Performance analysis of multi-access edge computing networks for heterogeneous IoT systems

Abstract

The emergence of Multi-access Edge Computing (MEC) has drastically transformed the computing capabilities of communication systems. By facilitating the convenience of low latency and on-demand services for delay-sensitive tasks, these platforms have also contributed towards mitigating the communication overheads to a larger extent. An emerging application area which has largely benefitted with the adoption of MEC is the Internet of Things (IoT) paradigms. However, with the exponential rise in interconnected IoT devices requesting provisioning of resources to the MEC may lead to traffic bottlenecks in the MEC networks. This further leads to lower response times and degradation in the network’s Quality of Service (QoS). In view of modeling request processing mechanism and monitoring of network QoS, a Markov process driven framework to model MEC-based IoT systems is proposed. By exploiting the analytical inference from pure birth–death processes and queue theoretic models the steady-state solutions for MEC networks is derived. Further, critical performance metrics like mean number of requests in the network, traffic intensity of requests, fluctuations in number of requests, and the mean response time of the network are provided. The corresponding results obtained for the above analysis are validated through numerical illustrations to prove the convergence of the framework and its impact on different network parameters.