Security- and Location-Aware Optimal Virtual Machine Management for 5G-Driven MEC Systems

Abstract

Mobile Edge Computing (MEC) over a densified 5G wireless network deployment arises as an effective response to the skyrocketing demand for latency-driven mobile computing applications which can meet virtual computing, storage, and networking resources one-hop away from their locations. In this context, the design of virtual machine (VM) management mechanisms comes out as key issue in cloud orchestration process given the need to provide these applications with the nearest VMs resources. The design becomes even more challenging when security is taken into account in the resource allocation problem due to need to shield the computation task with additional VMs. In this paper, we propose an optimal security- and location-aware VMM mechanism that efficiently manages the placement of computation tasks in a multi-cloud environment considering MEC and the Back-up cloud. The proposed method is based on the framework of Semi-Markov Decision Process (SMDP) that optimally and dynamically controls the assignment of VMs based on type of services and their stochastic features, cost of security services (which accounts for additional VMs), cost of blocking (which seeks to minimize any breach of availability), and cost of location (which seeks to minimize the latency). The analysis of the optimal policy as well as of the blocking probability and cloud utilization indicates the effectiveness of the proposed VMM mechanism.