Idle Computational Resources Research Articles

Researching the CNN Collaborative Inference Mechanism for Heterogeneous Edge Devices.

Convolutional Neural Networks (CNNs) have been widely applied in various edge computing devices based on intelligent sensors. However, due to the high computational demands of CNN tasks, the limited computing resources of edge intelligent terminal devices, and significant architectural differences among these devices, it is challenging for edge devices to independently execute inference tasks locally. Collaborative inference among edge terminal devices can effectively utilize idle computing and storage resources and optimize latency characteristics, thus significantly addressing the challenges posed by the computational intensity of CNNs. This paper targets efficient collaborative execution of CNN inference tasks among heterogeneous and resource-constrained edge terminal devices. We propose a pre-partitioning deployment method for CNNs based on critical operator layers, and optimize the system bottleneck latency during pipeline parallelism using data compression, queuing, and "micro-shifting" techniques. Experimental results demonstrate that our method achieves significant acceleration in CNN inference within heterogeneous environments, improving performance by 71.6% compared to existing popular frameworks.

Mobility-aware Vehicular Cloud formation mechanism for Vehicular Edge Computing environments

Rapid advancements in vehicular technology and increased vehicle modernization have led to the emergence of intelligent and interconnected entities. As a result, the Vehicular Edge Computing (VEC) paradigm has gained prominence. This paradigm enables the provision of cloud computing services close to vehicular users by utilizing the idle computational resources of vehicles to execute tasks that require computing power beyond what is available locally. Aggregating these computational resources in the vehicular context is known as Vehicular Cloud (VCloud) formation. However, leveraging and aggregating these resources poses several challenges due to the dynamic nature of the vehicular environment. One of the main challenges is the efficient selection of vehicles to assume management roles in the distribution of computational power within the group, often referred to as leading vehicles. This research presents a mobility-aware mechanism called PREDATOR to enhance the VCloud formation process. In this mechanism, the Roadside Unit (RSU) provides vehicular mobility predictions, enabling the selection of the most stable vehicles within the RSU coverage area to assume leadership roles in the VCloud. In this context, vehicle stability is associated with a vehicle’s time within the RSU coverage area, known as dwell time. PREDATOR employs a microscopic perspective to select vehicles with the longest dwell time in the VCloud, allowing for efficient management of computational resource utilization. Simulation results have demonstrated that PREDATOR not only increases the VCloud lifetime but also minimizes leader changes, reduces network message exchange, mitigates packet collisions, and facilitates the effective utilization of aggregated vehicular resources compared to state-of-the-art approaches.

Online Distributed Learning-Based Load-Aware Heterogeneous Vehicular Edge Computing

Vehicular edge computing is an emerging enabler in strengthening driving efficiency and traffic safety. However, both performance bottlenecks and low resource efficiency of heterogeneous edge servers arise concurrently because of the inhomogeneous load distribution among the servers. Further, the unsaturated infrastructure coverage situation can deteriorate the concurrent issues. Although transmitting raw task data with large sizes among heterogeneous edge servers can relieve the concurrent issues, it distinctly degrades the core network’s efficiency, especially during rush hours. Meanwhile, it cannot settle the unsaturated coverage situation. To relieve the concurrent issues without degrading the core network’s efficiency, we introduce an aerial relay station that can flexibly relay vehicular tasks to nearby heterogeneous edge servers. The long-term task scheduling problem without any prior environment knowledge for the considered vehicular edge system is crucial but still up in the air. We formulate the system latency minimization problem as a partially observable stochastic game. Then a model-free multi-agent reinforcement learning algorithm is developed to search the real-time load-aware scheduling policy. Besides, we design a practical factor named offloading latency gain to assist the training process of the learning algorithm. Simulation experiments show that our proposed algorithm can better exploit idle computation resources of heterogeneous edge infrastructures and significantly reduce the average system latency up to 15-20% over existing algorithms.

In-Network Pooling: Contribution-Aware Allocation Optimization for Computing Power Network in B5G/6G Era

The coming B5G/6 G will bring a huge amount of challenging applications with zettabytes of information. Computing power network (CPN) can provide a promising solution by accelerating the proliferation of computing power from a set of data centers to a multitude of network edges. However, in dealing with resource-hungry and real-time applications, most of the existing research doesn't make good use of idle computing and caching resources, and is barely possible to evaluate the contribution of the individual providing the resource. Therefore, we propose the in-network pooling framework, derived from a novel modified deep reinforcement learning (DRL) scheme, in which the dynamic resource pool (RP) is first modeled to make full use of the idle network resources, then the jointly computing and caching problem is formulated as the maximization of long-term system utility. Finally, Attention-based Proximal Policy Optimization (APPO) is employed to solve the problem. Particularly, the integrated attention mechanism reveals the evaluation of the different RPs' contributions to the learning process. Experimental results also show the priorities of the proposed algorithm and outperform the other existing algorithms.

