Search

In a large cloud data center HPC system, a critical problem is how to allocate the submitted tasks to heterogeneous servers that will achieve the goal of maximizing the system’s gain defined as the value of completed tasks minus system operation costs. We consider this problem in the online setting that tasks arrive in batches and propose a novel deep reinforcement learning (DRL) enhanced greedy optimization algorithm of two-stage scheduling interacting task sequencing and task allocation. For task sequencing, we deploy a DRL module to predict the best allocation sequence for each arriving batch of tasks based on the knowledge (allocation strategies) learnt from previous batches. For task allocation, we propose a greedy strategy that allocates tasks to servers one by one online following the allocation sequence to maximize the total gain increase. We show that our greedy strategy has a performance guarantee of competitive ratio $\frac{1}{1+\kappa }$ 1 1 + κ to the optimal offline solution, which improves the existing result for the same problem, where $\kappa$ κ is upper bounded by the maximum cost-to-gain ratio of each task. While our DRL module enhances the greedy algorithm by providing the likely-optimal allocation sequence for each batch of arriving tasks, our greedy strategy bounds DRL’s prediction error within a proven worst-case performance guarantee for any allocation sequence. It enables a better solution quality than that obtainable from both DRL and greedy optimization alone. Extensive experiment evaluation results in both simulation and real application environments demonstrate the effectiveness and efficiency of our proposed algorithm. Compared with the state-of-the-art baselines, our algorithm increases the system gain by about 10% to 30%. Our algorithm provides an interesting example of combining machine learning (ML) and greedy optimization techniques to improve ML-based solutions with a worst-case performance guarantee for solving hard optimization problems. [ABSTRACT FROM AUTHOR]

Deterministic executions of computation-intensive and delay-sensitive tasks are vital for realizing Industry 4.0. Multi-access edge computing (MEC) enables critical tasks to be offloaded to the network edge for quickly responding and processing tasks locally. However, during workload peaks at the resource-limited edge, the best effort IP forwarding in the typical network architecture incurs long-tail latency at both the access and interconnection networks of MEC systems. The long-tail latency results in strict service level agreements (SLAs) unable to be guaranteed in end-to-end communication. In this article, we propose a task deterministic network (TDN) architecture that provides communication with bounded low latency and zero jitter for critical tasks among MEC systems. In TDN, the cross-domain collaboration mechanisms of the layer 2/layer 3 deterministic networking and 5G are jointly designed to eliminate the long-tail latency in typical MEC system networks. In addition, we design a seamless working scheme for the time-sensitive networking controller, the MEC orchestrator, and other coordinators to enable edge collaboration and global deterministic forwarding. Moreover, the main functions of TDN are detailed, including clock synchronization, network slicing, cycle mapping, and multi-layer collaboration. Finally, we present simulation results to substantiate the performance of TDN for accommodating critical service flows among MEC systems. [ABSTRACT FROM AUTHOR]

We constructed an open multi-access network platform using open source hardware and software. The open multi-access network platform is characterized by flexible utilization of network functions, integral management and control of wired and wireless access networks, zero-touch provisioning, intelligent resource monitoring, and dynamic task offloading. We also propose application-driven dynamic task offloading that utilizes intelligent resource monitoring to ensure effective task processing in edge and cloud servers. For this purpose, we developed a mobile application and server applications for the open multi-access network platform. To investigate the feasibility and availability of our developed platform, we experimentally and analytically evaluated the effectiveness of application-driven dynamic task offloading and intelligent resource monitoring. The experimental results demonstrate that application-driven dynamic task offloading could reduce real-time task response time and traffic over metro and core networks. [ABSTRACT FROM AUTHOR]

