Modeling and performance analysis of smart map application in the Multi-access Edge Computing paradigm

Abstract

Multi-access Edge Computing (MEC) recognized as an emerging technology that provides cloud computing services in the proximity of Radio Access Network (RAN). Mobile subscribers could offload their computation-intensive or memory-intensive tasks to the nearest Base Stations (BS) and deliver the cloud services from MEC servers. This technology can meet the requirements of low latency and location awareness. Smart map applications are an indispensable part of our digital life in large cities. In this paper, we introduce how the map application could gain from MEC technology. We concentrate on performance modeling and analysis of direction-finding in the map application. We study the impact of variation of workload, the number of servers, queue length, and connection failure on the task rejection probability and mean sojourn time. We model each phase of direction-finding service by the Markov Reward Model (MRM). The numerical results presented by applying models in the SHARPE software package. Furthermore, inter-dependencies between sub-models resolved by the fixed-point iteration technique. Moreover, the analytical results verified by the Discrete-Event Simulation (DES), which conducted in MATLAB. The result could help the MEC providers to assess their platform performance.