Vision-language models are increasingly deployed as computer-use agents (CUAs) that operate desktops and browsers.Top-performing CUAs are framework-based systems that decompose planning and execution, while endto-end screenshot-to-action policies are easier to deploy but lag behind on benchmarks such as OSWorld-Verified.GUI datasets like OSWorld pose two bottlenecks: they expose only a few hundred interactive, verifiable tasks and environments, and expert trajectories must be gathered by interacting with these environments, making such data hard to scale.We therefore ask how reinforcement learning from verifiable rewards (RLVR) can best exploit a small pool of exist expert trajectories to train end-to-end policies.Naïvely mixing these offpolicy traces into on-policy RLVR is brittle: even after format conversion, expert trajectories exhibit structural mismatch and distribution shift from the learner.We propose BEPA (Bi-Level Expert-to-Policy Assimilation), which turns static expert traces into policy-aligned guidance via self-rolled reachable trajectories under the base policy (LEVEL-1) and a pertask, dynamically updated cache used in RLVR (LEVEL-2).On OSWorld-Verified, BEPA improves UITARS1.5-7Bsuccess from 22.87% to 32.13% and raises a held-out split from 5.74% to 10.30%, with consistent gains on MMBench-GUI and Online-Mind2Web.Our code and data are available at

In edge inference systems, efficient model deployment and accurate forecasting of online service requests are critical for maintaining optimal performance and resource utilization. This study introduces a novel approach that integrates predictive enhancement techniques to improve model deployment strategies and online service request forecasting. By addressing the existing gap between prediction and decision-making processes, our method enables more responsive and adaptive edge computing environments. Experimental results demonstrate significant improvements in both deployment efficiency and forecasting accuracy, highlighting the potential of our approach to advance the state-of-the-art in edge inference system management.

Video analytics pipelines migrating to edge deployments are facing performance bottlenecks under limited bandwidth. Non-uniform intra-frame encoding emerges to further compress pixels without affecting the output of the server deep neural network (DNN), while it is inefficient in high-resolution video streaming at low bandwidth. The detail enhancement capability of neural super-resolution (SR) permits resolution downsampling and aggressive compression on edge devices for low-latency transmission. To exploit its accuracy potential, DNN-oriented non-uniform encoding is expected to be additionally aware of SR models. However, traditional codecs struggle to cope with both quality optimization for SR and global semantic features for DNN. We advocate neural codecs for coordinated encoding and enhancement, enabling analytic-oriented video streaming with optimal accuracy-delay tradeoffs. Our system, VidIQ, achieves quality-enhanced real-time video analytics by 1) improving the network architecture of neural codecs (at two granularity) to integrate SR models into a DNN-oriented analytics pipeline, and 2) adapting the multi-scale encoder and SR-decoder to scene dynamics (i.e., content and bandwidth variations) with the help of the monolithic controller to hold a performance advantage. Extensive evaluations showcase that VidIQ reduces end-to-end delay by 35.8% and improves analytics accuracy by 21.2% compared to the recent video compression, enhancement, and streaming baselines.

Video analytics pipelines (VAPs) have been a paradigm for large-scale video analytics. Due to temporal redundancy in video, frame filtering is widely used in VAPs to reduce analysis workload. However, existing works overlook a limitation: while inference operates only on selected frames, decoders must still process many redundant frames due to codec dependencies, leading to over-decoding trap. This limitation stems from the reference-based design in modern codecs, which require decoding preceding frames to reconstruct any selected one. As a result, over-decoding has become the practical bottleneck in VAPs using modern decoders, highlighting a critical but under-explored problem. To address this issue, we propose ParaDeco, a high-throughput video analytics framework featuring a novel frame-level parallel generative decoder. Unlike traditional decoders, ParaDeco adopts a decode-what-matters approach with decoupled frame dependencies. To decode arbitrary frames independently, ParaDeco generates frame-wise features as standalone skeletons using compressed video metadata, then predicts pseudo frames maintaining semantic consistency with original frames. Moreover, ParaDeco identifies which frames truly matter for analysis via delicate contribution-based frame filtering. We implement ParaDeco on a cloud server and evaluate it on large-scale real-world video datasets. Our experimental results show that ParaDeco achieves a 2.76× speedup on average compared to state-of-the-art VAPs.

To enable edge AI providers to recruit edge devices and use them to deploy AI models and provision inference services, we conduct a comprehensive mathematical and algorithmic study on a novel incentive and optimization mechanism based on repeated auctions. We first model and formulate a time-cumulative social cost optimization problem to capture the challenges of the trade-off between cost and accuracy, the dependency between adjacent auctions, and the need of achieving desired economic properties. Then, to solve this intractable non-linear integer program in an online manner, we design a set of polynomial-time algorithms that work together. Our approach dynamically chooses and switches winning bids under careful control, incorporates online learning to overcome posterior inference accuracy and workload queue dynamics, and leverages randomization to strategically convert fractional decisions of model placement and query dispatch into integers. We also allocate payments to meet the necessary and sufficient conditions for the desired economic properties. Further, we rigorously prove the constant competitive ratio, the sub-linear regret and fit, and the truthfulness and individual rationality for our proposed approach. Finally, through extensive experiments using real devices, AI models, and data traces, we have validated the substantial advantages of our proposed approach compared to the baselines and the state-of-the-art methods.

Deep neural network (DNN)-based real-time video analytics service, as a core module for numerous crucial applications such as augmented reality (AR), has garnered increasing research attention, where mobile edge computing (MEC) is often leveraged to mitigate its real-time processing burden on resource-constrained user devices. For Quality of Experience (QoE) optimization, latest works employ reinforcement learning (RL)-based methods to adaptively adjust configurations (e.g., resolution and frame rate), yet still presenting significant challenges. Firstly, we observe a substantial diversity in QoE patterns among users. Given that existing methods integrate a fixed QoE pattern in parameter training, it is intuitive to customize a policy network for each user. However, this necessitates significant training investment, failing to support on-the-fly deployment for new users. Secondly, given the dual dynamics from both the network and video content in edge video analytics system, existing methods often fall into the dilemma of fitting newly emerged and diverse system states with offline-trained fixed parameters. While it is promising to employ online learning algorithms, most of them struggle to catch up with the high dynamics. We hence proposeMystique. In real-time edge video analytics domain, it is the first meta-RL-based user-level configuration adaptation framework. Mystique establishes an initial model in offline meta training with model-agnostic meta-learning (MAML), enabling swift online adaptation to new users and system states through limited gradient updates from initial parameters. Comprehensive experiments illustrate that Mystique can improve QoE by 42% on average compared to prior works.

Reinforcement learning (RL) has emerged as a promising approach across various applications, yet its reliance on repeated trial-and-error learning to develop effective policies from scratch poses significant challenges for deployment in scenarios where interaction is costly or constrained. In this work, we investigate the offline-to-online RL paradigm, wherein policies are initially pretrained using offline historical datasets and subsequently fine-tuned with a limited amount of online interaction. Previous research has suggested that efficient offline pretraining is crucial for achieving optimal final performance. However, it is challenging to incorporate appropriate conservatism to prevent the overestimation of out-of-distribution (OOD) data while maintaining adaptability for online fine-tuning. To address these issues, we propose an effective offline RL algorithm that integrates a guidance model to introduce suitable conservatism and ensure seamless adaptability to online fine-tuning. Our rigorous theoretical analysis and extensive experimental evaluations demonstrate better performance of our novel algorithm, underscoring the critical role played by the guidance model in enhancing its efficacy.

Increased videos captured by widely deployed cameras are being analyzed by computer vision-based Deep Neural Networks (DNNs) on servers rather than being streamed for humans. Unfortunately, the conventional codecs (e.g., H.26x and MPEG-x) originally designed for video streaming lack content-aware feature extraction and hinder machine-centric video analytics, making it difficult to achieve the required high accuracy with tolerable delay. Neural codecs (e.g., autoencoder) now hold impressive compression performance and have been widely advocated in video streaming. While autoencoder shows transformative potential, the application in video analytics is hampered by low accuracy in detecting small objects of high-resolution videos and the serious challenges posed by multi-video streaming. To this end, we propose AdaStreamer with adaptive neural codecs to enable real machine-centric high-accuracy multi-video analytics. We also investigate how to achieve optimal accuracy under delay constraints via careful scheduling in Compression Ratios (CRs, the ratio of the compressed size to the original data size) and bandwidth allocation, and further propose a Markov-based Adaptive Compression and Bandwidth Allocation algorithm (MACBA). We have practically developed a prototype of AdaStreamer, based on which extensive experiments verify its accuracy improvement (up to 15%) compared to state-of-the-art coding and streaming solutions.

Computing power infrastructure has a significant impact on the training speed and cost during the training process of Large Language Models (LLMs). Due to differences in GPU hardware, network architecture, and training frameworks, the cost and unit price of computing power fluctuate greatly, and even within the same cloud service provider, there are huge differences in prices for different clusters. The feature of regularly saving checkpoints during the training process of large language models makes it possible to dynamically migrate tasks between GPU clusters during the training process. Based on this background, this article proposes a set of greedy strategy heuristic algorithms with the goal of minimizing training costs, which choose the best time and target for task migration, thereby achieving overall optimization of training costs. Through multiple rounds of experimental verification on real task samples and simulation data, the results show that the proposed greedy strategy based preemptive algorithm can reduce the overall training cost by more than 10% . This study provides new solutions and methods for efficient computing power supply for large language model training, which can help reduce training costs for computing infrastructure operators and customers.

In recent years, many crowdsourcing platforms have emerged, using the resources of recruited workers to perform diverse outsourcing tasks, where the video analytics attracts much attention due to its practical implications. For maximum profits, platforms carefully choose the workers and determine the video analytics configurations to ensure accuracy; meanwhile, workers possess the flexibility to tailor the configurations for their indivi-dual gains, which makes it hard for platforms to optimize their profits considering the platform-worker conflicts. In this paper, we design an incentive mechanism for Multi-leader game-based video Analytics upon CROwdsourcing, named MACRO, to over-come the above situation. Under that mechanism, we first formu-late the utility optimization problems for platforms and workers, respectively. We then propose a dual ascent-based method to op-timally determine the video analytics configurations for a multi-platform game, ensuring Pareto efficiency. Moreover, in the context of a multi-leader game involving platform-worker conflicts, we design an incentive function with its incentive factor update strategy and propose an ADMM-based approach for maximizing incentives that motivate workers to contribute to the platforms' profits. Rigorous proofs demonstrate the linear convergence of the MACRO to the multi-leader Stackelberg equilibrium. Trace-driven experiments show that MACRO improves the Pareto efficiency by 26.3%, outperforming other approaches.

Recent years have witnessed the unprecedented performance of convolutional networks in image super-resolution (SR). SR involves upscaling a single low-resolution image to meet application-specific image quality demands, making it vital for mobile devices. However, the excessive computational and memory requirements of SR tasks pose a challenge in mapping SR networks on a single resource-constrained mobile device, especially for an ultra-high target resolution. This work presents TileSR, a novel framework for efficient image SR through tile-granular parallel offloading upon multiple collaborative mobile devices. In particular, for an incoming image, TileSR first uniformly divides it into multiple tiles and selects the top-K tiles with the highest upscaling difficulty (quantified by mPV). Then, we propose a tile scheduling algorithm based on multi-agent multiarmed bandit, which attains the accurate offload reward through the exploration phase, derives the tile packing decision based on the reward estimates, and exploits this decision to schedule the selected tiles. We have implemented TileSR fully based on COTS hardware, and the experimental results demonstrate that TileSR reduces the response latency by 17.77-82.2% while improving the image quality by 2.38-10.57% compared to other alternatives.

Increased videos captured by widely deployed cameras are being analyzed by computer vision-based Deep Neural Networks (DNNs) on servers rather than being streamed for humans. Unfortunately, the conventional codecs (e.g., H.26x and MPEG-x) originally designed for video streaming lack content-aware feature extraction and hinder machine-centric video analytics, making it difficult to achieve the required high accuracy with tolerable delay. Neural codecs (e.g., autoencoder) now hold impressive compression performance and have been widely advocated in video streaming. While autoencoder shows transformative potential, the application in video analytics is hampered by low accuracy in detecting small objects of highresolution videos and the serious challenges posed by multivideo streaming. To this end, we propose AdaStreamer with adaptive neural codecs to enable real machine-centric highaccuracy multi-video analytics. We also investigate how to achieve optimal accuracy under delay constraints via careful scheduling in Compression Ratios (CRs, the ratio of the compressed size to the original data size) and bandwidth allocation, and further propose a Markov-based Adaptive Compression and Bandwidth Allocation algorithm (MACBA). We have practically developed a prototype of AdaStreamer, based on which extensive experiments verify its accuracy improvement (up to 15%) compared to stateof-the-art coding and streaming solutions.