Web virtual reality (WebVR) is gaining increasing attention as interactive VR experiences become more prevalent. The high energy consumption and strict quality requirements are still significant challenges for its application in the mobile edge network. In this context, content caching and task offloading are promising solutions to save the energy of WebVR. This paper proposes a hybrid decentralized offloading architecture in which the MEC server and multiple terminals jointly participate in the caching and processing of WebVR tasks. In this framework, the graph convolutional network and unsupervised clustering algorithm are jointly applied to process the WebVR service feature graph to achieve FoV-level content caching. Moreover, WebVR users with idle computing resources assist neighboring users with performing tasks. On this basis, we propose a distributed low-bound-based alternative direction method of multiplier (LADMM) algorithm to optimize the offloading mode and the allocation of tasks and computing power resources to minimize system energy consumption. The proposed offloading mode can lower energy consumption while maintaining a good balance between delay performance and resource utilization. [ABSTRACT FROM AUTHOR]

Natural disasters often cause huge and unpredictable losses to human lives and properties. In such an emergency post-disaster rescue situation, unmanned aerial vehicles (UAVs) are effective tools to enter the damaged areas to perform immediate disaster recovery missions, owing to their flexible mobilities and fast deployment. However, UAVs typically have very limited battery and computational capacities, which makes them harder to perform heavy computation tasks during the complicated disaster recovery process. This paper addresses the issue of the battery and computation resource limitation with a fog computing based UAV system. Specifically, we first introduce the vehicular fog computing (VFC) system in which the unmanned ground vehicles (UGVs) perform the computation tasks offloaded from UAVs. To avoid the transmission competitions yet enable cooperations among UAVs and UGVs, a stable matching algorithm is developed to transform the computation task offloading problem into a two-sided matching problem. An iterative algorithm is then developed which matches each UAV with the most suitable UGV for offloading. Finally, extensive simulations are carried out to demonstrate that the proposed scheme can effectively improve utilities of UAVs and reduce average delay through comparison with conventional schemes. [ABSTRACT FROM AUTHOR]

A variety of novel mobile applications are developed to attract the interests of potential users in the emerging 5G-enabled vehicular networks. Although computation offloading and task scheduling have been widely investigated, it is rather challenging to decide the optimal offloading ratio and perform adaptive task scheduling in high-dynamic networks. Furthermore, the scheduling policy made by the network operator may be violated, since vehicular users are rational and selfish to maximize their own profits. By considering the incentive compatibility and individual rationality of vehicular users, we present POETS, an efficient partial computation offloading and adaptive task scheduling algorithm to maximize the overall system-wide profit. Specially, a two-sided matching algorithm is first proposed to derive the optimal transmission scheduling discipline. After that, the offloading ratio of vehicular users can be obtained through convex optimization, without any information of other users. Furthermore, a non-cooperative game is constructed to derive the payoff of vehicular users that can reach the equilibrium between users and the network operator. Theoretical analyses and performance evaluations based on real-world traces of taxies demonstrate the effectiveness of our proposed solution. [ABSTRACT FROM AUTHOR]

Dynamically adaptive streaming over HTTP requires a large-scale server to transcode various bitrate versions in which different preset parameters can be used to provide different video qualities at each resolution. When transcoding servers contain a heterogeneous mix of CPUs and GPUs, the task scheduler must choose a processor and preset parameter for each transcoding task to meet the transcoding deadlines while achieving the best possible video quality. We apply regression analysis to sample variable-bit-rate videos to provide accurate (mean absolute percentage error values from 1.3% to 13.9%) model for predicting bitrate, transcoding time and video quality at each resolution on different processors. We build this into a greedy allocation and scheduling algorithm which first satisfies deadlines with low video quality, and then redistributes the workload to improve that quality while continuing to meet the deadlines. This scheme was both simulated and implemented on a testbed server. It satisfies all deadlines while outperforming standard algorithms by between 3.12% and 15.59% in terms of popularity-weighted video quality divided by bitrate. [ABSTRACT FROM AUTHOR]

