Optimization of Computation Resource for Container-Based Multi-MEC Collaboration System

Mobile Edge Computing (MEC) has potential to improve performance of 5G networks in terms of latency and bandwidth. Recently, the architecture of distributed MEC host integrated with 5G networks is proposed by ETIS, providing a new paradigm for implementing edge networks. In this paper, considering the performance optimization and user mobility in edge networks, Docker container is applied to MEC host to form a container-based multi-MEC collaboration system (CMMCS) so that it can optimize computation resources among MEC hosts while guaranteeing QoS. Then the resource allocation optimization is formulated to the system cost minimization and the general pattern of virtual machine resource consumption is considered. After that, we prove that the system cost minimization problem is an NP-hard problem, and a DABF scheme based container deployment algorithm is proposed to figure out the optimal solution. Simulation results demonstrate that our proposed algorithm achieves significant reduction on system cost in CMMCS compared to other baselines in any network environment.