Privacy preserving federated learning for full heterogeneity

Federated learning is a novel distribute machine learning paradigm to support cooperative model training among multiple participant clients, where each client keeps its private data locally to protect its data privacy. However, in practical application domains, Federated learning still meets several heterogeneous challenges such data heterogeneity, model heterogeneity, and computation heterogeneity, significantly decreasing its global model performance. To the best of our knowledge, existing solutions only focus on one or two challenges in their heterogeneous settings. In this paper, to address the above challenges simultaneously, we present a novel solution called Full Heterogeneous Federated Learning (FHFL). Firstly, we propose a synthetic data generation approach to mitigate the Non-IID data heterogeneity problem. Secondly, we use knowledge distillation to learn from heterogeneous models of participant clients for model aggregation in the central server. Finally, we produce an opportunistic computation schedule strategy to exploit the idle computation resources for fast-computing clients. Experiment results on different datasets show that our FHFL method can achieve an excellent model training performance. We believe it will serve as a pioneer work for distributed model training among heterogeneous clients in Federated learning.

UAV-Aided Computation Offloading in Mobile-Edge Computing Networks: A Stackelberg Game Approach

Unmanned aerial vehicles (UAVs) are considered as a promising method to provide additional computation capability and wide coverage for mobile users (MUs), especially when MUs are not within the communication range of the infrastructure. In this article, a UAV-aided mobile-edge computing (MEC) network, including one UAV-MEC server, one BS-MEC server, and several MUs, is investigated for computation offloading, in which the edge service provider (ESP) manages two kinds of servers. It is considered that MUs have a large number of computation tasks to conduct, while the ESP has idle computational resources. MUs can choose to offload their tasks to the ESP to reduce their pressure and cost, and the ESP can make a profit by selling computational resources. The interaction among the ESP and MUs is modeled as a Stackelberg game, and both the ESP and MUs want to maximize their utility. The proposed game is analyzed by using the backward induction method, and it is proved that a unique Nash equilibrium can be achieved in the game. Then, a gradient-based dynamic iterative search algorithm (GDISA) is proposed to get the approximate optimal solution. Finally, the effectiveness of GDISA is verified by extensive simulations, and the results show that GDISA performs better than other benchmark methods under different scenarios.

Vehicular Edge-Computing Framework for Making Use of Parking and Charging Electric Vehicles

In big cities, there are more and more parking lots and charging piles for electric vehicles, and the resources of parking and charging vehicles can be aggregated to provide strong computing power for vehicular edge computing (VEC). In this paper, we propose a VEC framework that uses charging vehicles in parking lots to assist edge servers in processing computational tasks, and an edge crowdsourcing platform (ECP) is designed to manage and integrate the idle computation resources of electric vehicles in parking lots to provide computation services for requesting vehicles. Based on game theory, we first model the interactions among the edge server, the ECP and the requesting vehicles as a Stackelberg game, and theoretically prove the existence of a Nash equilibrium for this Stackelberg game. Then, a genetic algorithm-based game-strategy solving algorithm is proposed to find the optimal strategy for the edge server and ECP. The simulation results demonstrate that the performance of our proposed solution is better than other traditional solutions. Compared with the solution without ECP, our solution can increase the utilities of the edge server and the requesting vehicle by 13.3% and 10.99%, respectively.

DRL-Based Resource Allocation for Computation Offloading in IoV Networks

Due to the dynamic nature of a vehicular fog computing environment, efficient real-time resource allocation in an Internet of Vehicles (IoV) network without affecting the quality of service of any of the onboard vehicles can be challenging. This article proposes a priority-sensitive task offloading and resource allocation scheme in an IoV network, where vehicles periodically exchange beacon messages to inquire about available services and other important information necessary for making the offloading decisions. In the proposed methodology, the vehicles are stimulated to share their idle computation resources with the task vehicles, whereby a deep reinforcement learning algorithm based on soft actor–critic is designed to classify the tasks based on priority and computation size of each task for optimally allocating the power. Furthermore, we also design deep deterministic policy gradient (DDPG) and twin delayed DDPG (TD3) algorithms for the considered framework. In particular, the algorithms work toward achieving the optimal policy for task offloading by maximizing the mean utility of the considered network. Extensive numerical results under different network conditions, along with comparison among the three algorithms, are presented to validate the feasibility of distributed reinforcement learning for task offloading in future IoV networks.