Continuously orchestrating in-network model aggregations for federated learning faces fundamental challenges such as the combinatorial nature of traffic reduction, the dynamic trade-offs between system overhead and model convergence, and the unpredictable inputs from uncertain system environments. In this work, we model a nonlinear mixed-integer program to optimize the long-term total cost of federated learning computation overhead, traffic reduction, network delay, and programmable switch reconfigurations over time. To attack the lexicographic minimax, submodular, and online nature of this problem, we propose a polynomial-time algorithmic framework to judiciously designate the timing of reconfigurations, while designing and invoking a linearized transformation for selecting routing paths, a greedy sub-algorithm for selecting aggregation locations, and an online learning sub-algorithm for controlling federated learning convergence. We demonstrate our rigorous mathematical insights behind our algorithms, and prove the competitive ratio as the performance guarantee. Using trace-driven evaluations, we have validated our approach's superiority over existing methods.

Over the last decade, the mobile crowdsourcing has become a paradigm to conduct the manual annotation and further analytics by recruited workers, with their rewards depending on the result quality. Existing dispatchers cannot precisely capture the resource-quality trade-off for video analytics, because the configurations supported by recruited workers are limited, and workers’ availability changes over time. To determine the most suitable configurations as well as workers for video analytics, we formulate a non-linear mixed program in long term, maximizing the crowdsourcing profit. Based on previous results under various configurations and workers, we design an algorithm via a series of subproblems to decide the configurations adaptively upon the prediction of workers’ feedbacks. Such prediction is based on volatile multi-armed bandit to capture workers’ availability and stochastic changes on resource uses. Furthermore, we extend the proposed algorithms to the multi-worker selection scenario where the platform needs to determine a candidate worker set instead of a single worker for video analytics. Via rigorous proof, the regret is ensured upon the Lyapunov optimization and the bandit, measuring the gap between the online decisions and the offline optimum. Extensive trace-driven experiments show that our proposed algorithm improves the profit by 37% compared with other algorithms.

The detection of objects via neural networks plays a key role in various video analytics, but consumes huge resources. Due to the limited computing capability at edges, such real-time detections should be precisely used for the objects that need the most attention. Unfortunately, as the target objects keep moving, existing systems fail to consider both the distribution of objects and object movements over regions, and existing tracking mechanisms are easily affected by background content. Therefore, we propose spatial and temporal detection with attention for analytics, to increase the quality of detections for those targets. However, the attention shift over regions, the uncertainty of detections, and the constrained edge resources essentially hamper us from efficient analytics. We propose an adaptive partition planner to divide the frame into regions to achieve spatial attention. Afterwards, we design a detection planner to orchestrate the detection model temporally for each region by an online mechanism, via a queue-based adaptation. The spatial and temporal attention are integrated to maximize the accumulative detection accuracy. Via rigorous proof, both dynamic regret regarding detection accuracy and the real-time requirement for the video analytics are ensured. The testbed experiments confirm the superiority of our approach over multiple state-of-the-art algorithms.

The usage of live streaming services has led to a substantial increase in live video traffic. However, the perceived quality of experience of users is frequently limited by variations in the upstream bandwidth of streamers. To address this issue, several adaptive bitrate (ABR) algorithms have been developed to mitigate bandwidth variations. Nevertheless, the ability of users to enjoy high-quality live streams remains limited. While neural-enhanced approaches, such as super-resolution, offer significant quality improvements, frame-oriented super-resolution leads to excessive inference delay that violates the real-time feature of live streaming. In response, we propose ViChaser, which examines block-oriented super-resolution for live streaming. ViChaser performs neural super-resolution on potential blocks of interest in the media server, corresponding to the user’s viewpoint, and uses online learning to adapt to the dynamic content of the video. Additionally, ViChaser utilizes the Lyapunov framework to efficiently allocate uplink bandwidth for original low-quality live video and high-quality labels. The experimental results demonstrate that ViChaser achieves 1.2–1.5 dB higher video quality in Peak-Signal-to-Noise-Ratio than WebRTC and increases processing speed by 11–16 fps relative to LiveNAS.

Offline reinforcement learning (RL) aims to learn effective policies from recorded data without further interactions in the environments that are often costly or risky. Model-based algorithms, which begin by constructing an environmental model and then learn the policy under the model, have become a promising approach. However, most existing works have been over-conservative to avoid the out-of-distribution error induced by the model generated samples, leading to poor performance instead. In this work, we propose a novel model-based offline RL method, named Simple Double Validation (SDV). The main idea of SDV is to introduce an additional guidance model to assist the agent in determining the rationality of the states, combined with an advantage weighting factor to avoid effects that could potentially mislead the models due to suboptimal samples. In this way, the agent can be guided to more favourable states with reliable decisions. We evaluated SDV on the widely studied offline RL benchmarks and demonstrated its state-of-the-art performance. At the same time, our work introduces the idea of double validation and model advantage weighting into the field of model-based offline RL, providing new insights for future research.

Network visualization is essential for network operators to diagnose ongoing network failures and understand the quality of the network. In-Band Network Telemetry (INT) supports network visualization by inserting P4 switch state information (e.g., queue length, hop latency, and link utilization) into the specific INT packets. In order to achieve network-wide coverage, paths of INT packets need to be delicately designed to ensure non-overlapping, high performance, and low overhead. However, existing INT path planning solutions ignore capturing the dynamic network status and fail to obtain the optimum. In this paper, we model the INT path planning based on the directed Edge Cover problem upon the dynamic network status, with the objective of minimizing the INT path latency. Although it is actually a non-linear integer program problem, by adopting delicate transformations, we design approximation algorithms with performance guarantees. We implement our system prototype INTView based on INTCollector upon real devices, i.e., Barefoot Wedge100BF and Inspur Rack. Extensive evaluation upon realistic settings shows that our proposed algorithm achieves 2x performance improvement regarding the completion time, compared with state-of-the-art schemas.

The widespread use of 5G and artificial intelligence applications has led to strong momentum in Mobile Edge Computing (MEC). With MEC, we can offload compute-intensive tasks to edge servers that are closer to the user, thereby reducing the long latency incurred by data transmission via WAN. Although many works have investigated task offloading decisions under service caching, the state of services is an equal, if not more important, research area of MEC, yet receive much less attention. In general, the arrival of tasks exhibit a distribution over time. Besides the necessary energy consumption in processing tasks offloaded to edge servers, a large amount of energy is required for maintaining services cached on servers. When more and more services become idle, they will incur a non-negligible additional energy. In this paper, we focus on an interesting but currently less studied problem in MEC, namely online service caching and state management in MEC. We propose DCSO, a bounded online algorithm that considers dynamic service caching and state management of services to minimize long-term cost in MEC systems. Meanwhile, our algorithm achieves a 2 competitive ratio in state management. Trace-driven simulations show that our algorithm reduces the overall cost efficiently while keeping low computation latency.

Delayed Gradient Averaging (DGA) has gained massive attention for improving the training efficiency of Federated Learning (FL) at edge networks, by allowing local computation in parallel to communication. However, it faces multiple challenges due to data distribution across heterogeneous edge devices and dynamic network environments. To address these challenges, we present A-DGA, a novel communication learning parallel federated learning algorithm designed for fluctuating, heterogeneous, and high-latency network conditions. Our proposed A-DGA dynamically sets extra local updates based on network status, avoiding inefficient training upon the outdated gradients. Theoretical analysis demonstrates that A-DGA outperforms DGA in terms of convergence rates given the fixed time slot length. We further conduct massive evaluations under various conditions, including different datasets, models, and data distribution. The experimental results show that A-DGA performs the best, compared to the state-of-the-art methods, achieving an acceleration factor of approximately x2 ∼ x4 and x1.2 ∼ x2, compared to FedAvg and DGA, respectively. Besides, our A-DGA also reduces energy consumption by about 40% ∼ 50% compared to DGA.

Performing machine learning inference at the network edge, named Edge Inference, showing benefits like low latency, reduced data traffic, and improved user privacy, has attracted massive attention. Computing Power Network (CPN) creates opportunities for edge inference, but also poses multiple challenges for service providers, including computing power selections from different enterprises, the time-coupled decision for resource provisioning and the unpredictable CPN network status. To overcome these challenges, this study formulates a time-varying integer program problem whose goal is to minimize long-term costs, involving switching costs, operational costs, communication costs, and queuing costs. Then we design a group of polynomial-time online algorithms to make online decisions, by taking into account stochastic inputs. Our algorithms adaptively make control decisions by solving delicately constructed subproblems based on the inputs predicted via online learning. Specifically, we first obtain fractional solutions which are transformed into integers for deployment with expectations preserving. Furthermore, we conduct a rigorous proof and establish the competitive ratio which highlights the difference between the performance of our proposed algorithms and the offline optimum. Our comprehensive evaluations, using datasets from real systems, demonstrate that our algorithms outperform multiple alternatives, up to an average of 35% cost reduction, confirming the effectiveness of our algorithms.

The inference workload redistribution is a technique for evacuating inference requests from hot edges to idle edges in edge collaborative systems, thereby achieving inference workload balancing for inference on different edges. However, with the continuous development of edge accelerators, the resource utilization of edge accelerators in executing inference requests in series is often low, and when executing multiple inference requests in parallel, it faces uncertain execution delays, different response-time Service Level Objectives (SLOs), and the generality of inference workloads in heterogeneous edge collaborative systems. To address these issues, for the first time in the domain of inference workload redistribution, we propose a Batch-aware Inference workload Redistribution and Parallel execution scheme, called BIRP, to reduce the additional latency caused by waiting for a single inference task during serial execution, thereby improving the overall inference accuracy. BIRP uses the Multi-Armed Bandit (MAB) algorithm to adjust hyperparameters of the Throughput Improvement Ratio (TIR) function online for improving the overall inference accuracy. For nonlinear terms in the problem, BIRP uses a piecewise linear approximation to convert it into a Quadratic Programming (QP) problem, ensuring the effectiveness of BIRP in theory. We prototype BIRP on an edge collaborative system composed of three heterogeneous edges. Based on real inference workload trace, we validate the superiority of our algorithm compared to the state-of-the-art model selection-based inference workload redistribution algorithm, with an overall inference loss reduction of at least 32.9% and the failure rate of SLO has been reduced to 19.8% of alternatives.

In-band Network Telemetry (INT) is proposed to detect networks via injecting specific probes to collect the hop-by-hop metadata within programmable switches. But there exist multiple challenges to conducting INT at Computing Power Network, such as control decisions of different INT frequencies, and the unforeseeable INT query workloads. In this study, we formulate an online non-linear time-varying integer programming problem that aims to maximize the overall quality of service through both frequency selection and INT query workload distribution. To achieve this, we propose an online learning, INTService, which utilizes a primal-dual mechanism to make fractional decisions. At last, extensive evaluations show that our proposed INTService exhibits up-lift performance 40% on average over other state-of-the-art algorithms.

Provisioning machine learning inference near the users at the network’s edge is emerging as a promising area for Edge AI. Due to the excessive energy consumption of edge devices, their owners often lack the motivation to actively contribute to their edge resources. To tackle this issue, we propose an incentive mechanism based on auctions, which enables edge device owners to submit their bids and compensates such bids via reward payments, ensuring the minimization of inference accuracy loss. We formulate a nonlinear mixed-integer program problem with the objective of minimizing the social cost, including accuracy loss cost, edge device cost, and service provider cost in the edge inference system. Then an Online Learning algorithm is devised to find the solutions, based on the primal-dual. To calculate the remuneration, we design a payment allocation algorithm based on the bid-winning probabilities. Our rigorous theoretical analysis shows that our algorithm designed achieves sub-linear growth on dynamic regret and dynamic fit over time while preserving the economic properties of truthfulness and individual rationality. Finally, multiple experiments validate the efficacy of the proposed auction mechanism algorithm from various perspectives compared with three other existing algorithms.

