区块链赋能的边缘异构计算系统中资源调度研究  

Resource management in blockchain-enabled heterogeneous edge computing system

作　　者：张平[1] 李世林 刘宜明 秦晓琦 许晓东[2] ZHANG Ping;LI Shilin;LIU Yiming;QIN Xiaoqi;XU Xiaodong(State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China;National Engineering Laboratory for Mobile Network Technologies,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]北京邮电大学网络与交换技术国家重点实验室,北京100876 [2]北京邮电大学移动网络技术国家工程实验室,北京100876

出　　处：《通信学报》2020年第10期1-14,共14页Journal on Communications

基　　金：国家重点研发计划基金资助项目(No.2018YFB1800800);北京市自然科学基金资助项目(No.19L2033)。

摘　　要：在区块链赋能的移动边缘计算(BMEC)系统中,针对各类新型计算任务并行性需求的差异,提出了一种基于异构计算的BMEC系统模型,通过调用异构计算架构中并行计算能力不同的处理器,实现区块链业务与用户业务的高效处理。通过综合考虑异构处理器调度、计算资源分配以及带宽资源分配,将通信及计算资源受限下的系统效用最大化问题建模为混合整数非线性问题。为了快速求解该问题,将所提模型进一步解耦为业务驱动的异构处理器调度问题和资源联合分配问题,并提出了基于拉格朗日对偶理论的联合优化算法。仿真结果表明,所提算法可以有效提升BMEC系统的系统效用。In blockchain-enabled mobile edge computing(BMEC)systems,a new class of blockchain application related computation tasks was introduced to the system.Due to the differences of parallelism among computation tasks,heterogenous computing framework was introduced to suitably split various computation tasks on processors with vastly different processing power to achieve efficient task execution.Under the limited computation and communication resources,a system-wide utility maximization problem by jointly considering heterogeneous processor scheduling,computation and bandwidth resource allocation was formulated as a mixed-integer nonlinear programming problem.To solve the problem efficiently,the formulated problem was transformed into two sub-problems,namely application-driven heterogeneous processor scheduling and joint resource allocation,and a Lagrange-dual based algorithm was proposed.Simulation results show that the proposed scheme can effectively improve the system-wide utility of the BMEC system.

关 键 词：移动边缘计算 区块链 异构计算 计算卸载 

分 类 号：TN92[电子电信—通信与信息系统]