Stackelberg-Game-Based Computation Offloading Method in Cloud–Edge Computing Networks

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

Joint Task Partition and Resource Allocation for Multiuser Cooperative Mobile Edge Computing

Exploiting the idle computation resources distributed at wireless devices (WDs) can enhance the mobile edge computing (MEC) computation performance. This paper studies a multiuser cooperative computing system consisting of one local user and multiple helpers, in which the user solicits multiple nearby WDs acting as helpers for cooperative computing. We design an efficient orthogonal frequency-division multiple access- (OFDMA-) aided three-phase transmission protocol, under which the user’s computation-intensive tasks can be executed in parallel by local computing and offloading. Under this setup, we study the energy consumption minimization problem by optimizing the user’s task partition, jointly with the communication and computation resources allocation for task offloading and results downloading, subject to the user’s computation latency constraint. For the nonconvex problem, we first transform the original problem into a convex one and then use the Lagrange duality method to obtain the globally optimal solution. Compared with other benchmark schemes, numerical results validate the effectiveness of the proposed joint task partition and resource allocation (JTPRA) scheme.

A cluster-based cooperative computation offloading scheme for C-V2X networks

The explosive growth of computation-intensive and delay-sensitive applications in vehicular edge network demands high computation resources and energy. With the cellular-vehicle to everything (C-V2X) technology, these applications can be transferred to another node for processing. However, offloading all these tasks to the mobile edge computing (MEC) server becomes a bottleneck. A cluster-based cooperative task offloading scheme for C-V2X networks is proposed. The proposed scheme explores the collaboration between the MEC servers and vehicles with idle computation resources known as the vehicular edge computing (VEC) servers. Vehicles can utilize the direct sidelink vehicle to vehicle (V2V) communication for task offloading to the VEC server and the cellular vehicle to network (V2N) communication for task offloading to the MEC server. In the proposed scheme, vehicles are grouped to form clusters in order to enhance link reliability between nodes for efficient task offloading. We then formulate an optimization problem to minimize energy consumption and task offloading latency, which is NP-hard. Accordingly, an improved matching algorithm is proposed to solve this problem by optimizing offloading decisions and resource allocation. According to our results, the proposed scheme reduces offloading latency and energy consumption compared to other existing schemes.

Joint task offloading and resource allocation in vehicle-assisted multi-access edge computing

Multi-access Edge Computing (MEC) has significant advantages in improving resource efficiency of Internet of Things (IoT) and 5G networks, however its limited resources cannot meet the demand of data communication and computation capability during off-peak time. Incentivizing intelligent vehicles with idle computation resources as vehicle edge nodes (VENs) to provide computation offloading for nearby user equipments (UEs) is an appealing idea. Thus, we propose a vehicle-assisted MEC (VMEC) paradigm, where tasks can be offloaded to MEC server and VENs. In this paper, we first establish a differentiated pricing model based on different states of resources and a dynamic incentive model according to the demands of UEs. Then, we formulate a Stackelberg game between UEs and MEC service provider (MEC SP) to obtain the optimal offloading strategy and pricing scheme. A gradient-based resource allocation iteration algorithm (GRAIA) is designed for the Nash equilibrium solution. Finally, considering the matching between UEs and vehicles, we present a reverse auction-based task scheduling algorithm (RATSA) to choose VENs. The simulation results demonstrate that the proposed scheme can achieve significant performance improvement and is superior to the existing schemes in improving system utility.

Task Offloading Optimization for UAV-Assisted Fog-Enabled Internet of Things Networks

Recently, unmanned aerial vehicles (UAVs) have been considered as an efficient way to provide enhanced coverage or relaying services to Internet of Things devices (IDs) in wireless systems with limited or no infrastructure. In this article, a UAVs-assisted fog-enabled Internet of Things (IoT) network is studied, in which moving UAVs are equipped with computing capabilities to offer task offloading opportunities to IDs. Besides, there are two types of IDs, namely, requested-IDs (R-IDs), which has task offloading requirement, and free-IDs (F-IDs), which could offload tasks for R-IDs with idle computation resources. Two offloading links are considered: 1) the device-to-device (D2D) link and 2) the ground-to-air (G2A) link, which are responsible for both the uplink and downlink offloading procedure. To minimize the total network overhead, we jointly optimize the UAV trajectory, transmission power, and computation offload radios, while satisfying Quality-of-Service (QoS) requirements of R-IDs. The optimization problem is nonconvex, and the UAV-assisted task offloading optimization algorithm is proposed to obtain the local optimal solutions, which decomposes the original problem into two parallel subproblems and solved alternately. Finally, simulation results demonstrate that the proposed algorithm could achieve superior performance in terms of the network overhead compared with algorithms in the literature.