Decentralized federated learning across edge networks can leverage blockchain with consensus mechanisms for training information exchange among participants over costly and distrustful wide-area networks. However, it is non-trivial to optimally operate the blockchain to support decentralized federated learning due to the complex cost structure of blockchain operations, the balance between blockchain overhead and model convergence, and the dynamics and uncertainties of edge network environments. To overcome these challenges, we formulate a non-linear time-varying integer program that jointly places blockchain nodes and determines the number of training iterations to minimize the long-term blockchain computation and communication cost. We then design an online polynomial-time approximation algorithm that decomposes the problem and solves the subproblems alternately on the fly using only estimated inputs. We rigorously prove the sublinear regret of our approach. We further implement our approach with a prototype system, and conduct extensive trace-driven experiments to validate the superiority of our approach over other alternatives.

Conducting federated learning across distributed sites with In-Band Network Telemetry (INT) based data collection faces critical challenges, including control decisions of different frequencies, convergence of the models being trained, and resource provisioning coupled over time. To study this problem, we formulate a non-linear mixed-integer program to optimize the long-term INT overhead, resource cost, and federated learning cost. We then design polynomial-time online algorithms to solve this problem with only observable inputs on the fly, featuring laziness-aware resource adaption, online-learning-based INT flow selection and model aggregation control, as well as expectation-preserving randomized dependent rounding. We rigorously prove the parameterized-constant competitive ratio of our approach against the offline optimum, and the time-averaged constraint violation that vanishes in the long run. With extensive trace-driven evaluations, we confirm the superiority of our approach over other alternative approaches for reducing total cost and the efficacy of our trained models for solving real machine learning problems, reducing the real-time cost by 34% on average.

Loosely coupled and highly cohesived microservices running in containers are becoming the new paradigm for application development. Compared with monolithic applications, applications built on microservices architecture can be deployed and scaled independently, which promises to simplify software development and operation. However, the dramatic increase in the scale of microservices and east-west network traffic in the data center have made the cluster management more complex. Not only does the scale of microservices cause a great deal of pressure on cluster management, but also cascading QoS violations present a substantial risk for SLOs (Service Level Objectives). In this paper, we propose a Microservice-Oriented Topology-Aware Scheduling Framework (MOTAS), which effectively utilizes the topologies of microservices and clusters to optimize the network overhead of microservice applications through a heuristic graph mapping algorithm. The proposed framework can also guarantee the cluster resource utilization. To deal with the dynamic environment of microservice, we propose a mechanism based on distributed trace analysis to detect and handle QoS violations in microservice applications. Through real-world experiments, the framework has been proved to be effective in ensuring cluster resource utilization, reducing application end-to-end latency, improving throughput, and handling QoS violations.

In past years, artificial intelligence related services and applications have boomed, which require high computation, high bandwidth and low latency. Edge computing is regarded as an appropriate solution for them, especially video analytics. In this paper, we study the multi-server multi-user heterogeneous video analytics offloading problem, where users select appropriate edge servers and then offload their raw video data to the servers for essential analytics. To deal with the cooperation and conflicts among users and get a stable situation where each user has no incentive to change the offloading decision unilaterally, we formulate the video analytics offloading problem as a multiplayer game. Based on the goal of minimizing the overall delay, we design the potential optimal server selection strategy and then propose a game theory-based algorithm, through which the Nash equilibrium can be reached. Furthermore, we analyze its near-optimal performance via rigorous proof. Finally, extensive trace-driven experiments show that our method improves the overall delay by 48% on average, compared with other algorithms.

Crowdsourcing enables a paradigm to conduct the manual annotation and the analytics by those recruited workers, with their rewards relevant to the quality of the results. Existing dispatchers fail to capture the resource-quality trade-off for video analytics, since the configurations supported by various workers are different, and the workers’ availability is essentially dynamic. To determine the most suitable configurations as well as workers for video analytics, we formulate a non-linear mixed program in a long-term scope, maximizing the profit for the crowdsourcing platform. Based on previous results under various configurations and workers, we design an algorithm via a series of subproblems to decide the configurations adaptively upon the prediction of the worker rewards. Such prediction is based on volatile multi-armed bandit to capture the workers’ availability and stochastic changes on resource uses. Via rigorous proof, the regret is ensured upon the Lyapunov optimization and the bandit, measuring the gap between the online decisions and the offline optimum. Extensive trace-driven experiments show that our algorithm improves the platform profit by 37%, compared with other algorithms.

With the rapid development of Machine Learning (ML), Machine-Learning-as-a-Service (MLaaS) clusters appear in large numbers to support cloud platforms services, which adopt virtual machine (VM) to improve the availability, resilience and security. However, low energy efficiency is a major problem in such clusters. Previous work focused on reducing the number of physical machines by centralizing resources migration. Nevertheless, for ML tasks with frequent memory switching, blind migration is not worth the cost because the remaining time is less than the migration time, since the migration time can not be ignore due to the memory intensive of ML tasks. Therefore, this paper explores how the remaining time and memory replacement states in ML tasks, which we summarize as residue, affect migration, and proposes an online residue-aware migration algorithm based on Lyapunov optimization. Through rigorous proof, the gap between the algorithm and the optimal solution is ensured. Extensive simulations show that the proposed algorithm is better than the previous migration.

Edge computing is a new computing paradigm rising gradually in recent years. Applications, such as object detection, virtual reality and intelligent cameras, often leverage Deep Neural Networks (DNN) inference technology. The traditional paradigm of DNN inference based on cloud suffers from high delay because of the limited bandwidth. From the perspective of service providers, caching DNN models on the edge brings several benefits, such as efficiency, privacy, security, etc.. The problem we concerned in this paper is how to decide the cached models and how to allocate processors of edge servers to reduce the overall system cost. To solve it, we model and study the DNN Model Caching and Processor Allocation (DMCPA) problem, which considers user-perceived delay and energy consumption with limited edge resources. We model it as an integer nonlinear programming (INLP) problem, and prove its NP-Completeness. Since it is considered as a long-term average optimization problem, we leverage the Lyapunov framework to develop a novel online algorithm DMCPA-GS-Online with Gibbs Sampling. We give the theoretical analysis to prove that our algorithm is near-optimal. In experiments, we study the performance of our algorithm and compare it with other baselines. The simulation results with the trace dataset from real world demonstrate the effectiveness and adaptiveness of our algorithm.

User-perceived quality of experience (QoE) is critical as mobile video streaming experiences a substantial growth. User's demands are becoming diversified where accelerated play-back is the preference of a considerable part of users. However, the limited and fluctuate mobile bandwidth is often not capable of satisfying user's demand of watching video at 2x or higher speed because of consequential frequent rebuffering. Previous adaptive bitrate (ABR) algorithms hardly consider the variety of user playback rates. In this work, we fully exploit the relation between user-perceived, i.e., subjective video quality and the characteristic of video content. The result of our motivational experiments shows that viewers are less sensitive to the bitrate variation and playback rate alternation if there is higher degree of motion in the video. With above guidelines, we adaptively adjust the quality configuration and playback rate to significantly reduce the rebuffering while achieving similar or even higher subjective quality. Then we formulate subjective quality and playback rate adaption as a QoE maximization problem and propose the content based subjective quality and playback rate adaptation algorithm (CSP) utilizing Lyapunov optimization technique. Via rigorous proof, the time-average QoE achieved by CSP is in$O(1/V)$gap compared to optimal value, where$V$is the control parameter. Extensive evaluations confirm the superiority of our proposed algorithm over other state-of-the-art algorithms under both normal and accelerated playback rate.

The growing demands for cloud computing have led to high power consumption in data centers. To save power, existing works attempt to reduce the number of active servers via scheduling or migrating VMs. Meanwhile, owing to the lightweight, highly-portable and scalable properties, containers have been widely used in data centers. However, comparing with VMs, containers show the migratability property, i.e., some containers can not be consolidated by migration. Thus, one has to take this property into consideration when making scheduling and migrating decisions. To this end, this paper studies the energy-efficient container orchestration problem in data centers. We frame container scheduling and migration as a single problem, and take the migratability of containers into account. We propose an online container orchestration algorithm, based on Lyapunov optimization and Markov approximation. It works without requiring future information and achieves a provable performance guarantee. Simulation results show that our algorithm can effectively reduce the power consumption of data centers.

Federated learning is emerging as a major learning paradigm, which enables multiple devices to train a model col-laboratively and to keep the privacy of data. However, substantial computation-intensive iterations are performed on devices before the training completion, which incurs heavy consumption of the energy. Along with the stabilization of those model parameters being trained, such on-device training iterations are redundant gradually over time. Thus, we propose to scale the update results obtained from reduced iterations as the substitute for on-device training, based on current model status and device heterogeneity. We thus formulate a time-varying integer program, to minimize cumulative energy consumption over devices, subject to a long-term constraint regarding the model convergence. We then design a polynomial-time online algorithm upon system dynamics, which essentially balances the energy consumption and the model quality being trained. Via rigorous proofs, our approach only incurs sub linear regret, compared with its optimum, and ensures related model convergence. Extensive testbed experiments for real training confirm the superiority of our approach, over multiple alternatives, under various scenarios, decreasing at least 30.2% energy consumption, while preserving the accuracy of the model.

The detection of objects via neural networks plays a key role in various video analytics, but consumes huge resources. Due to the limited on-device computing capability, such real-time detections should be precisely used for the objects that need the most attention. Unfortunately, as the target objects keep moving, existing systems fail to conduct adaptive detections over multiple regions in a video, and existing tracking mechanisms are easily affected by background contents. Therefore, we propose to design attention-aware on-device detection for analytics, to increase the quality of detections for those targets. However, the uncertainty of detections, the attention shift over regions, and the provisioning of on-device resources essentially hamper us from efficient analytics. We formulate such a scenario as a non-linear integer program in long-term scope, to maximize the detection accuracy. Afterwards, we design an online mechanism to orchestrate the detection model for each region in the video to cope with the moves of the targets, via a queue-based adaptation and the randomized rounding. Via rigorous proof, both dynamic regret regarding detection accuracy and the real-time requirement for the video analytics are ensured. The testbed experiments confirm the superiority of our approach over multiple state-of-the-art algorithms.

Major cities worldwide have millions of cameras deployed for surveillance, business intelligence, traffic control, crime prevention, etc. Real-time analytics on video data demands intensive computation resources and high energy consumption. Traditional cloud-based video analytics relies on large centralized clusters to ingest video streams. With edge computing, we can offload compute-intensive analysis tasks to nearby servers, thus mitigating long latency incurred by data transmission via wide area networks. When offloading video frames from the front-end device to an edge server, the application configuration (i.e., frame sampling rate and frame resolution) will impact several metrics, such as energy consumption, analytics accuracy and user-perceived latency. In this paper, we study the configuration selection and bandwidth allocation for multiple video streams, which are connected to the same edge node sharing an upload link. We propose an efficient online algorithm, called JCAB, which jointly optimizes configuration adaption and bandwidth allocation to address a number of key challenges in edge-based video analytics systems, including edge capacity limitation, unknown network variation, intrusive dynamics of video contents. Our algorithm is developed based on Lyapunov optimization and Markov approximation, works online without requiring future information, and achieves a provable performance bound. We also extend the proposed algorithms to the multi-edge scenario in which each user or video stream has an additional choice about which edge server to connect. Extensive evaluation results show that the proposed solutions can effectively balance the analytics accuracy and energy consumption while keeping low system latency in a variety of settings.

In the multi-access edge computing environment, app vendors deploy their services and applications at the network edges, and edge users offload their computation tasks to edge servers. We study the user-perceived delay-aware service placement and user-allocation problem in edge environment. We model the MEC-enabled network, where the user-perceived delay consists of computing delay and transmission delay. The total cost in the offloading system is defined as the sum of service placement, edge server usage and energy consumption cost, and we need to minimize the total cost by determining the overall service-placing decision and user-allocation decision, while guaranteeing that the user-perceived delay requirement of each user is fulfilled. Our considered problem is formulated as a Mixed Integer Linear Programming problem, and we prove its NP-hardness. Due to the intractability of the considered problem, we propose a LOCal-search based algorithm for USer-perceived delay-aware service placement and user-allocation in edge environment, named LOCUS, which starts with a feasible solution and then repeatedly reduces the total cost by performing local-search steps. After that, we analyze the time complexity of LOCUS and prove that it achieves provable guaranteed performance. Finally, we compare LOCUS with other existing methods and show its good performance through experiments.