With the recent advancement of vehicular ad-hoc networks (VANETs) or the internet of vehicles (IoVs), vehicles are getting more powerful and generating huge amount of traffic data, including computation-intensive and delay-sensitive applications in the vehicular edge computing (VEC) networks, which are difficult to be processed by an individual vehicular node. These resource-demanding tasks can be transferred to another vehicular node with idle computing resources for processing. Due to high mobility and limited resources of vehicular nodes, it is challenging to execute lengthy computation-intensive tasks until completion within the delay constraint. There is a need to provide an efficient task offloading strategies to support these applications. In this paper, an efficient distributed task offloading scheme is proposed to select nearby vehicles with idle computing resources, to process the tasks in parallel by considering some vital metrics, including link reliability, distance, available computing resources, and relative velocity. In order to complete the lengthy computation-intensive tasks in vehicular edge computing networks, a task is divided into several subtasks before offloading. The performance of the proposed scheme is evaluated in several VEC network conditions. Results show that the proposed computation task offloading scheme achieves better performance in latency, throughput, resource utilization and packet delivery ratio than the existing schemes. [ABSTRACT FROM AUTHOR]

Mobile computing network is envisioned as a powerful framework to support the growing computation-intensive applications in the era of the Internet of Things (IoT). In this paper, we exploit the potential of a multi-layer network via a two-dimensional (2-D) task offloading scheme, which enables horizontal cooperations among the edge nodes. To minimize the average task offloading delay for all the mobile users, we formulate a mixed non-linear programming (MINLP) by jointly optimizing the 2-D offloading decisions and communication/computational resource allocation. To address this very challenging problem, we exploit the unique algorithmic structure of the optimal branch-and-bound (B&B) algorithm, and propose a novel Gaussian process imitation learning (GPIL) method to learn how to discover the shortcut for node searching in the B&B enumeration tree and significantly accelerate the B&B algorithm. When the network key parameters change, we further propose a novel recursive GPIL (RGPIL) method to agilely adapt to the new scenario with a fast policy update, where the new posterior distribution can be recursively updated based on a few new training data. Our simulation results show that the proposed method can achieve a near optimal solution with a significantly reduced complexity (e.g., a reduction of 98.7% in the number of searched nodes for a typical case). On this basis, the advantage of 2-D offloading scheme over the conventional schemes is also verified. [ABSTRACT FROM AUTHOR]

Mobile edge computing (MEC) has recently emerged as an effective paradigm to enhance the computing capability of capability-limited mobile devices (MDs). In this article, we consider an MEC system consisting of a number of MEC servers and one MD which generates a series of tasks characterized by their dependency relationships. We study computation scheduling and offloading problem of the tasks. To improve the task processing performance, we first propose a parallel transmission and execution (PTE) scheme, based on which we design a computation scheduling and offloading algorithm. Considering the fairness among tasks in terms of task transmission and execution time, we formulate the computation scheduling and offloading problem as a constrained worst-case latency optimization problem which minimizes the maximum completion time of all the tasks. As the original optimization problem cannot be solved conveniently, we first categorize the tasks into high priority tasks (HPTs), medium priority tasks (MPTs) and low priority tasks (LPTs) based on their task execution status and causal relationship. Then, a dynamic priority-based computation scheduling and offloading algorithm is proposed, which designs computation scheduling and offloading strategy for dynamically-changed HPTs and MPTs, respectively. In particular, for HPTs, a multiple knapsack-based heuristic algorithm is proposed, and a task weight and data size-based computation scheduling and offloading algorithm is further proposed for MPTs. Numerical results demonstrate the effectiveness of the proposed scheme. [ABSTRACT FROM AUTHOR]

Distributed stochastic optimization has important applications in the practical implementation of machine learning and signal processing setup by providing means to allow interconnected network of processors to work towards the optimization of a global objective with intermittent communication. Existing works on distributed stochastic optimization predominantly assume all the processors storing related data to perform updates for the optimization task in each iteration. However, such optimization processes are typically executed at shared computing/data centers along with other concurrent tasks. Therefore, it is necessary to develop efficient optimization methods that possess the flexibility to share the computing resources with other ongoing tasks. In this work, we propose a new first-order framework that allows this flexibility through a probabilistic computing resource allocation strategy while guaranteeing the satisfactory performance of distributed stochastic optimization. Our results, both analytical and numerical, show that by controlling a flexibility parameter, our suite of algorithms (designed for various scenarios) can achieve the lower computation and communication costs of distributed stochastic optimization than their inflexible counterparts. This framework also enables the fair sharing of the common resources with other concurrent tasks being processed by the processing network. [ABSTRACT FROM AUTHOR]

