data in DTNs. In this paper, we propose a dynamic segmented network coding scheme to efficiently exploit the transmission opportunity that is scarce in DTNs. In particular, we adopt a dynamic segment size control mechanism, which makes the segmentation adapt to the dynamics of the network. A lower bound of the expected delivery delay for bulk-data dissemination using segmented network coding is also derived. Both analytical and simulation results validate the high performance of our proposal. Several other interesting findings are also observed.

Network coding has shown the promise of significant throughput improvement. In this paper, we study the network throughput using network coding and explore how the maximum throughput can be achieved in a two-way relay wireless network. Unlike previous studies, we consider a more general network with arbitrary structure of overhearing status between receivers and transmitters. To efficiently utilize the coding opportunities, we invent the concept of network coding cliques (NCCs), upon which a formal analysis on the network throughput using network coding is elaborated. In particular, we derive the closed-form expression of the network throughput under certain traffic load in a slotted ALOHA network with basic medium access control. Furthermore, the maximum throughput as well as optimal medium access probability at each node is studied under various network settings. Our theoretical findings have been validated by simulation as well.

In Delay Tolerant Networks (DTNs), the connections between mobile nodes are intermittent and constant end-to-end paths rarely exist. In order to achieve high delivery ratio, most existing DTN routing algorithms favor the ``multi-hop forwarding'' fashion where a message can be forwarded by multiple relay nodes in the hope that one of the employed relay nodes can deliver the message to the destination node. Since aggressively employing relay nodes may incur intolerable delivery cost to resource-constrained mobile nodes in DTNs, it is significant to design a cost-efficient routing protocol that can achieves high delivery ratio. In this paper, we first utilize the small-world feature to limit the maximum forwarding hops in a reasonable way and then propose a greedy relay node selection strategy. A cost-efficient social-aware forwarding algorithm called TBSF (the-best-so-far) is then presented. Extensive simulations on real mobility traces are conducted to evaluate the performance of TBSF. Simulation results show that, in comparison with several well-known routing algorithms, TBSF can achieve high performance in terms of delivery ratio and delivery delay while with much lower delivery cost.

To understand the delivery performance of message dissemination in Disruption Tolerant Networks (DTNs), various methods have been proposed in the literature. However, existing work shares a common simplification that the pairwise meeting rate between any two mobile nodes is exponentially distributed. In this paper, instead of relying on such assumption, we jointly consider the transmission range and Random Direction Mobility (RDM) model to stochastically analyze delivery performance of epidemic routing in terms of percolation ratio and delivery delay. Furthermore, we study a controlled epidemic routing, in which any message stays at a mobile node longer than a predefined lifetime should be removed from the node. It can be considered as an age-structure process described by the Susceptible-Infectious-Recovered (SIR) model. To the best of our knowledge, we are the first to characterize the message propagation process by applying the Delay Differential Equations (DDEs) in DTNs. The correctness of our analysis is validated by extensive simulations.

After a decade of extensive research on application-specific wireless sensor networks (WSNs), the recent development of information and communication technologies make it practical to realize new WSNs paradigm known as software-defined sensor networks (SDSNs). SDSNs are able to adapt to various application requirements and to fully explore the communication, computation and sensing resources of WSNs. Sensor nodes in SDSNs can be dynamically reprogrammed for different sensing tasks via the over-the-air-programming technique. In this paper, we introduce the concept of SDSNs and outline several pioneering related work and enabling technologies for the realization of SDSNs.

A large number of data centers have been deployed and available for public renting with the rapid development of cloud computing recently. Meanwhile, the proliferation of automotive electronics has made rich resources in various forms of computation and communication. It is challenging but of great significance to make use of these resources in an efficient way. In this paper, we propose a two-tier data center architecture that leverages the excessive storage resources in parking lots. Such resources form an auxiliary vehicular data center (VDC) such that the pressure on the conventional data center can be mitigated and the total communication cost be reduced. After modeling the dynamics of available resources in a parking lot with a finite capacity, we propose three VDC management policies (i.e., non-replication, simple replication and network coding based replication) and derive their total communication cost in closed form. The high efficiency of the two-tier data center architecture and the accuracy of our analysis are validated via extensive simulations.