Efficiently analyzing geo-distributed datasets is emerging as a major demand in a cloud-edge system. Since the datasets are often generated in closer proximity to end users, traditional works mainly focus on offloading proper tasks from those hotspot edges to the datacenter to decrease the overall completion time of submitted jobs in a one-shot manner. However, optimizing the completion time of current job alone is insufficient in a long-term scope since some datasets would be used multiple times. Instead, optimizing the data distribution is much more efficient and could directly benefit forthcoming jobs, although it may postpone the execution of current one. Unfortunately, due to the throwaway feature of data fetcher, existing data analytics systems fail to re-distribute corresponding data out of hotspot edges after the execution of data analytics. In order to minimize the overall completion time for a sequence of jobs as well as to guarantee the performance of current one, we propose to re-distribute the data along with task offloading, and formulate corresponding ε-bounded data-driven task scheduling problem over wide area network under the consideration of edge heterogeneity. We design an online schemarunData, which offloads proper tasks and related data via piggybacking to the datacenter based on delicately calculated probabilities. Through rigorous theoretical analysis,runData is proved concentrated on its optimum with high probability. We implementrunData based on Spark and HDFS. Both testbed results and trace-driven simulations show that run Data re-distributes proper data via piggybacking and achieves up to 37 percent reduction on average response time compared with state-of-the-art schemas.

Mobile Edge Computing (MEC) has become an attractive solution to enhance the computing and storage capacity of mobile devices by leveraging available resources on edge nodes. In MEC, the arrivals of tasks are highly dynamic and are hard to predict precisely. It is of great importance yet very challenging to assign the tasks to edge nodes with guaranteed system performance. In this article, we aim to optimize the revenue earned by each edge node by optimally offloading tasks to the edge nodes. We formulate the revenue-driven online task offloading (ROTO) problem, which is proved to be NP-hard. We first relax ROTO to a linear fractional programming problem, for which we propose the Level Balanced Allocation (LBA) algorithm. We then show the performance guarantee of LBA through rigorous theoretical analysis, and present the LB-Rounding algorithm for ROTO using the primal-dual technique. The algorithm achieves an approximation ratio of$2(1+\xi)\ln (d+1)$with a considerable probability, where$d$is the maximum number of process slots of an edge node and$\xi$is a small constant. The performance of the proposed algorithm is validated through both trace-driven simulations and testbed experiments. Results show that our proposed scheme is more efficient compared to baseline algorithms.

The concept and application of smart city brings us enormous computing resources, but it also brings considerable computing heterogeneity. In this paper, we propose a brand new scenario of batch task offloading for vehicular cloud which maximizes the performance of offloading strategy in case of satisfying the deadline restraints and reliability requirements. To capture the sources of the service delay better, we split the task offloading process into several components, and focus on the optimization of routing and task execution. We model our problem (Batch Task Offloading for Vehicle Cloud problem with reliability restraints) in a submodular function maximizing perspective and show its NP-hardness. We present a novel greedy algorithm for offline scenario and analyze its theoretical performance. We prove our algorithm is$\left( {1 - {1 \over e}} \right)$-approximate in polynomial time. Furthermore, we study the problem in the online scenario using bandit submodular set maximizing model. We present an algorithm with regret bound$O(\text{log}^{2}\ T)$running in a small amount of arms which can reduce the computing complexity and the number of selected arms under the same performance. We also evaluate our algorithm in a microscopic and continuous traffic simulation platform, SUMO. The result shows our algorithms (both online and offline) outperforms other existing algorithms over 20%.

With the rapid development of edge computing, edge clusters need to deal with a tremendous amount of tasks, making some edge clusters overloaded, which further translates into task completion lag. Previous works usually copy the tasks from overloaded edges to idle edges so as to reduce the task queuing and computing delay. However, the completion delay of tasks copied to different edges cannot be predicted before the replication decision is made, which affects the overall task replication performance. In this paper, we propose an online task replication algorithm based on the predictions derived from multi-armed bandit. Via rigorous proof, the regret is ensured to be sub-linear upon the bandit, measuring the gap between the online decisions and the offline optimum. Extensive simulations are conducted to confirm the superiority of the proposed algorithm over state-of-the-art replication strategies.

Operating federated learning optimally over distributed cloud-edge networks is a non-trivial task, which requires to manage data transference from user devices to edges, resource provisioning at edges, and federated learning between edges and the cloud. We formulate a non-linear mixed integer program, minimizing the long-term cumulative cost of such a federated learning system while guaranteeing the desired convergence of the machine learning models being trained. We then design a set of novel polynomial-time online algorithms to make adaptive decisions by solving continuous solutions and converting them to integers to control the system on the fly, based only on the predicted inputs about the dynamic and uncertain cloud-edge environments via online learning. We rigorously prove the competitive ratio, capturing the multiplicative gap between our approach using predicted inputs and the offline optimum using actual inputs. Extensive evaluations with real-world training datasets and system parameters confirm the empirical superiority of our approach over multiple state-of-the-art algorithms.

Real-time on-device object detection for video analytics fails to meet the accuracy requirement due to limited resources of mobile devices while offloading object detection inference to edges is time-consuming due to the transference of video data over edge networks. Based on the system with both on-device object tracking and edge-assisted analysis, we formulate a non-linear time-coupled program over time, maximizing the overall accuracy of object detection by deciding the frequency of edge-assisted inference, under the consideration of both dynamic edge networks and the constrained detection latency. We then design a learning-based online algorithm to adjust the threshold for triggering edge-assisted inference on the fly in terms of the object tracking results, which essentially controls the deviation of on-device tracking between two consecutive frames in the video, by only taking previously observable inputs. We rigorously prove that our approach only incurs sub-linear dynamic regret for the optimality objective. At last, we implement our proposed online schema, and extensive testbed results with real-world traces confirm the empirical superiority over alternative algorithms, in terms of up to 36% improvement on detection accuracy with ensured detection latency.

Since the real-time video analytics with high accuracy requirement is resource-consuming, the profiles regarding such resource-accuracy trade-off are needed before the analytics for better resource allocation at resource-constrained edges. With the inner changes of the video contents, outdated profiles fail to capture the trade-off dynamically over time, which requires the profiles to be updated periodically and incurs an overwhelming resource overhead. Thus, we present Soudain, which dynamically adjusts the configurations in profiles and corresponding profiling intervals to capture the inner changes of multiple video streams at edges. Upon the fine-grained decisions for profiles, we propose an integer program to maximize the accuracy of video analytics in a long-term scope with resource constraint, and then design an algorithm to adjust the profiles in an online manner. We implement Soudain upon the server with GPU. Our testbed evaluations confirm that, by using the live video streams derived from real-world traffic cameras, Soudain ensures the real-time requirement and achieves up to 25% improvement on the detection accuracy, compared with multiple state-of-the-art alternatives.

Mobile cloud gaming (MCG), which is potential to deliver high-quality gaming experience to users anywhere and anytime, suffers from tremendous wide-area traffic and long network delays. Mobile edge computing (MEC), where cloud computing capabilities are pushed to the network edge, can help by providing gaming services in the proximity to users. However, since the quality of experience (QoE), i.e., the gaming experience, is easily impaired by long network delays and low frame rates, the performance of MEC-enabled MCG highly depends on the placement of gaming services and the allocation of related bandwidth. Furthermore, due to the erratic mobility of users, migrating services to follow such mobility decreases the impairment but incurs extra system cost, leading to the performance-cost tradeoff. To address these challenges, in this paper, we jointly investigate service placement and bandwidth allocation for MEC-enabled MCG. Considering the system dynamics, we propose to minimize the QoE impairment in a long time scope under a cost constraint for long-term migrations. To solve the problem, we develop an online two-layer iterative algorithm OnTrial. Rigorous theoretical analyses demonstrate that OnTrial achieves a near-optimal performance and bounds the potential violation of the migration cost constraint. Simulation results show that OnTrial outperforms other algorithms by at least 25% on the long-term QoE impairment.

Performing federated learning continuously in edge networks while training data are dynamically and unpredictably streamed to the devices faces critical challenges, including the global model convergence, the long-term resource budget, and the uncertain stochastic network and execution environment. We formulate an integer program to capture all these challenges, which minimizes the cumulative total latency of stream learning on device and federated learning between devices and the edge server. We then decouple the problem, design an online learning algorithm for controlling the number of local model updates via a convex-concave reformulation and rectified gradient-descent steps, and design a bandit learning algorithm for selecting the edge server for global model aggregations by incorporating the budget information to strike the exploit-explore balance. We rigorously prove the sub-linear regret regarding the optimization objective and the sub-linear constraint violation regarding the maximal on-device load, while guaranteeing the convergence of the global model trained. Extensive evaluations with real-world training data and input traces confirm the empirical superiority of our approach over multiple state-of-the-art algorithms.

Instead of relying on remote clouds, today's Augmented Reality (AR) applications usually send videos to nearby edge servers for analysis (such as objection detection) so as to optimize the user's quality of experience (QoE), which is often determined by not only detection latency but also detection accuracy, playback fluency, etc. Therefore, many studies have been conducted to help adaptively choose best video configuration, e.g., resolution and frame per second (fps), based on network bandwidth to further improve QoE. However, we notice that the video content itself has significant impacts on the configuration selection, e.g., the videos with high-speed objects must be encoded with a high fps to meet the user's fluency requirement. In this article, we aim to adaptively select configurations that match the time-varying network condition as well as the video content. We design Cuttlefish, a system that generates video configuration decisions using reinforcement learning (RL). Cuttlefish trains a neural network model that picks a configuration for the next encoding slot based on observations collected by AR devices. Cuttlefish does not rely on any pre-programmed models or specific assumptions on the environments. Instead, it learns to make configuration decisions solely through observations of the resulting performance of historical decisions. Cuttlefish automatically learns the adaptive configuration policy for diverse AR video streams and obtains a gratifying QoE. We compared Cuttlefish to several state-of-the-art bandwidth-based and velocity-based methods using trace-driven and real world experiments. The results show that Cuttlefish achieves a 18.4-25.8 percent higher QoE than the others.

The booming of Convolutional Neural Networks (CNNs) has empowered lots of computer-vision applications. Due to its stringent requirement for computing resources, substantial research has been conducted on how to optimize its deployment and execution on resource-constrained devices. However, previous works have several weaknesses, including limited support for various CNN structures, fixed scheduling strategies, overlapped computations, high synchronization overheads, etc. In this article, we present DeepSlicing, a collaborative and adaptive inference system that adapts to various CNNs and supports customized flexible fine-grained scheduling. As a built-in functionality, DeepSlicing has supported typical CNNs including GoogLeNet, ResNet, etc. By partitioning both model and data, we also design an efficient scheduler, Proportional Synchronized Scheduler (PSS), which achieves the trade-off between computation and synchronization. Based on PyTorch, we have implemented DeepSlicing on the testbed with real-world edge settings that consists of 8 heterogeneous Raspberry Pi's. The results indicate that DeepSlicing with PSS outperforms the existing systems dramatically, e.g., the inference latency and memory footprint are reduced up to 5.79× and 14.72×, respectively.

Software Defined Networking (SDN) is beneficial to many applications, such as intra-datacenter communication, inter-datacenter transportation, etc., due to its centralized control. However, this centralized control frequently makes the controller a bottleneck, due to the large amount of interactions between the controller and switches. In this paper, we characterize such interactions as control traffic, and propose RouteStitch to minimize such kind of traffic. RouteStitch exploits existing route entries in switches to build new paths. To this end, RouteStitch first builds a graph model to describe existing route entries. Then, on such a model, a novel minimum color-alternation routing problem is defined to minimize control traffic, after which an optimal algorithm is proposed on a fixed routing path. For general paths, an O(log2L)-competitive online algorithm is designed to build new paths in an online manner that preserves fundamental property of switch Ternary Content Addressable Memory (TCAM) capacity and allowed maximum hop length L. Extensive simulation results based on realistic topology show that RouteStitch has good performance in terms of reducing control traffic, by 40%.

Federated learning achieves the privacy-preserving training of models on mobile devices by iteratively aggregating model updates instead of raw training data to the server. Since excessive training iterations and model transferences incur heavy usage of computation and communication resources, selecting appropriate devices and excluding unnecessary model updates can help save the resource usage. We formulate an online time-varying non-linear integer program to minimize the cumulative resource usage over time while achieving the desired long-term convergence of the model being trained. We design an online learning algorithm to make fractional control decisions based on both previous system dynamics and previous training results, and also design an online randomized rounding algorithm to convert the fractional decisions into integers without violating any constraints. We rigorously prove that our online approach only incurs sub-linear dynamic regret for the optimality loss and sub-linear dynamic fit for the long-term convergence violation. We conduct extensive trace-driven evaluations and confirm the empirical superiority of our approach over alternative algorithms in terms of up to 27% reduction on the resource usage while sacrificing only 4% reduction on accuracy.