With the explosive growth in the scale of cloud computing infrastructures, reliability and energy efficiency have become important concerns considering the great complexity of cloud data centers. There is an urgent need for efficient task assignment that can dispatch tasks to appropriate cloud data center servers, which is critical to achieve reliability and energy efficiency in current cloud data centers. Most of the research on task assignment focuses on only one of the objectives of reliability and energy efficiency, while the two objectives are intrinsically conflicting with each other. In this paper, we deal with the problem of task assignment in data centers, with the objective of minimizing the energy consumption while providing failure tolerance to task execution failure. We propose a reliability-aware and energy-efficient task replica assignment algorithm based on running task replicas at a low speed and enabling multiple task replicas to share the same server resources. Each task in a job processed by the cloud computing platform has two instances: main task and task replica (shadow). Each main task runs on an individual server, and the task replica associated with the main task is assigned on a different server. The main tasks run at the full server speed, while the task replicas run at a lower rate than the main tasks. The task replicas can be mapped onto dedicated backup servers or be assigned to the servers on which the main tasks are running. Multiple task replicas can share the same server resources to reduce the number of servers required. We conduct experiments through simulations. Experimental results demonstrate that the proposed algorithm can effectively reduce the energy consumption, while achieving a good balance between the number of servers used and job completion time. [ABSTRACT FROM AUTHOR]

Mobile edge computing (MEC) is a promising solution to support emerging delay-sensitive mobile applications, such as self-driving, augment/virtual reality, and various Internet of Things (IoT) applications. By deploying MEC servers at network edge, e.g., close to cellular base stations (BSs), the computational tasks generated by these applications can be offloaded to edge nodes (ENs) and be quickly executed there. At the same time, with the projected large number of IoT devices, the communication and computational resources allocated to each user can be quite limited, making it challenging to provide low-latency MEC services. In this paper, we investigate the problem of task partitioning and user association in an MEC system, aiming to minimize the average latency of all users. We assume that each task can be partitioned into multiple subtasks that can be executed on local devices (e.g., vehicles), MEC servers, and/or cloud servers; each user can be associated with one of the nearby ENs. The subtasks can be independent of or dependent on each other. For each case, we formulate the joint optimization of task partitioning ratios and user association as a mixed integer programming problem. Each problem is solved by decomposing it into two subproblems. The lower-level subproblem is task partitioning under a given user association, which can be solved optimally. The higher-level subproblem is user association, we propose a dual decomposition-based approach and a matching-based approach to derive near-optimal solutions. Simulation results show that compared to benchmark schemes, the proposed schemes reduce the average latency by about 50% and 40% for the cases of independent and dependent subtasks, respectively. [ABSTRACT FROM AUTHOR]

In vehicular crowdsourcing, task requesters rely on a server to distribute spatial crowdsourcing tasks to on-road vehicular workers based on interests and locations. To protect the privacy of the interests and locations, both requesters and workers prefer to encrypt the information before uploading them to the server. However, such an encryption-before-outsourcing paradigm makes it a challenging issue to conduct the task matching. In this paper, we propose a Privacy-Preserving Task Matching (PPTM) with threshold similarity search via vehicular crowdsourcing. We first propose an interest-based PPTM by transforming vehicular workers’ interests into binary vectors. By using Symmetric-key Threshold Predicate Encryption (STPE) and proxy re-encryption, PPTM achieves privacy-preserving multi-keyword task matching with Jaccard similarity search in multi-worker multi-requester setting. Furthermore, by comparing the Euclidean distances between workers and requesters against a pre-defined threshold, PPTM preserves the location privacy of workers and requesters that only reveals the comparison results to the crowdsourcing server. The security analysis and extensive experiments demonstrate that PPTM protects the confidentiality of locations and interests of requesters and workers while achieving the efficient task matching. [ABSTRACT FROM AUTHOR]