Recently, BitTorrent (BT) communities are rapidly evolving towards Private Torrent (PT) sites. The efficiency of such content sharing system depends on how well it can satisfy the requirements of users. The traditional way to improve usability is to make users extend their seeding duration by incentives like Share Ratio Enforcement (SRE). However, our measurement-based studies show its inefficiency in meeting various download requests from users. For example, some available contents have not been requested for long and some requests are not available, recognized as the “available but no request” and “request but not available” problems respectively. To make the content availability match users' download demands, we propose new incentive polices that only reward the seeding behaviors for needed torrents. This is achieved by providing users the rewarding information of seeding various contents that vary based on their interests and timeliness to the community. Our trace-driven simulation result shows the proposed “Max Cost Interval” and “Max Time Interval” policies can significantly improve the content over existing policies in a more cost-efficient way.

Network Coding (NC) has shown the promise of significant throughput improvement to wireless networks. Meanwhile, Multiple Packet Reception (MPR) has been proved as an efficient way to combat the Multiple Access Interference (MAI) problem and to make wireless networks truly scalable. Their outstanding advantages to wireless network performance have attracted considerable research interests. In this paper, we find that the greedy use of MPR may take side effect to NC and degrade the overall network throughput for All-to-All (A2A) communication in two-way relay networks. To fully explore the potential of both MPR and NC, the transmission should be carefully scheduled. The optimal scheduling to achieve the maximum A2A throughput is then formulated as an Integer Programming (IP) problem. A heuristic transmission scheduling algorithm with low computational complexity is also proposed. Simulation results show that our heuristic algorithm performs very closely to the optimal solution obtained by solving the IP problem, and achieves a significantly improved performance over the greedy scheduling that always exhausts the MPR capacity. Furthermore, a non-strict monotonic relationship between the MPR capacity and the maximum A2A throughput has been also discovered from our experimental results.

Delay/Disruption Tolerant Network (DTN) differs from the conventional networks in that it has no continuous or contemporaneous connections among wireless nodes. Its inherent characteristic of intermittent connections makes existing routing solutions hardly to be applied directly. Epidemic routing using random linear network coding has been studied and proved as an efficient way for delivering small amount of data. To our best knowledge, we are the first to study high performance reliable transmission for bulk or stream-like data in DTNs. In this paper, we propose a dynamic segmented network coding scheme to efficiently exploit the transmission opportunity that is scarce in DTNs. In particular, we adopt a dynamic segment size control mechanism, which makes the segmentation adapt to the dynamics of the network. The high performance of our proposal is validated by extensive simulations. Several other interesting findings are also observed from the simulation results.

Content distribution in vehicular networks is essential to many emerging applications. The issues such as content distribution from road side units (RSUs) to vehicles or the cooperation between vehicles have drawn a lot of interests in the literature. However, little work is on packets distribution from content providers to RSUs and many related issues are still under-investigated. In this paper, we consider the problem of minimizing the distribution cost, which is defined as the number of packets that shall be dispatched to RSUs, for deadline-constrained content distribution in vehicular networks. The problem is first formulated as an integer programming problem, based on a link-coloring concept. Then, a heuristic algorithm with low computational complexity is proposed. The high efficiency of the proposed algorithm is extensively validated by the fact that it performs close to the optimal solution obtained by the CPLEX solver.

Many services and applications in vehicular ad-hoc networks (VANETs) require preserving and secure data communications. To improve driving safety and comfort, the traffic-related status information will be broadcasted regularly and shared among drivers. Without the security and privacy guarantee, attackers could track their interested vehicles by collecting and analyzing their traffic messages. Hence, anonymous message authentication is an essential requirement of VANETs. On the other hand, when a vehicle is involved in a dispute event of warning message, the certificate authority should be able to recover the real identity of this vehicle. To deal with this issue, we propose a new privacy-preserving authentication protocol with authority traceability using elliptic curve based chameleon hashing. Compared with existing schemes, our approach possesses the following features: (1) mutual and anonymous authentication, (2) unlinkability, (3) authority tracking capability and (4) high efficiency. We also demonstrate the merits of our proposed scheme through extensive security analysis and performance evaluation.