In this paper, we consider a fundamental problem: given one mobile charger that can charge multiple sensor nodes simultaneously, how we can schedule it to charge a given WSN to maximize the energy usage effectiveness (EUE)? We propose a novel charging paradigm-Overlapped Mobile Charging (OMC)- the first of its kind to the best of our knowledge. Firstly, OMC clusters sensor nodes into multiple non-overlapped sets using k-means evaluated by the Davies-Bouldin Index, such that the sensor nodes in each set have similar recharging cycles. Secondly, for each set of sensor nodes, OMC further divides them into multiple overlapped groups, and charges each group at different locations for different time durations to make sure that each overlapped sensor node just receives its required energy from multiple charging locations.

With the development of deep learning, artificial intelligence applications and services have boomed in the recent years, including recommendation systems, personal assistant and video analytics. Similar to other services in the edge computing environment, artificial intelligence computing tasks are pushed to the network edge. In this paper, we consider the multi-user edge-assisted video analytics task offloading (MEVAO) problem, where users have video analytics tasks with various accuracy requirements. All users independently choose their accuracy decisions, satisfying the accuracy requirement, and offload the video data to the edge server. With the utility function designed based on the features of video analytics, we model MEVAO as a game theory problem and achieve the Nash equilibrium. For the flexibility of making accuracy decisions under different circumstances, a deep reinforcement learning approach is applied to our problem. Our proposed design has much better performance compared with some other approaches in the extensive simulations.

Real-time analytics on video data demands intensive computation resources and high energy consumption. Traditional cloud-based video analytics relies on large centralized clusters to ingest video streams. With edge computing, we can offload compute-intensive analysis tasks to the nearby server, thus mitigating long latency incurred by data transmission via wide area networks. When offloading frames from the front-end device to the edge server, the application configuration (frame sampling rate and frame resolution) will impact several metrics, such as energy consumption, analytics accuracy and user-perceived latency. In this paper, we study the configuration adaption and bandwidth allocation for multiple video streams, which are connected to the same edge node sharing an upload link. We propose an efficient online algorithm, called JCAB, which jointly optimizes configuration adaption and bandwidth allocation to address a number of key challenges in edge-based video analytics systems, including edge capacity limitation, unknown network variation, intrusive dynamics of video contents. Our algorithm is developed based on Lyapunov optimization and Markov approximation, works online without requiring future information, and achieves a provable performance bound. Simulation results show that JCAB can effectively balance the analytics accuracy and energy consumption while keeping low system latency.

The execution and completion of analytics jobs can be significantly inflated by the slowest tasks contained. Despite task replication is well-adopted to reduce such straggler latency, existing replication strategies are unsuitable for geo-distributed analytics environments that are highly dynamic, uncertain, and heterogeneous. In this paper, we firstly model the task replication and scheduling problem over time, capturing the geo-analytics features. Afterwards, we design an online algorithm, GeoClone, to select tasks to replicate and select sites to execute the task replicas in an irrevocably online manner, through jointly considering the execution progress of each job and the resource performance in each site. We rigorously prove the competitive ratio to exhibit the theoretical performance guarantee of GeoClone, compared against the offline optimal algorithm which knows all the inputs at once beforehand. Finally, we implement GeoClone with Spark and Yarn for experiments and also conduct extensive large-scale simulations, which confirms GeoClone's practical superiority over multiple state-of-the-art replication strategies.

Provisioning machine learning inference as a service at the mobile network edge for distributed users in an online setting faces multiple challenges, including the accuracy-resource trade-off for model selection, the time-coupled decision for model distribution, and the unpredictable user inference workload. To overcome such challenges, we firstly model an online time-varying non-linear integer program of maximizing the overall service's inference accuracy through dynamic model instance selection, delivery and workload distribution. Afterwards, we design an online learning algorithm to make fractional control decisions, which alternates between minimizing an outer problem and maximizing an inner problem of an equivalent convex-concave formulation by only taking previously observable inputs. We further design a randomized rounding algorithm to convert the fractional decisions into integers. We rigorously prove that our approach only incurs sub-linear dynamic regret for the optimality loss and sub-linear dynamic fit for the long-term constraints violation. Finally, we conduct extensive evaluations with real- world data and confirm the empirical superiority of our approach over state-of-the-art algorithms in terms of up to 30% reduction on accuracy loss and 34% reduction on constraints violation.

Mobile Augmented Reality (MAR) applications usually contain computation-intensive tasks which far outstrip the capability of mobile devices. One way to overcome this is offloading computation-intensive MAR tasks to remote clouds. However, the wide area network delay is hard to reduce. Thanks to edge computing, we can offload MAR tasks to nearby servers. Prior studies focus on either single-task MAR applications offloading or dependent tasks offloading for a single user. In this article, we study the offloading decision of MAR applications from multiple users, each of which is comprised of a chain of dependent tasks, over a generic cloud-edge system consisting of a group of heterogeneous edge servers and remote clouds. We formulate the Multi-user Multi-task MAR Application Scheduling (M3AS) problem, which is NP-hard. We present Mutas, an efficient scheduling algorithm that jointly optimizes server assignment and resource management. We also consider the online version of M3AS and present OnMutas. Extensive evaluations demonstrate that both Mutas and OnMutas can significantly reduce the service delays of MAR applications when compared to three other heuristics.

Computation offloading makes sense to the interaction between users and compute-intensive applications. Current researches focused on deciding locally or remotely executing an application, but ignored the specific offloading proportion of application. A full offloading cannot make the best use of client and server resources. In this paper, we propose an innovative reinforcement learning (RL) method to solve the proportional computation problem. We consider a common offloading scenario with time-variant bandwidth and heterogeneous devices, and the device generates applications constantly. For each application, the client has to choose locally or remotely executing this application, and determines the proportion to be offloaded. We formalize the problem as a long-term optimization problem, and then propose a RL-based algorithm to solve it. The basic idea is to estimate the benefit of posible decisions, of wihch the decision with the maximum benefit is selected. Instead of adopting the original Deep Q Network (DQN), we propose Advanced DQN (ADQN) by adding Priority Buffer Mechanism and Expert Buffer Mechanism, which improves the utilization of samples and overcomes the cold start problem, respectively. The experimental results show ADQN's high feasibility and efficiency compared with several traditional policies, such as None Offloading Policy, Random Offloading Policy, Link Capacity Optimal Policy, and Computing Capability Optimal Policy. At last, we analyse the effect of expert buffer size and learning rate on ADQN's performance.

The emerging Non-Volatile Memory (NVM) technology has given rise to an opportunity to accelerate big data analysis. In this paper, we investigate the joint job and data scheduling problem in private cloud data center with a hybrid storage system, and we propose Labeling Scheduler, a flexible labeling-based approach for jointly scheduling. The core idea of the approach is to introduce the labeling system to characterize the features of big data analysis jobs and data objects, and conduct data replacement dynamically between NVM and disk. To the best of our knowledge, this is the first work to introduce the labeling methodology to the big data analysis problem in the cloud data center with a hybrid storage system. We conduct extensive simulations and the simulation results show that the Labeling Scheduler has a significant improvement on system utility compared to the method without labeling information. In addition, the Labeling Scheduler guarantees a high NVM hit rate, which is valuable for NVM endurance enhancement.

Popular data analytics systems including Apache Hadoop, Dryad, and Apache Spark abstract jobs as directed acyclic graphs (DAGs). Speeding up completions for DAG jobs matter in practice in order to support real-time decisions. State-of-the-art works propose clairvoyant schedulers to optimize these goals, however, they assume complete job information as a prior knowledge which includes the precise DAG structure, and fine-grained resource requirement and duration time of each task. This assumption limits their applicability. In this paper, to be more practical, we relax the complete prior knowledge assumption and rely solely on partial prior information, based on which, we design a semi-clairvoyant task scheduler Cobra operating within each job. When managing resources for a job, Cobra adaptively adjusts its resource desires in a multiplicative-increase multiplicative-decrease manner on the basis of the its current resource utilization and the presence of current waiting tasks. When assigning tasks to run on the allocated resources, Cobra strives to satisfy task locality preference by tolerating each task waiting for some time that is bounded by a parameterized threshold. Even with the partial prior job information, when a set of jobs in which each employing Cobra as its task scheduler, run on a cluster that employs the fair job scheduler, we theoretically prove the produced makespan and average job response time are O(1)-competitive in different settings. We implement our design in Spark on YARN system, and use experiments from both real deployments and simulations on Google's trace to verify the performance promotion and sensitivity of Cobra.

Wireless energy transfer has been widely studied in recent decades, with existing works mainly focused on maximizing network lifetime, optimizing charging efficiency, and optimizing charging quality. All these works use a charging model with the linear superposition, which may not be the most accurate. We apply a nonlinear superposition model, and we consider the Fast Charging Scheduling problem (FCS): Given multiple chargers and a group of sensors, how can the chargers be optimally scheduled over the time dimension so that the total charging time is minimized and each sensor has at least energy E ? We prove that FCS is NP-complete and propose a 2-approximation algorithm to solve it in one-dimensional (1D) line. In a 2D plane, we first consider a special case of FCS, where the initial phases of all chargers are the same, and propose an algorithm to solve it, which has a bound. Then we propose an algorithm to solve FCS in a general 2D plane. Unlike other algorithms, our algorithm does not need to calculate the combined energy of every possible combination of chargers in advance, which greatly reduces the complexity. Extensive simulations demonstrate that the performance of our algorithm performs almost as good as the optimal algorithm.

Remote Direct Memory Access (RDMA) technique allows the messaging service that directly access the memory on remote machines, which provides low CPU overhead, low latency, and high throughput network transmission. On the other hand, however, due to the limited cache space in RDMA NIC (RNIC), it is still challenging to achieve effective and fair resource sharing across different applications. To address this problem, we present a scalable RDMA as a service to manage resource and deliver fair scheduling to applications' requests. We study the thread contention and preemptive schedule issues at end-hosts, and report the corresponding performance degradation through experiments. Then, we introduce Avatar, a model to manage memory and Queue Pairs (QPs) resource for a large number of connections, which eliminates the lock contention and provides fair data scheduling for applications with different priorities. Finally, we implement Avatar and demonstrate that Avatar can support a thousand of connections, improve the fairness and reduce the requests completion time up to 50% in comparison with the native RDMA.

Edge computing is gaining popularity these years, more service providers are shifting their services from clouds to the edge for better QoS provision. Recent studies in NFV also tend to deploy Network Function Virtualization (NFV) services in the edge network. However, the NFV deployment in the edge network is a challenging problem and differs from the similar problem in data centers. We mainly focus on a new NFV Chain Placement (NCP) problem in the paper. It is well known that the edge of the network is dynamic, and edge computing aims to utilize the physical edge resources efficiently and quickly. We first prove that the NCP problem is NP-complete. Then we propose a novel metric that can better measure the balance condition of the physical resources. We also analyze its advantages. Finally, we design an efficient algorithm, MINI, based on this metric. We evaluate MINI using extensive simulations. The results show that MINI has great advantages over a genetic algorithm (GA) in terms of physical resource utilization, acceptance rate, and running time.

In data center networks (DCN), large scale flows produced by parallel computing frameworks form many coflows semantically. Most inter-coflow schedulers only focus on the remaining data of coflows and attempt to mimic Shortest Job First (SJF). However, a coflow may consist of multiple stages. In this paper, we consider the Multi-stage Inter-Coflow Scheduling problem and try to give an efficient online scheduling scheme. We first explore a short-sighted algorithm with the greedy strategy. This gives us an insight into utilizing the network resources. Based on that, we propose a far-sighted heuristic, which schedules sub-coflows to occupy network bandwidth in turn. Through simulations in various network environments, we show that, compared to a state-of-the-art scheduler - Varys, a multi-stage aware scheduler can reduce the coflow completion time by up to \pmb4.81× even though it is short-sighted. Moreover, the far-sighted scheduler can improve the performance by nearly \pmb7.95 × reduction.

Modern datacenter networks employ Load-balancing (LB) in the large-scale multi-tier topology to ensure high network utilization as well as low flow completion time. This paper presents the design and evaluation of Nem, a robust Erasure Coding (EC) based load balancing scheme at end-host to spread data across multiple paths. Our design is based on two key insights. First, both theory and implementation have shown that redundancy is a powerful technique to reduce latency in networked system. Second, the commercial RDMA network interface card supports EC offload which can dramatically reduce the CPU consumption. Nem is an optimal user-level LB design, which leveraging redundant fine-grained data blocks and high speed lossless RDMA network to realize effective load balancing transmission. Evaluation over many workloads shows that Nem is adaptive to the asymmetric networks, and achieves better performance compared to the state-of-art host-based load balancing mechanism.