Peer-to-Peer (P2P) resource sharing promotes local resource-hungry task offloading to other mobile devices and balances the resource consumption between mobile devices. Most of existing P2P task offloading systems aims to solve the resource sharing between one pair exclusively without considering the cost of resource supply and the strategic behaviors of mobile users. In this paper, we propose two user models for the P2P task offloading system: honest user model and strategy user model. For the honest user model, we formulate the resource allocation maximization problem with latency and energy consumption constraints as an Integer Linear Programming. We show that the solution for honest user model can output 189% resource transactions of that for the strategic users. For the strategy user model, we propose a double auction-based P2P task offloading system, and design a truthful multi-resource transaction mechanism to maximize the number of resource transactions. We first group the mobile users based on the connected components to improve the efficiency of double auction. Then we utilize the McAfee Double Auction to price the resource transactions. Finally, we split each winning mobile user of double auction into multiple virtual mobile users, and use the matching approach to calculate the resource allocation. Through both rigorous theoretical analysis and extensive simulations, we demonstrate that the designed multi-resource transaction mechanism satisfies the desirable properties of computational efficiency, individual rationality, budget balance, truthfulness for resource request/supply, and general truthfulness for bid/ask price. [ABSTRACT FROM AUTHOR]

In the context of Multi-access Edge Computing (MEC), the task sharing mechanism among edge servers is an activity of vital importance for speeding up the computing process and thereby improving the user experience. The distributed resources in the form of edge servers are expected to collaborate with each other in order to boost overall performance of a MEC system. However, there are many challenges to adopt global collaboration among the edge computing server entities and two of them are especially significant: ensuring trust among the servers and developing a unified scheme to enable real-time collaboration and task sharing. In this article, a blockchain framework is proposed to provide a trusted collaboration mechanism among edge servers in a MEC environment. In particular, a permissioned blockchain scheme is investigated to support a trusted design that also provides incentives for collaboration. Finally, Caliper tool and Hyperledger Fabric benchmarks are used to conduct an experimental evaluation of the proposed blockchain scheme embebbed in a MEC framework. [ABSTRACT FROM AUTHOR]

In this paper, sub-channel scheduling, task assignment and power allocation are investigated for orthogonal multiple access (OMA)-based and non-orthogonal multiple access (NOMA)-based mobile edge computing (MEC) systems. Based on different channel conditions and computational capacities, computational tasks are partially offloaded to the MEC server via OMA or NOMA protocols. In order to minimize the total energy consumption, an optimization problem under the task execution latency constraint is formulated and divided into two sub-problems. By utilizing matching theory, the formulated sub-channel allocation problem is solved by a proposed low-complexity algorithm, where the joint optimization of task assignment and power allocation is performed at each iteration. Based on the delay constraint, some insights are obtained, and the closed-form solutions of task assignment coefficients and transmit power are derived. Furthermore, the offloading strategy in both OMA and NOMA schemes is analyzed, which shows that the optimal task assignment coefficient is decided by the energy consumption efficiency (ECE). Simulation results indicate that: i) the proposed sub-channel allocation algorithm and derived closed-form solutions can significantly improve the MEC system in terms of the energy consumption; ii) the provided offloading strategy can be dynamically and efficiently employed with different channel conditions and computational capacities. [ABSTRACT FROM AUTHOR]

In the Internet of Health Things (IoHT)-based e-Health paradigm, a large number of computational-intensive tasks have to be offloaded from resource-limited IoHT devices to proximal powerful edge servers to reduce latency and improve energy efficiency. However, the lack of global state information (GSI), the adversarial competition among multiple IoHT devices, and the ultra reliable and low latency communication (URLLC) constraints have imposed new challenges for task offloading optimization. In this article, we formulate the task offloading problem as an adversarial multi-armed bandit (MAB) problem. In addition to the average-based performance metrics, bound violation probability, occurrence probability of extreme events, and statistical properties of excess values are employed to characterize URLLC constraints. Then, we propose a URLLC-aware Task Offloading scheme based on the exponential-weight algorithm for exploration and exploitation (EXP3) named UTO-EXP3. URLLC awareness is achieved by dynamically balancing the URLLC constraint deficits and energy consumption through online learning. We provide a rigorous theoretical analysis to show that guaranteed performance with a bounded deviation can be achieved by UTO-EXP3 based on only local information. Finally, the effectiveness and reliability of UTO-EXP3 are validated through simulation results. [ABSTRACT FROM AUTHOR]

