Probabilistic computation offloading for mobile edge computing in dynamic network environment

Abstract Explosive increase in mobile devices, IoT devices, connected cars, etc., are straining cloud computing servers and network devices. It is envisioned that the mobile edge computing (MEC), which is a new paradigm for providing computation at the edge of the network to support wireless devices to offload computational intensive tasks to MEC servers for execution, is the best possible solution currently. Different users have tasks/applications requiring various computational power with distinct computing target latency for smooth running of the applications. Moreover, tasks arriving rate at the MEC server for computation varies depending upon the time of the day and users density. In such varying environment, it is necessary to consider probabilistic approach to offload tasks for successful mobile edge computing. In this paper, we formulate successful computation probability, successful communication probability and successful edge computing probability for offloading tasks to the MEC server. We then analyze by simulation how the formulated probabilities vary for different sizes of task, task’s target latency and task arrival rate at the MEC server helping users to make offloading decision.