Many organizations and companies have deployed not only datacenters but also large number of geo-distributed heterogeneous edges to provide fast data analytics services. Since large volume of data transmission across WAN can be costly, existing works mainly focus on pre-processing data in-place to avoid transmission. However, the heterogeneity of edges on either local computing capacity or network bandwidth limits the efficient use on scarce resource, which may result in long task completion time. To cope with dynamic demands on scarce resource, we take the heterogeneity of both computing capacity and network bandwidth of geo-distributed edges into consideration when assigning data analytical tasks and their associated data between the central datacenter and edges such that the overall latency can be reduced. We formulate the geo-distributed data-task joint scheduling problem (GJS), show its NP-hardness, and propose a near-optimal randomized scheduling algorithm (ran-GJS). ran-GJS can be proved concentrated around its optimum value with high probability, i.e., 1--O(e--t2) where t is the concentration bound by using Martingale Analysis. The experimental results obtained form both extensive simulations and Yarn-based prototype show that ran-GJS significantly speeds up the geo-distributed analytics with a gain on average completion time of at least 28% over state-of-the-art baseline algorithms.

Typical data analytics systems abstract jobs as directed acyclic graphs (DAGs). It is crucial to maximize throughput and speedup completions for DAG jobs in practice. Existing works propose clairvoyant schedulers optimizing these goals, however, they assume complete job information as a prior knowledge which limits their applicability. Instead, we remove the complete prior knowledge assumption and rely solely on a partial prior information, which is more practical. And we design a semi-clairvoyant task scheduler Cobra working within each job. Cobra adaptively adjusts its resource desires in a multiplicative-increase multiplicative-decrease (MIMD) manner according to nearly past resource utilizations and the current waiting tasks. On the other hand, Cobra seeks to satisfy task locality preferences by allowing each task to wait for some time that is bounded by a parameterized threshold. Surprisingly, even with the partial prior job information, we theoretically prove, Cobra, when working with the widely used fair job scheduler, is O(1)-competitive with respect to both makespan and average job response time. We experimentally validate that the performance promotion of Cobra in both real system deployment and trace-driven simulations.

Electric vehicle (EV) is a promising technological tool for diminishing environmental impact caused by gasoline-consumed transportation. Due to the limited battery capacity, EVs need to be charged frequently in a static charging station and thus waste large amounts of time being out of service. Research previously conducted in this topic have proposed solutions for deployment of charging lanes that can charge in-motion EVs. However, they cannot guarantee that every EV can be operational in their respective entire route. Meanwhile, we observe that EVs have repetitive motions and may cyclically encounter with each other which no prior research having been investigated. In addition, the development on the circuit design of energy transmit antennas can render EVs to be able to bi-directionally, highly efficiently transfer energy between themselves. These two observations enable us to distribute energy among EVs in a collaborative and interactive manner. We consider the cases of both loss-less and lossy energy transfer between EVs. In both cases, we formulate the problem of minimizing the time needed (or energy transferred) to reach a given energy distribution into a series of linear programming problems. When compared with a state-of-the-art algorithm, extensive simulation results show that the proposed algorithms can reduce the balancing time and energy loss by up to 70.60% and 36.59%, respectively.

Wireless power transfer is a promising technology used to extend the lifetime of, and thus enhance the usability of, energy-hungry battery-powered devices. It enables energy to be wirelessly transmitted from power chargers to energy-receiving devices. Existing studies have mainly focused on maximizing network lifetime, optimizing charging efficiency, minimizing charging delay, etc. In this paper, we consider wireless charging service provision in a two-dimensional target area and focus on optimizing charging quality, where the power of each charger is adjustable. We first consider the charger Placement and Power allocation Problem with Stationary rechargeable devices (SP3): Given a set of stationary devices and a set of candidate locations for placing chargers, find a charger placement and a corresponding power allocation to maximize the charging quality, subject to a power budget. We prove that SP3is NP-complete, and propose an approximation algorithm. We also show how to deal with mobile devices (MP3), cost-constrained power reconfiguration (CRP), and optimization with more candidate locations. Extensive simulation results show that, the proposed algorithms perform very closely to the optimum (the gap is no more than 4.5, 4.4, and 5.0 percent of OPT in SP3, MP3, and CRP, respectively), and outperforms the baseline algorithms.

Nowadays cloud computing provides an effective way of implementing infrastructure as a service (IaaS). However virtualized data centers still face many challenges, such as low resource utilization of physical machines (PMs) and imbalanced server loads. Virtual machine (VM) consolidation based on live migration allows administrator to dynamically redeploy VMs into PMs for better resource utilization. Common VM consolidation methods usually focus on one challenge, and pay little attention to others or just ignore them, while effective VM redeployment should make tradeoffs between these challenges, and more importantly, should not let other challenges become worse. On the other hand, since VM live migration leads to performance degradation of applications, consolidation work should control migration cost. In this paper, we provide a manner to comprehensively consider power consumption, load balancing, communication delay and migration cost during VM redeployment. And we formalize VM consolidation as a multiobjective optimization problem, then solve this problem with an improved genetic algorithm. Simulation experiments based on real world workload trace show that compared with single objective optimization approaches our method effectively make tradeoffs between optimized objectives and has better overall performance, which is more practical in real data centers.

The lines between IaaS, PaaS, and SaaS are becoming blurred as datacenter providers seek to create cloud platforms that can widen their appeal to developers. With this kind of hybrid datacenter, resource requests from tenants are increasingly transforming into hybrid requests that may simultaneously demand IaaS, PaaS, and SaaS resources. This paper tackles the challenge of modeling and deploying hybrid tenant requests in datacenter networks, for which we coin ``networklet" to represent a set of VMs that collaboratively provide some PaaS or SaaS service. Through extracting networklets from tenant requests and thus sharing them between multiple tenants, we can achieve a win-win situation for datacenter providers and tenants. Extensive evaluations show that, the proposed model and deployment algorithm indeed improve DCN resource utilization while maintaining performance guarantee.

In today's datacenters, resource requests from tenants are increasingly transforming into hybrid requests that may simultaneously demand IaaS, Paas, and SaaS resources. This paper tackles the challenge of modeling and deploying hybrid tenant requests in datacenters, for which we coin "networklet" to represent a set of VMs that collaboratively provide a PaaS or SaaS service. Through extracting networklets from tenant requests and thus sharing them between tenants, we can achieve a win-win situation for datacenter providers and tenants.

Many corporations rely on Internet service provider (ISP) network to provide reliable communication services. However, the current communication networks are vulnerable to disruptive events, such as natural disaster or power outage. Such disastrous events may destroy multiple network facilities in a specific region and result in a long term recovery of ISP networks. The disconnected communication will lead to enormous economic loss even if corporation's infrastructure is not directly destroyed during the disaster. Therefore, corporations need a self-rescue mechanism to actively respond to the emergency instead of simply relying on the ISP. This paper proposes Ambula, an easy-to-deploy platform to realize fast congestion-aware recovery for corporation's communication lifeline. Our platform leverages current widely-deployed public cloud services to build a scalable peer-to-peer overlay routing system. By so doing, the corporation is capable of controlling the packets forwarding path to bypass the affected region and congested routes. To this end, Ambula first carefully selects a small set of virtual machines (VMs) from geographically distributed public clouds, and then apply the self-developed congestion-aware routing protocol to achieve automatic and fast routing recovery. Simulations on both random generated and real network topologies show that the high recovery ratio of 80% can be achieved. The congestion avoidance algorithm can significantly reduce the impact of congestion. Our prototype on Emulab shows it can recover within hundreds of milliseconds. To the best of our knowledge, no effective disaster recovery mechanism currently exists for corporations during emergency. Ambula will facilitate the business continuity management of corporations in present of hazard events.

Hardware middleboxes are widely used in current cloud datacenter to provide network functions such as firewalls, intrusion detection system, load balancers, etc. Unfortunately, they are expensive and unable to offer customized functions for individual tenant. To overcome this issue, there is an increasing interest in deploying software middleboxes to enable flexible security, network access functionality. This paper addresses the software middleboxes placement problem with minimum bandwidth guarantee. We first specify the model of tenants' requirement that specifies the need for virtual machines of application and middleboxes, as well as communication traffic. A virtual middlebox placement algorithm called MISSILE is then proposed to offer predictable network performance for each accepted tenant, and minimize datacenter bandwidth utilization. Extensive simulation results based on current large-scale datacenter networks verify that MISSILE is effective and provides network performance guarantee for tenants.

In this paper, we investigate the joint job and data assignment problem in cloud data centers with non-volatile memory (NVM) for makespan minimization. Through extensive analysis, we find that there is an indicator variable that characterizes the hardness of the problem. Depending on the value of the indicator variable, we classify our problem into three cases: inf-case, opt-case, and nph-case. We first show that there is no feasible assignment under the inf-case. For the opt-case, we present an optimal algorithm. We show that a mixed data assignment with diversified popularity achieves high memory utilization. For the nph-case, we first prove the problem's NP-hardness and then propose a heuristic algorithm and a 2-approximation algorithm to tackle it. We conduct extensive simulations, and we find that the performance of the heuristic algorithm is better than the 2-approximation algorithm and that it is nearly the same as the theoretical optimal solution.

The various flows in production datacenters usually can be classified into two types: bandwidth-hungry and delay-sensitive. To improve their performance, datacenter networks require effective load balancing and flow control protocols, respectively. However, as the two techniques are typically employed separately in current datacenters, they are unable to optimize the network in a coordinated way. In this work, we argue that the adaptive routing, in load balancing sense, and the flow control, in congestion control sense, could be tightly coupled at the transport layer to handle the complex datacenter traffic. We design OmniFlow, a novel transfer protocol which aims to achieve a proper balance between throughput and latency in a datacenter. Firstly, it can simultaneously and precisely measure the queueing latencies on multiple paths between two hosts, which enables it to have more visibility of the path congestion and have better control of the transmission states. Secondly, OmniFlow adaptively integrates the load balancing and flow control modules and shares the same congestion metrics (i.e. queueing latencies) between them. Based on different network conditions, it either dynamically reroutes flows to utilize the bisection bandwidth or proactively adjusts flow rates to bound queueing occupancies. The results of extensive experiments show that OmniFlow can provide both low average and tail latency for small flows without sacrificing the throughput of elephant flows.

Wireless power transfer provides a promising way to extend the battery lifetime of our energy-hungry rechargeable devices. Previous studies have envisioned using mobile vehicles/robots/drones equipped with high capacity batteries as mobile chargers to replenish those devices, and they mainly focus on maximizing network lifetime, optimizing efficiency of charging scheduling, minimizing total charging delay, etc. However, existing methods may be insufficient and inflexible when the energy consumption of rechargeable devices fluctuates overtime, or when rechargeable devices are sparse. In this paper, we consider how to efficiently provide flexible wireless charging using pre-planned charging itineraries. We present the Itinerary Selection and Charging Association (ISCA) problem: given a set of rechargeable devices and a set of candidate charging itineraries, how can we select itineraries and determine a corresponding charging association to minimize the amount of energy which is due to mobile chargers' movement and wireless charging loss, so that every device gets its required energy. We prove that ISCA is NP-complete by reducing the set cover problem to it. We start solving this problem by first looking at the case in which an itinerary can only be used once, and we propose an algorithm with approximation ratio of O (lnM) and a practical heuristic algorithm, where M is the number of devices. For the general case in which an itinerary may be used multiple times, we propose an approximation algorithm of factor 10 using the Primal-Dual schema. Evaluations results from real field experiments and extensive simulations show that the proposed algorithms have near-optimal performance and PDA reduces the amount of wasted energy by up to 65 percent compared with a set cover-based algorithm.

Existing data center transport protocols are usually based on the Processor Sharing (PS) policy and/or the Shortest Remaining Processing Time (SRPT) policy. PS divides link bandwidth equally between competing flows, thus it fails to achieve optimal average flow completion time (FCT). SRPT prioritizes flows that have the shortest remaining processing time and provides near-optimal average FCT, but it may cause long flows to suffer unfair delays, or even starve them. In fact, these two types of policies represent two directions in the design space: PS prefers fairness (in terms of starvation freedom) while SRPT favors efficiency (in terms of average FCT). In this paper, we propose a novel metric, expansion ratio, which enables us to strike a balance between SRPT and PS. We design MERP that achieves efficient flow scheduling without starvation. MERP takes care of both average and tail FCTs by minimizing the expansion ratio of competing flows in a lexicographically manner. MERP controls the sending rate of competing flows via synchronized virtual deadlines and routes flows in a downstream-aware manner that reacts quickly to link failures. We evaluate MERP using extensive NS2-based simulations. Results show that, under various traffic loads, MERP reduces the tail FCT significantly with a negligible increase of average FCT compared with pFabric, and MERP reduces the average FCT notably compared with ECMP and CONGA when link failures occur.