Autonomous driving has the potential to make transportation systems safer, greener and more efficient. To realize autonomous driving, a great deal of in-car cutting-edge applications such as augmented reality, dynamic path planning and cognitive driving systems are required, which need significant computational resources and near realtime response. To cope with this new paradigm, vehicular fog computing (VFC) has emerged recently, which migrates the computing from congested base stations (or cloud servers) to nearby vehicles with under-utilized computational resources. in VFC, the designs of server recruitment and task offloading strategies under information asymmetry and uncertainty pose new technical challenges. in this article, we propose a two-stage VFC framework to address these challenges. The framework consists of a contract theory based vehicular computational resource management mechanism, and a matching-learning based task offloading mechanism. Simulation results demonstrate that the proposed framework can boost the performance of VFC in terms of resource utilization efficiency and task offloading delay. [ABSTRACT FROM AUTHOR]

In this paper, we consider parallel and sequential task offloading to multiple mobile edge computing servers. The task consists of a set of inter-dependent sub-tasks, which are scheduled to servers to minimize both offloading latency and failure probability. Two algorithms are proposed to solve the scheduling problem, which are based on genetic algorithm and conflict graph models, respectively. Simulation results show that these algorithms provide performance close to the optimal solution, which is obtained through exhaustive search. Furthermore, although parallel offloading uses orthogonal channels, results demonstrate that the sequential offloading yields a reduced offloading failure probability when compared to the parallel offloading. On the other hand, parallel offloading provides less latency. However, as the dependency among sub-tasks increases, the latency gap between parallel and sequential schemes decreases. [ABSTRACT FROM AUTHOR]

With the booming development of IoT and massive smart MDs springing up in daily life, the conflict between resource-hungry IoT applications and resource-constrained MDs becomes increasingly prominent. To cope with compute-intensive applications and big data, MCC combining AI was adopted as a workable solution. Nevertheless, considering MCC's long transmission latency and the ultra-low latency requirements of most IoT applications, traditional MCC combining AI is not applicable any more in the era of IoT. Migrating cloud computing capabilities to the edge, and integrating AI with it, are envisioned to be a promising paradigm, which gives rise to the so-called edge intelligence. As a pivotal technique in edge computing, task offloading can effectively improve the MDs' computation and energy efficiency. However, existing research on task offloading mostly focused on fixed scenarios and cannot deal with varying situations, where user privacy protection was neglected either. Toward this end, we introduce machine learning into task offloading at the edge, and design an intelligent task offloading scheme. Extensive numerical results demonstrate that our proposed scheme cannot only have good adaptability and security, but also achieve high prediction accuracy and low processing delay, compared to traditional offloading schemes. [ABSTRACT FROM AUTHOR]

Low latency/delay is one of the most critical requirements for the application of vehicular networks. However, frequent real-time information update caused by vehicles high mobility is liable to aggravate the delay. Meanwhile, the task migration between different vehicular edge computing (VEC) servers results in an amount of delay if the computing cannot be completed before the vehicle moves out of the coverage of the current VEC server. In this paper, the problem is concluded as when and to whom to offload the task for VEC, which is formulated as a finite horizon Markov decision process (MDP) to minimize the delay with respect to the communication, computing, handover and migration. Through characterizing the time-space correlation of vehicles mobility, the curse of dimensionality problem in MDP is resolved. Meanwhile, a general expression of the transition probabilities is derived. On this basis, the specific results of highway, 2-D street and real-data scenarios are provided as well. For practical implementation considerations, the transition probabilities are commonly uncertain primarily due to random driver behavior, inaccurate sample data and complex path environment. Under this uncertain environment,a robust time-aware MDP-based task offloading algorithm (RTMDP) is proposed, which has been proved to perform well even under the high uncertain transition probabilities by simulation results. [ABSTRACT FROM AUTHOR]

