Virtual Machine Placement Algorithm for Supporting Multiple Applications to Mobile Edge Computing

In Mobile Edge Computing (MEC) environment, deploying application services in the form of Virtual Machines (VM) at the edge of the network can effectively reduce the service response delay and reduce the data traffic of core network. There have been many solutions to the problem of optimal allocation of edge network resources, but few studies consider the optimal deployment of VM that provide users with multiple application services to mobile edge networks. To this end, two heuristic algorithms are proposed, Fitness-based Heuristic Placement Algorithm (FHPA) and Divide-and-Conquer Based Heuristic Placement Algorithm (DCBHPA). By distributing VMs that support multiple application services to the MEC network, these two algorithms aim to minimize the data traffic in MEC architecture. Besides, FHPA and DCBHPA define respectively different fitness computing models based on the network connection characteristics of edge servers, as well as the differences in users' application requests. Thus, VM configuration can be realized through the sub-problem division mechanism. Compared with the baseline algorithms, the simulation results show that the proposed algorithms can better control the system data traffic and improve effectively the utility of edge network service resources.