Performance Analysis of an Edge Computing System for Real Time Computations and Mobile Users

Recently, the ever increasing emergence of intelligent systems requiring fast data processing, has focused attention on high responsiveness techniques for computation of delay sensitive applications. As regards real time computation, the mobility often plays a very central role, requiring highly responsive information processing and interpretation of the data, retrieved by environment, in order to apply proper strategies for decision making. This paper deals with the evaluation of the influence of the users mobility on the performance of an edge computing system. The paper provides the analysis of real time computations requests through the application of the queuing theory. Each edge node (EN) coverage area has been modeled as M/M/C system with reneging, where the queuing requests are served with FIFO policy, in which the migration of the vehicles requiring processing from a EN to a contiguous one has been taken into account by introducing handoff procedure. The aim of this paper is the minimization of the number of processors used by ENs such that the minimum quality of service constraints, in terms of probability that a computation is blocked and the probability that a handoff attempt fails. Comparisons between simulation and analytical results have been provided for different distributions of the rise of new real time computation requests, confirming the validity of the estimation accuracy of the proposed analysis