In this paper, we propose OmniFlow, a novel transport protocol which combines load balancing and flow control at the transport layer to optimize datacenter transfers. OmniFlow outweighs previous solutions in two aspects. Firstly, it can simultaneously and precisely measure the queueing latencies on multiple paths between two hosts. Secondly, OmniFlow adaptively integrates the load balancing and flow control modules and shares the same congestion metrics (i.e. queueing latencies) between them. Based on different network conditions, it either dynamically reroutes flows to utilize the bisection bandwidth or proactively adjusts flow rates to bound queueing occupancies. Our results show that OmniFlow can provide both low latency for small flows without sacrificing the throughput of elephant flows.

Virtualization technology has brought new vitality to data centers but also brought some thorny issues. Virtualization technology creates an abstract intermediate layer, separating the upper layer applications from the underlying infrastructure, which cause some difficulties to the upper layer applications for the effective use of resources. Based on this observation, we propose a dynamic load balancing system in virtualized environment. The system adjusts the weights of virtual machines in real-time in order to balancing physical servers' load and improve data center efficiency. In addition, by monitoring physical machine failure information, we can evacuate physical machines as soon as they failed to ensure high availability of the data center. We design a virtualized environment monitoring system and propose an efficient algorithm. We evaluate the proposed system with real implementations which show DLB has rather good performance.

In computing clouds, burstiness of a virtual machine (VM) workload widely exists in real applications, where spikes usually occur aperiodically with low frequency and short duration. This could be effectively handled through dynamically scaling up/down in a virtualization-based computing cloud; however, to minimize energy consumption, VMs are often highly consolidated with the minimum number of physical machines (PMs) used. In this case, to meet the dynamic runtime resource demands of VMs in a PM, some VMs have to be migrated to some other PMs, which may cause potential performance degradation. In this paper, we investigate the burstiness-aware server consolidation problem from the perspective of resource reservation, i.e., reserving a certain amount of extra resources on each PM to avoid live migrations, and propose a novel server consolidation algorithm, QUEUE. We first model the resource requirement pattern of each VM as a two-state Markov chain to capture burstiness, then we design a resource reservation strategy for each PM based on the stationary distribution of a Markov chain. Finally, we present QUEUE, a complete server consolidation algorithm with a reasonable time complexity. We also show how to cope with heterogenous spikes and provide remarks on several extensions. Simulation and testbed results show that, QUEUE improves the consolidation ratio by up to 45 percent with large spike size and around 30 percent with normal spike size compared with the strategy that provisions for peak workload, and achieves a better balance between performance and energy consumption in comparison with other commonly-used consolidation algorithms.

To improve users' experience, TCP short flows that are heavily used in interactive services should be completed as soon as possible. In current data centers, large flows and head-of-line blocking in switches hinder short flows from completion, which leads to long-tailed flow completion times (FCT). Replicating short flows with multiple equal-cost paths is a promising way to reduce FCT. However, the original flow and its replicated one are quite likely to be routed to the same path (ECMP hash collision), which increases both the mean and 99-percentile FCT significantly. What's more, inadequate replication leaves many other less-congested equal-cost paths unused and limits the performance while excess replication degrades throughput of large flows. To solve these problems, we propose SmartRep, a scheme consisting of an efficient and effective traceroute based hash collision avoidance method and an algorithm to decide the optimal number of replicated flows for different short flows. SmartRep can be easily implemented in software and readily deployed in data centers. Extensive NS2 simulations show that our approach improves previous replication-based work by 25%-50% in both mean and 99th percentile FCT, and meanwhile imposes negligible impact on large flows.

Wireless power transfer is a promising technology to extend the lifetime of, and thus enhance the usability of, the energy-hungry battery-powered devices. It enables energy to be wirelessly transmitted from power chargers to energy receiving devices. Existing studies have mainly focused on maximizing network lifetime, optimizing charging efficiency, minimizing charging delay, etc. Different from these works, our objective is to optimize charging quality in a 2-D target area. Specifically, we consider the following charger Placement and Power allocation Problem (P3): Given a set of candidate locations for placing chargers, find a charger placement and a corresponding power allocation to maximize the charging quality, subject to a power budget. We prove that P3is NP-complete. We first study P3with fixed power levels, for which we propose a (1-1/e)-approximation algorithm; we then design an approximation algorithm of factor 1-1/e / 2L for P3, where e is the base of the natural logarithm, and L is the maximum power level of a charger. We also show how to extend P3in a cycle. Extensive simulations demonstrate that, the gap between our design and the optimal algorithm is within 4.5%, validating our theoretical results.

Today's public cloud providers typically deploy their small sized data centers in multiple geographically different locations, so as to improve data center power usage effectiveness and locate resources closer to users. A major challenge is resource allocation. Many results have been reported regarding this issue from the perspectives of virtual machine consolidation, network-aware virtual machine placement, traffic engineering, dynamic capacity provisioning, and so on. However, there has not been any focus on stable resource allocations, where no resource request or data center has any migration incentives. To the best of our knowledge, this paper is the first attempt at gaining a better understanding of the structure of the Stable rEsource Allocation (SEA) problem. We introduce a formal problem statement and develop two algorithms for the 1-dimensional (1-D) and 2-D cases, respectively. Simulation results show that the proposed algorithms have good scalability and convergence.

Virtual networks that allow tenants to explicitly specify their computing as well as networking resources are recently proposed to be better interfaces between cloud providers and tenants. Many virtual networks have time-varying resource demands, as evidenced in prior studies [1-3]. New opportunities emerge when such variation is exploited. In this paper, we design a novel resource demand model for tenants to flexibly trade off between application performance and cost, and propose a work-conserving allocation algorithm, WCA, for deploying virtual networks with time-varying resource demands. WCA places virtual nodes in a first-fit fashion, and places virtual links through path-splitting. In each physical node or link, by opportunistically sharing physical resources among multiple variable parts of resource demands, physical utilization can be improved, and more virtual networks can be deployed concurrently. Our evaluation results show that WCA achieves a 4% higher physical resource utilization and rejects 18% less virtual network requests than a state-of-the-art algorithm [4].

With the rapid development of virtualization techniques, modern data centers move into a new era of cloud in recent years. Despite numerous advantages such as high resource utilization and rapid service scalability, current virtualization techniques don't guarantee perfect performance isolation among virtual machines sharing the physical machine, which may lead to unstable and unpredictable user-perceived application performance in clouds. Therefore, understanding and modeling performance interference among collocated applications is of utmost importance. However, the hypervisor and guest OSes usually run independent resource schedulers and are invisible into each other, thereby making accurately characterizing performance interference a non-trivial work. In this paper, we first present a comprehensive experimental study on performance interference of different combinations of benchmarks, observing that virtual CPU floating overhead between multiple physical CPUs, and VMEXITs, i.e., the control transitions between the hypervisor and VMs, constitute the key source of performance interference. In order to characterize the performance interference effects, we measure both the application-level and VM-level characteristics from the collocated applications and then build a novel interference prediction framework based on kernel canonical correlation analysis. Our evaluations first show the practicability of KCCA in finding reliable correlation, and further confirm the high accuracy and great applicability of our interference model with a low prediction error of no more than 7.9%.

Flow scheduling is one of the primary issues in data centers. The efficiency of flow scheduling affects the user experience significantly, since the service latency is determined by the flow completion time (FCT). Current transport protocols are based on the Processor Sharing (PS) policy by dividing the link bandwidth equally. As short flows may be blocked by long flows under PS policy, these protocols could not meet latency requirements. The Shortest Remaining Processing Time (SRPT) policy provides a near-optimal solution in term of reducing average FCT. However, this flow scheduling policy may cause long flows suffering unfair delays. In this paper, we propose MERP (Minimizing Expansion Ratio Protocol) for flow scheduling, which aims to navigate the tradeoff between fairness and the average FCT. We propose expansion ratio as a new metric to describe the gap between reality and ideal in terms of acquired link resources by a flow. Larger expansion ratio indicates that the flow obtains fewer bottleneck link resources. We strive to make the flow scheduling relatively fair by reducing the largest expansion ratio of all the flows. And we further demonstrate that reducing expansion ratio is more reasonable than reducing average FCT for services with the Partition/Aggregate pattern. And to meet the scalability requirements, we implement distributed MERP through flow preemption and explicit rate control. The experiments indicate that MERP could significantly reduce the completion time for 99.9th percentile flows under high load.

Internetware challenges distributed systems in aspects from operating platforms, programming models, to engineering approaches, etc. Cloud computing based on virtualization is now a popular paradigm which can meet the dynamic resource allocation requirements of Internetware. Software entities dispersed on distributed nodes over the Internet, now are evolving into self-contained, autonomous software services. These software entities which are often deployed on virtual machines (VMs), are coordinated dynamically to achieve flexible design objectives. To improve the utilization of infrastructure resource, VMs processing components of a software should be consolidated to fewer physical machines (PMs). However, as the increasing trends of communication-intensive softwares, data traffic among VMs should be considered as well. And for the sake of safety and QoS (Quality of Service), certain VMs (e.g. backup nodes) are mutually-exclusive which means some VMs require to be placed on different PMs. In this paper, we investigate the online software placement problem with the target to minimize the network traffic cost, while taking into account the mutually-exclusiveness of VMs. We provide a formal problem description and its NP-hardness analysis. The proposed algorithm places the VMs that have heavy traffic on the same PM, while isolating the mutually-exclusive VMs simultaneously, which can guarantee high effectiveness and reliability for softwares in Internetware, respectively. The simulations show our algorithm reduces the traffic cost by 29% compared against the existing approaches.

Network virtualization has emerged as a promising approach to overcome the ossification of the Internet. A major challenge in network virtualization is the so-called virtual network embedding problem, which deals with the efficient embedding of virtual networks with resource constraints into a shared substrate network. A number of heuristics have been proposed to cope with the NP-hardness of this problem; however, all of the existing proposals reserve fixed resources throughout the entire lifetime of a virtual network. In this paper, we re-examine this problem with the position that time-varying resource requirements of virtual networks should be taken into consideration, and we present an opportunistic resource sharing-based mapping framework, ORS, where substrate resources are opportunistically shared among multiple virtual networks. We formulate the time slot assignment as an optimization problem; then, we prove the decision version of the problem to be NP-hard in the strong sense. Observing the resemblance between our problem and the bin packing problem, we adopt the core idea of first-fit and propose two practical solutions: first-fit by collision probability (CFF) and first-fit by expectation of indicators' sum (EFF). Simulation results show that ORS provides a more efficient utilization of substrate resources than two state-of-the-art fixed-resource embedding schemes.

Network virtualization currently becomes an important technology in optimizing resource management in a datacenter. The major research issue in this area is the virtual network embedding problem (VNEP) concerned with mapping virtual networks (VNs) onto a shared substrate network (SN) with some constrains satisfied. In this paper, by proposing the new concepts of “node sharing” and “partially accepting”, we try to break the “atomicity” of the traditional solutions of VNEP, which usually treat a VN request as an inseparable whole. We consider a more realistic scenario where virtual nodes in a VN request are of different importance and accordingly devise a mapping algorithm called Best Effort Algorithm (BEA). Our algorithm can make the most of the resources through node sharing at idle period of the substrate network and provide the best-effort mapping by partially accepting VN requests when physical resources become scarce. Final simulations demonstrate the effectiveness of our algorithm and show that our novel ideas can lead to higher acceptance ratio and long-term average revenue.

In a social network, information runs from word-of-mouth based on the relationship of the users. The influence maximization is to find a limited number of initial users (nodes) to spread the information, so that the maximum number of other users could accept the information, which is a useful technique for marketing, information monitoring and advertising in a social network. Diffusion model of social networks imitates the process of information spreading in social networks, and Independent Cascade (IC) Model and Linear Threshold (LT) Model, are well-known stochastic information influence models. In this paper, we extend the classical IC model according to the observation of users' behaviors in social networks and propose an effective influence maximization algorithm based on this extended IC model. This novel algorithm calculates the influence probability of each node in sub-graphs that other nodes can engendered to it iteratively. The simulation experiments on real social network datasets show that our algorithm is much faster than the greedy hill-climbing algorithm, while the results are very close to the greedy algorithm and out-perform the other heuristic algorithms.