Contemporary resource constraint mobile devices face great challenges to execute computationally expensive tasks of low-latency Tactile Internet applications in a time-efficient manner. The so-called HART has emerged as a promising platform to address the burden of resource-limited mobile devices by allowing collaboration among humans, cloud agents, and robots for task execution. To facilitate a better task execution performance, dynamic task offloading onto suitable actors represents an important research problem, taking time-varying task, robot, and cloud agent properties into account. To tackle this open problem, this article first discusses characteristics, applications, and challenges of HART tasks. Next, this article investigates the problem of whether and where a HART task should be offloaded and presents a COTA strategy by taking task offloading gain and communication overhead into account. For HART-centric task offloading, this article leverages central cloud, mobile edge computing (MEC) servers, and multi-robot FiWi networks. Finally, numerical results are presented to examine the performance of the proposed COTA scheme over FiWi based Tactile Internet infrastructures. [ABSTRACT FROM AUTHOR]

Vehicular fog computing has emerged as a cost-efficient solution for task processing in vehicular networks. However, how to realize effective server recruitment and reliable task offloading under information asymmetry and uncertainty remains a critical challenge. In this paper, we adopt a two-stage task offloading framework to address this challenge. First, we propose a convex–concave-procedure-based contract optimization algorithm for server recruitment, which aims to maximize the expected utility of the operator with asymmetric information. Then, a low-complexity and stable task offloading mechanism is proposed to minimize the total network delay based on the pricing-based matching. Furthermore, we extend the work to the scenario of information uncertainty and develop a matching-learning-based task offloading mechanism, which takes both occurrence awareness and conflict awareness into consideration. Simulation results demonstrate that the proposed algorithm can effectively motivate resource sharing and guarantee bounded deviation from the optimal performance without the global information. [ABSTRACT FROM AUTHOR]

There is a growing interest in the development of in-network dispersed computing paradigms that leverage the computing capabilities of heterogeneous resources dispersed across the network for processing a massive amount of data collected at the edge of the network. We consider the problem of task scheduling for such networks, in a dynamic setting in which arriving computation jobs are modeled as chains, with nodes representing tasks, and edges representing precedence constraints among tasks. In our proposed model, motivated by significant communication costs in dispersed computing environments, the communication times are taken into account. More specifically, we consider a network where servers can serve all task types, and sending the outputs of processed tasks from one server to another server results in some communication delay. We first characterize the capacity region of the network, then propose a novel virtual queueing network encoding the state of the network. Finally, we propose a Max-Weight type scheduling policy, and considering the stochastic network in the fluid limit, we use a Lyapunov argument to show that the policy is throughput-optimal. Beyond the model of chains, we extend the scheduling problem to the model of the directed acyclic graph (DAG) which imposes a new challenge, namely logic dependency difficulty, requiring the data of processed parents tasks to be sent to the same server for processing the child task. We propose a virtual queueing network for DAG scheduling over broadcast networks, where servers always broadcast the data of processed tasks to other servers, and prove that Max-Weight policy is throughput-optimal. [ABSTRACT FROM AUTHOR]

In this paper, we consider the problem of task offloading in a software-defined access network, where IoT devices are connected to fog computing nodes by multi-hop IoT access-points (APs). The proposed scheme considers the following aspects in a fog-computing-based IoT architecture: 1) optimal decision on local or remote task computation; 2) optimal fog node selection; and 3) optimal path selection for offloading. Accordingly, we formulate the multi-hop task offloading problem as an integer linear program (ILP). Since the feasible set is non-convex, we propose a greedy-heuristic-based approach to efficiently solve the problem. The greedy solution takes into account delay, energy consumption, multi-hop paths, and dynamic network conditions, such as link utilization and SDN rule-capacity. Experimental results show that the proposed scheme is capable of reducing the average delay and energy consumption by 12% and 21%, respectively, compared with the state of the art. [ABSTRACT FROM AUTHOR]