Virtual Edge: Exploring Computation Offloading in Collaborative Vehicular Edge Computing

Vehicular edge computing (VEC) has been a new paradigm to support computation-intensive and latency-sensitive services. However, the scarcity of computational resources is still a challenge. Making efficient use of sporadic idle computational resources on smart vehicles in the vicinity to extend the resource capability of each vehicle is an important research issue. In this paper, we propose Virtual Edge, which is an efficient scheme to utilize free computational resources of multiple vehicles as a virtual server to facilitate collaborative vehicular edge computing. We design a virtual edge formation algorithm that considers both the stability of virtual edge and the computational resources available at the vehicles constituting the virtual edge. The prediction of the link duration between vehicles reduces the number of computation offloading failures caused by unexpected link disconnections. Extensive simulations with realistic vehicle movements are conducted to show the advantage of the proposed scheme over existing baselines in terms of the completion ratio of computation offloading tasks and average task execution time.

NOMA-Aided Mobile Edge Computing via User Cooperation

Exploiting the idle computation resources of mobile devices in mobile edge computing (MEC) system can achieve both channel diversity and computing diversity as mobile devices can offload their computation tasks to nearby mobile devices in addition to MEC server embedded access point (AP). In this paper, we propose a non-orthogonal multiple-access (NOMA)-aided cooperative computing scheme in a basic three-node MEC system consisting of a user, a helper, and an AP. In particular, we assume that the user can simultaneously offload data to the helper and the AP using NOMA, while the helper can locally compute data and offload data to the AP at the same time. We study two optimization problems, energy consumption minimization and offloading data maximization, by joint communication and computation resource allocation of the user and helper. We find the optimal solutions for the two non-convex problems by some proper mathematical methods. Simulation results are presented to demonstrate the effectiveness of the proposed schemes. Some useful insights are provided for practical designs.

Intelligent Resource Allocation for Utility Optimization in RSU-Empowered Vehicular Network

Intelligent transportation system (ITS) has attracted extensive attention in both academia and industry for its potential benefits. For example, ITS is dedicated to convenient, economical and environmentally friendly service provisioning for the drivers and passengers in vehicles via advanced technologies including artificial intelligence (AI), knowledge mining, depth fusion, etc. Besides, several newly emerging computing paradigms revolved around ITS such as vehicular cloud and vehicular fog computing are proposed to fully exploit idle computing and communication resources within connected vehicles. As the number of vehicular applications is explosively increasing, it has posed great challenges to the limited capabilities of vehicle loaded computer systems and communication facility. Accordingly, more intelligent resource allocation strategies are needed for computationally intensive and time sensitive vehicular applications. In this paper we propose a road side unit (RSU) empowered vehicular network that consists of three hierarchical layers-vehicular cloud, RSU-enabled cloudlet, and central cloud, respectively. RSU is enhanced with edge servers such that it can intelligently respond to the resource requests in a real time fashion. To this end, an approximate but efficient resource allocation strategy is proposed that can intelligently optimize the utility value from the perspective of RSU-enabled cloudlet. Extensive experiments are carried out to evaluate the performance of the strategy. The results reveal that the proposed algorithm DbHA shows great advantages over other approaches such as the genetic algorithm (GA) and particle swarm optimization (PSO) in both respects (i.e., performance and response latency).

Artificial Intelligence-Empowered Edge of Vehicles: Architecture, Enabling Technologies, and Applications

With the proliferation of mobile devices and a wealth of rich application services, the Internet of vehicles (IoV) has struggled to handle computationally intensive and delay-sensitive computing tasks. To substantially reduce the latency and the energy consumption, application work is offloaded from a mobile device to a remote cloud or a nearby mobile edge cloud for processing. Compared with remote clouds, mobile edge clouds are located at the edge of the network. Therefore, mobile edge computing (MEC) has the advantages of effectively utilizing idle computing and storage resources at the edge of the network and reducing the network transmission delay. In addition, mobile devices are increasingly moving toward intelligence. To satisfy the service experience and service quality requirements of mobile users, the vehicle Internet is transforming into the intelligent vehicle Internet. Artificial intelligence (AI) technology can adapt to rapidly changing dynamic environments to provide multiple task requirements for resource allocation, computational task scheduling, and vehicle trajectory prediction. On this basis, combined with MEC technology and AI technology, computing and storage resources are placed on the edge of the network to provide real-time data processing while providing more efficient and intelligent services. This article introduces IoV from three aspects, namely, MEC, AI and the advantages of combining the two, and analyzes the corresponding architecture and implementation technology. The application of MEC and AI in IoV is analyzed and compared with current approaches. Finally, several promising future directions in the field of IoV are discussed.

