Performance Analysis of Heterogeneous Mobile Edge Computing Networks with Multi-core Server

Abstract

The mobile edge computing (MEC) architecture enables the transfer of the computing capability from the cloud to the edge of the network which further improves the network performance. Most initial works are based on the analysis of single-core MEC servers. In this paper we theoretically investigate the performance of MEC network with multi-core edge servers by combining stochastic geometry and queueing theory. By considering the statistical distribution of waiting delay and service delay of the computation tasks on MEC servers, we obtain the successful computing probability in a MEC heterogeneous network model. Combining the successful communication probability and the successful computing probability, we derive the total successful edge computing probability. The accuracy of theoretical derivation is verified through simulation. We compare the successful computing probability, successful edge computing probability and mean latency of single-core edge server and dual-core edge server. The theoretical analysis of the performance of MEC networks provides a certain basis for the design, deployment and optimization of MEC technologies.