Burstiness is a common pattern of virtual machines (VMs)'s workload in production data centers, where spikes usually occur aperiodically with low frequency and last shortly. Since virtualization technology enables elastic resource provisioning in a computing cloud, the bursty workloads could be handled effectively through dynamically scaling up/down. However, to cut back energy consumption, VMs are usually highly consolidated with the minimum number of physical machines (PMs) used. In this case, to meet the runtime expanding demands of the resources (spikes), some VMs have to be migrated to other idle PMs, which is costly and causes performance degradation potentially. In this paper, we investigate the elastic resource provisioning problem and propose a novel VM consolidation mechanism with resource reservation which takes burstiness into consideration as well as energy consumption. We model the resource requirement pattern as the popular ON-OFF Markov chain to represent burstiness, based on which a reservation strategy via queuing theory approach is given for each PM. Next we present a complete VM consolidation scheme with resource reservation within reasonable time complexity. The experiment result show that our algorithms improve the consolidation ratio by up to 45% with large spike size and around 30% with normal spike size compared to those provisioning for peak workload, and a better balance of performance and energy consumption is achieved in comparison with other commonly used consolidation algorithms.

Network virtualization provides a promising way to overcome Internet ossification. A major challenge is virtual network mapping, i.e., how to embed multiple virtual network requests with resource constraints into a substrate network, such that physical resources are utilized in an efficient and effective manner. Since this problem is known to be NP-complete, a variety of heuristic algorithms have been proposed. In this paper, we re-examine this problem and propose a virtual network mapping framework, ORS TA, which is based on Opportunistic Resource Sharing and Topology-Aware node ranking. Opportunistic resource sharing is taken into consideration at the entire network level for the first time and we develop an online approximation algorithm, FFA, for solving the corresponding time slot assignment problem. To measure the topology importance of a substrate node, a node ranking method, MCRank, based on Markov chain is presented. We also devise a simple and practical method to estimate the residual resource of a substrate node/link. Extensive simulation experiments demonstrate that the proposed framework enables the substrate network to achieve efficient physical resource utilization and to accept many more virtual network requests over time.

Cloud computing is a newly emerging reliable and scalable paradigm in which customers pay for cloud resources they use on demand. However, current auto-scaling mechanisms in cloud lack the critical self-adaption policy which helps application providers decide on when and how to reallocate resources. Furthermore, virtualization techniques can not ensure an absolute isolation between multiple virtual machines sharing the same physical resource, which leads to some customers paying unfairly for heavy-loaded resource under a widely-adopted fixed pricing scheme.

In traditional stop-and-wait strategy for reliable communications, such as ARQ, retransmission for the packet loss problem would incur a great number of packet transmissions in lossy wireless ad-hoc networks. We study the reliable multicast lifetime maximization problem by alternatively exploring the random linear network coding in this paper. We formulate such problem as a min-max problem and propose a heuristic algorithm, called maximum lifetime tree (MLT), to build a multicast tree that maximizes the network lifetime. Simulation results show that the proposed algorithms can significantly increase the network lifetime when compared with the traditional algorithms under various distributions of error probability on lossy wireless links.

Network virtualization has emerged as a powerful way to fend off the current ossification of the Internet. A major challenge is virtual network mapping, which is to assign substrate resources to virtual networks (VNs) such that some predefined constraints are satisfied and substrate resources are utilized in an effective and efficient manner. Due to the NP-completeness of this problem, a variety of heuristic algorithms have been proposed. However, existing solutions rarely consider the inefficient utilization of bandwidth resources due to the network traffic fluctuation. In this paper, we study the opportunistic bandwidth sharing in a single physical link among multiple virtual links from different VNs. We formulate the problem of assigning time slots to dispensable sub-flows with constraints on the performance guarantee and the objective of minimizing the number of time slots used, as an optimization problem. Two heuristic algorithms HA-I and HA-II, which consider the problem from different perspectives, are presented. Extensive simulations are conducted to evaluate the effectiveness and efficiency of our algorithms.

Service composition is a useful technology to achieve dynamic requests. But in cloud, the cloud provider should guarantee the QoS of the user request as well. Previous work investigated how to select proper available replicas (the similar functional services located on different physical nodes) to effectively achieve the composite services. However, as the requirements for some service grow, even the optimal selection strategy could not satisfy all the QoS requests, if it is only based on the existed replicas. In this case, cloud provider should deploy more replicas to meet the growing requests. Some existing literatures have addressed the service redeployment problem, aiming to optimize the overall or average performance; however, few of them can guarantee QoS of each request. This paper investigates QoS-aware service redeployment problem (SRP), with objective to minimize the redeployment cost. We show that, it is NP-hard to decide whether there exists a feasible solution of SRP. Thus we propose a novel heuristic algorithm SRA, which can find a solution such that most of the requests can be satisfied, while the deployment cost is minimized. Experimental results show that our approach is effective and efficient.

Network virtualization has emerged as the most promising approach to overcome the current ossification of the Internet. A key problem in it is how to efficiently and effectively make use of the substrate network resources by embedding multiple virtual networks with various constraints. Due to its NP-hardness, many heuristic approaches have been proposed. However, most of them simplify the problem by relaxing some constraints, resulting in their designs in conflict with practical limitations. Furthermore, some other important design issues, like load balancing and the response time requirements of different applications, have been usually ignored as well. In this paper, we propose FELL, a Flexible virtual network Embedding algorithm with guaranteed Load baLancing for the general problem. Based on simulated annealing, FELL can flexibly control the tradeoff between results accuracy and running time to meet various requirements of different applications by changing parameters. Load balancing enables substrate network to avoid resource fragmentation and further increases the profit of infrastructure providers. A novel cost criterion that reflects the impact of distribution of allocated resources for an embedding is designed to conduct embedding process to guarantee it. The splittable flow is supported to obtain better resource utilization by making use of small pieces of available bandwidth. We also design some key functions including generating initial and neighbor solutions. The effectiveness of our algorithm is finally validated by our simulations.

Context data is important information for the behaviors of the applications in a pervasive space. To effectively restore huge amount of data, tree-liked layered storage architecture are proposed, where the leaf nodes collect the data from the sensing devices located in its domain. However, the sensing devices may be moving among different domains. In order to integrate the data from the same device, related leaf nodes should upload and store the data to a certain up-layer node, called host node. This paper presents a deep study of the data storage problem and proposes an online algorithm DHS to dynamically select the host node, which reduces the communication cost significantly. We prove the correctness of the algorithm theoretically. The experiment results also show that DHS is correct and effective.

The wireless sensor networks (WSNs) are normally operated in unattended, harsh, or hostile environment. Some strategies have been proposed to establish pairwise keys to protect the sensitive data and the sensor readings, but most existing schemes either suffer the large-scale node capture attacks, or cannot provide full and direct key establishment. In this paper, we present a permutation-based multi-polynomial scheme for pairwise key establishment in wireless sensor networks. This scheme is highly robust to the large-scale node capture attacks. Even after a large number of nodes have been compromised, the pairwise keys shared by non-compromised nodes remain secure. It also guarantees that any two nodes can directly establish a pairwise key without exposing any secret to other nodes. This full and direct key establishment features enable our scheme every adaptive to support node addition and mobility, which are particularly beneficial to the real-world applications. Our performance analysis also shows that the proposed scheme incurs low communication, storage, and computation overhead.

A lot of network services and applications are being designed to support quality-of-service (QoS) routing. One of the key problems is to find a feasible path that satisfies multiple QoS requirements, i.e., Multi-Constrained Path (MCP) problem which is known to be NP-complete. Many heuristic and approximation algorithms have been proposed to solve this problem. However, most of them converted it into the classic shortest path problem by transforming multiple QoS weights into single weight. In this paper, we propose a binary graph reduction (BGR) algorithm to convert network graph into a simplified graph by removing redundant edges before constructing a routing path. BGR improves system performance from two aspects: i) BGR decreases the decision-making time, which then decreases the delay of end users; ii) BGR removes some abnormal anti-heuristic redundant edges from network graph, so that the actual routing algorithm could get a better result. We use a multimedia delivery system to illustrate these advantages in this paper. Simulation results also approve the efficiency of the proposed BGR algorithm.

Pervasive computing, the new computing paradigm aiming at providing services anywhere at anytime, poses great challenges on dynamic service composition. Existing service composition methods can hardly meet the requirements of dynamic characteristic and heterogeneity in pervasive computing environment. In this paper, we propose a parameter-based service model to accurately describe pervasive services. Based on the model, pervasive services are aggregated in a two-layer graph according to both semantic and syntactic information of the input and output parameters. Moreover, we design a novel service composition scheme to accomplish the user task while satisfy the QoS requirements. Both theoretical analysis and simulation experiments show that this service composition mechanism is effective in pervasive environment.

Service-oriented multimedia processing system is an effective approach to allow dynamic and customized service requests in a pervasive environment. This paper studies the context-aware multimedia processing problem under the consideration of the QoS from two aspects: delay and media quality. The second metric is introduced based on the factor that different display devices can adapt to different media quality to the same request. We propose a new measurement, the media quality change ratio, to quantitively describe the adaptability of source media to meet quality requirement of different devices. We further investigate the important issue on how to find an optimal service path according to the users' requirements which are formulated as functional paths. Previous work on the path construction is only based on one fixed functional path. We propose a heuristic algorithm CMpath that explores multiple functional paths and thus can generate a service path to meet both low-delay and media quality requirements. Our experimental results also validate the proposed algorithm.

Service composition is a useful technique to assemble light, independent services to meet the complicated and dynamic requirements. Previous research has addressed the quality-of-service (QoS) aware composition path selection problem. However, as the requests growing, the selected service composition path may violate the QoS requirements. In this case, more service replicas should be deployed on suitable nodes to improve the QoS. But which service component should be selected and where these service replicas should be deployed is a challenge. In this paper, we make a deep study on this service replication problem. We give a detailed description of service replication triggering time. And then, we propose LDCS (Longest Delay Service Component Selection) to select the bottle-neck service component by evaluating the real-time performance of all these components. Finally, we employ MACP (Maximum Available Capacity Path) algorithm to select a suitable node to deploy this service replica. Simulation results approve that our approach is effective and efficient.

Service composition is an effective approach for multimedia delivery in pervasive environment. In previous works, there is one fixed functional path which is composed of several underlying services in a certain order. Actually, there are several functional paths delivering different quality level multimedia from the source to the end user. Due to the dynamicity and mobility of pervasive space, system should generate a reliable and low-delay service path for multimedia delivery in real-time. Since some multimedia service components change the data transmission volume which has a deep impact on the transmission delay, it makes the media delivery problem equal to Multi-Constrained Path problem which is known to NP-Complete. We propose an efficient algorithm LD/RPath(Lowest Delay/Reliability Path) for adaptive multimedia delivery. LD/RPath generates a low-delay service path based on several functional paths with reliability guarantee. Experiment results show that LD/RPath has a good performance and it is an effective algorithm for multimedia delivery in pervasive space.

Service composition is an effective approach for large- scale multimedia delivery. One of the challenge issues is how to choose services to build a multimedia delivery path from source to destination based on user's requirements. In previous works, user's requirement is represented as one fixed functional path which is composed of several functional components in a certain order. Actually, there may be several functional paths (deliver different quality level multimedia data, e.g. image pixel, frame rate) that can meet one request. And due to the diversity of devices and connections in pervasive environment, system should choose a suitable media quality delivery path in accordance with context, instead of based on one fix functional path. This paper proposes LDpath which aims at delivering multimedia data to end users with lowest delay. It chooses services to build delivery path hop-by-hop and the generated path matches one of the possible functional paths, which essentially achieves the context aware multimedia delivery. Furthermore, the amount of data transmission deeply affects delay, thus, LDpath considers data volume changing as one of the metrics of service selection. Experimental results show that LDpath is an effective approach for multimedia delivery in pervasive space.

Service composition is an effective method to achieve flexible pervasive application. According to the relationship of messages and behaviors, we define a message oriented activity based Petri net (Moap) model to describe service, which supports concurrent processes and the reuse of composite service. And user's loose requirement is denoted as goal which is composed of expected input, output and key behaviors. Based on Moap, an automatic service composition algorithm autoSC is proposed to automatically create a specification of composite process to achieve goal. Finally, we analyze the effectiveness of autoSC and give a detailed comparison with other methods.