Delay/Disruption Tolerant Network (DTN) differs from the conventional networks in that it has no continuous or contemporaneous connections among wireless nodes. Its inherent characteristic of intermittent connections makes existing routing solutions hardly to be applied directly. Epidemic routing using random linear network coding has been studied and proved as an efficient way for light data delivery. In this paper, we propose a pipelined segmented network coding approach using double buffer to provide best-effort services for bulk or stream-like data dissemination in DTNs. Simulations were conducted to verify its performance in terms of packet delivery ratio and throughput. Several other interesting findings are also observed.

Network coding has shown the promise of significant throughput improvement. In this paper, we study the throughput of two-hop wireless network coding and explore how the maximum throughput can be achieved under a random medium access scheme. Unlike previous studies, we consider a more practical network where the structure of overhearing status between the intended receivers and the transmitters is arbitrary. We make a formal analysis on the network throughput using network coding upon the concept of network coding cliques (NCCs). The analysis shows that the maximum normalized throughput, subject to fairness requirement, is n/n+m, where n is the number of transmitters and m is the number of NCCs in a 2-hop wireless network. We have also found that this maximum throughput can be achieved under a random medium access scheme when the medium access priority of the relay node is equal to the number of NCCs in the network. Our theoretical findings have been validated by simulation as well.

Delay/Disruptive Tolerant Network (DTN) differs from the traditional networks in that it has no continuous or contemporaneous connections but only intermittent connections among wireless nodes and thus it is viewed as an opportunistic networks. DTNs emerge as a good alternative to provide services to a variety of applications in highly challenged environments. However, the characteristics of DTNs make existing solutions infeasible to be applied directly and new solutions are required to be explored, such as multicast which has been extensively studied before in internet and mobile ad hoc networks. Network coding has been proved been proved as an efficient way to improve the performance of multicast in traditional networks. In this paper, we use simulations to study how multicast with network coding performs in DTNs in terms of delivery delay under various application requirements (i.e., the amount of content to distribute and the number of multicast group members) and the network settings (i.e., the popularity of the network and the contact rate). Some empirical results are provided in this paper as well.

Traversing Network Address Translation (NAT) for Peer-to-Peer (P2P) communication has become a hot topic recently. Compared to UDP, establishing TCP connections for hosts behind different NATs is more complex. Thus, many TCP-based applications do not address TCP traversal through NATs. Some solutions suggest using delegates to relay all communications, or tunneling TCP over UDP. However, they require a big reform to network architecture, or using a non-standard TCP/IP stack. In this paper, we present a novel idea called TCPBridge. TCPBridge converts TCP traversal to UDP traversal without modifying any binaries of the TCP-based applications. Our design can be integrated with those P2P applications which have not solved TCP traversal problem, and extends them to support direct communications between NAT hosts. It deals with the problem of TCP traversal, so as to improve the usability of applications. We have implemented TCPBridge in several existing P2P systems. Statistics prove that TCPBridge is scalable and robust, and we believe it will benefit many other existing P2P applications.

Recently, pervasive computing has become a very hot research field. Future Creation Lab., Olympus Corp. and the University of Aizu are working on this future technology to facilitate people¿s everyday life. The system we have proposed is named UMP-Percomp, a ubiquitous multiprocessor-based pipeline processing architecture to support high performance pervasive application development. Until now, we have implemented a prototype system to evaluate the performance of the architecture. However, the structure of the prototype system has some limitations: lack of scalability, inefficient pipeline algorithm, and do not do anything if there is lack of available PEs. So we modify the design to change these disadvantages. We add a UDP server to each component to support scalability and component substitution. We design a parallel algorithm to maximum the usage of PEs, and we also consider the situation of lacking suitable PEs. Finally, we do an experiment on the new system to compare with the old system.