摘要
移动边缘计算环境下的计算卸载技术有助于解决移动终端在资源存储、计算性能等方面的不足。然而,在拥有大量计算资源的移动边缘计算环境中,边缘服务器、移动终端以及网络通信链路的不可靠性不可避免,而应用任务的执行失败对工作流任务调度将造成极大的影响。针对上述问题,首先借鉴社会学中的人际关系模型,同时考虑移动边缘计算环境下应用任务执行的特点,利用Bayes方法对移动终端、边缘服务器和云服务器的可信度分别进行评估,构建了移动边缘计算环境下各类计算资源之间的信任关系模型。其次,结合信任模型和基于多重计算卸载策略的时间开销计算方法,设计了评价计算卸载策略可靠性以及时间开销的适应度计算方法,提出了基于信任模型的可卸载边缘服务器选择算法。最后,结合工作流管理系统提出了移动边缘计算环境下基于信任模型的可靠多重计算卸载策略。通过仿真实验,证明了所提出的基于信任模型的可靠多重计算卸载策略算法能够以较小的时间开销为代价,有效提升应用任务的执行成功率。
Computation offloading technique is helpful for overcoming the shortage of mobile terminal in resource storage and computing performance.However,the unreliability of edge server,mobile termination and network link are inevitable in a mobile edge computing environment with large number of computing resources,and task execution failure will have a great impact on the performance of workflow management system.In view of the above problem,based on interpersonal trust model in sociology,a trust evaluation model was proposed by considering the computing resources past experience and the characteristics of mobile edge computing environment.On this basis,combined with trust evaluation model and multi-resources computation offloading strategy,the fitness computation method of task scheduling plan was designed to evaluate the reliability and time expenses.A trust based multi-resource computation offloading strategy task scheduling algorithm was presented to satisfy the tasks’trust requirement.The simulation experimental results showed that the proposed algorithms could effectively reduce the failure probability at the expense of a relatively small increase in the execution time.
作者
齐平
王福成
徐佳
李学俊
QI Ping;WANG Fucheng;XU Jia;LI Xuejun(Department of Mathematics and Computer Science,Tongling University,Tongling 244000,China;School of Computer Science and Technology,Anhui University,Hefei 230601,China)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2020年第6期1616-1627,共12页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(61672034,61300042)
安徽省高校自然科学基金重点项目(KJ2019A0704)。
关键词
移动边缘计算
资源分配
计算卸载
信任评估
工作流
mobile edge computing
resource allocation
computation offloading
trust evaluation
workflow