面向边缘协作的动态服务配置与迁移机制研究

Dynamic Service Configuration and Migration Mechanism for Edge Collaboration

随着物联网技术的发展,时延敏感型用户请求日益激增,边缘计算正在成为提供灵活可靠服务的有前途的范例。考虑到物联网设备有限的存储和计算资源与用户需求延迟之间的冲突,边缘计算激发了将用户请求卸载到网络边缘的边缘服务器的新一轮浪潮。用户请求任务卸载方案的制定对于边缘网络中物联网设备的执行时间和能耗是一个显着的挑战。针对该问题,提出了一种动态计算卸载策略,具体包括:①提出中间节点卸载策略。该节点具有足够的计算资源,并且能够通过动态最小化处理任务的时间和能量成本;②将任务服务配置节点选择问题建模为多维马尔可夫决策过程(MDP)空间,提出深度强化学习算法以减少MDP状态空间并实现快速决策。实验结果表明,所提策略可以显著改善任务请求处理的时延,并且可以在一定程度上减少能量消耗。

With the significant increase of requests from delay-sensitive users in the Internet of Things(IoT)network,Edge Computing(EC)is emerging as a promising paradigm to provide flexible and reliable services.Considering the conflict between the limited storage and computing resources of IoT devices and the latency of user’s requirements,EC spurs a popular wave of offloading user requests to edge servers at the edge of network.The orchestration for user-requested offloading schemes is a remarkable challenge for the execution time and energy consumption of IoT devices in edge networks.To address this issue,a dynamic computation offloading strategy is proposed,which consists of the following two parts:①the concept of intermediate nodes,which have sufficient computing resources and are able to minimize the time and energy cost of task processing by dynamically combining task offloading and service migration strategies;②the optimal intermediate node selection problem is modeled as multi-dimensional Markov Decision Process(MDP)space,and deep reinforcement learning algorithms are implemented to reduce the large MDP space and achieve fast decision-making.Experimental results show that this strategy performs better than existing baseline approaches on the reduction of delay and energy consumption.