Fast Switching With VM Migration in 5G Cloud Radio Access Networks

Cloud radio access network (C-RAN) is an important architecture for 5G networks. In C-RAN systems, baseband units (BBUs) are centralized in a BBU pool and remote radio heads (RRHs) are connected with the BBU pool through front-haul links. The BBU-RRH peer switching scheme is quite a key point in C-RAN especially for fast switching scenarios. A typical C-RAN system based on software defined radio (SDR) and virtual machine migration is introduced as the system model. Then we analyze the factors, including packet delay, and power consumption, that influence the performance of a C-RAN system. We conclude backlogs, idle computation resources, and switching costs could reflect the system performance adequately. They can evaluate radio resource usage, computation resource usage, and extra computation resource usage, respectively. Based on the evaluation model, the optimization algorithm is studied. In this paper, we propose a method to find the optimized load range for the BBUs. Taking switching quality into consideration, we propose a scheme that can adjust the BBU working load dynamically in an optimized range with less switching cost. The system performance variation under different switching intervals is also investigated. Finally, the effectiveness of the proposed optimization scheme is validated by simulation results.

Game Theory Based Opportunistic Computation Offloading in Cloud-Enabled IoV

With the growing popularity of the fifth-generation (5G) wireless systems and cloud-enabled Internet of Vehicles, vehicular cloud has been introduced as a novel mobile device computing mode, which enables vehicles to offload their computation-intensive tasks to neighbors. In this paper, we first present a 5G cloud-enabled scenario of vehicular cloud computing where a vehicular terminal works either as a service provider with idle computation resources or a requestor who has a computation-intensive task that can be executed either locally or offloaded to nearby providers via opportunistic vehicle-to-vehicle communications. Then, we study the following issues: 1) how to determine the appropriate offloading rate of requestors; 2) how to select the most appropriate computation service provider; 3) how to identify the ideal pricing strategy for each service provider. We address the above-mentioned problems by developing a two-player Stackelberg-game-based opportunistic computation offloading scheme under situations involving complete and incomplete information that primarily considers task completion duration and service price. We simplify the former case into a common resource assignment problem with mathematical solutions. For the latter case, Stackelberg equilibriums of the offloading game are derived, and the corresponding existence conditions are concretely discussed. Finally, a Monte-Carlo simulation-based performance evaluation shows that the proposed methods can significantly reduce the task completion duration while ensuring the profit of service providers, thus achieving mutually satisfactory computation offloading decisions.

High-Resolution Rapid Refresh Model Data Analytics Derived on the Open Science Grid to Assist Wildland Fire Weather Assessment

AbstractTerabytes of weather data are generated every day by gridded model simulations and in situ and remotely sensed observations. With this accelerating accumulation of weather data, efficient computational solutions are needed to process, archive, and analyze the massive datasets. The Open Science Grid (OSG) is a consortium of computer resources around the United States that makes idle computer resources available for use by researchers in diverse scientific disciplines. The OSG is appropriate for high-throughput computing, that is, many parallel computational tasks. This work demonstrates how the OSG has been used to compute a large set of empirical cumulative distributions from hourly gridded analyses of the High-Resolution Rapid Refresh (HRRR) model run operationally by the Environmental Modeling Center of the National Centers for Environmental Prediction. These cumulative distributions derived from a 3-yr HRRR archive are computed for seven variables, over 1.9 million grid points, and each hour of the calendar year. The HRRR cumulative distributions are used to evaluate near-surface wind, temperature, and humidity conditions during two wildland fire episodes—the North Bay fires, a wildfire complex in Northern California during October 2017 that was the deadliest and costliest in California history, and the western Oklahoma wildfires during April 2018. The approach used here illustrates ways to discriminate between typical and atypical atmospheric conditions forecasted by the HRRR model. Such information may be useful for model developers and operational forecasters assigned to provide weather support for fire management personnel.