Resource management for mobile edge computing using user mobility prediction

Abstract

This paper proposes a resource management for Mobile Edge Computing (MEC) using user mobility prediction. MEC is a new technology in which distributed edge servers (ESs) proceed the task of users in real-time. In MEC, ESs are distributedly installed and users have to decide which ESs they use to proceed the tasks. At this time, the users have to consider the connectivity of ESs. When the tasks of the user are distributedly proceeded with multiple ESs, and the connection of ESs is terminated by user mobility when users collect the computational results from ESs, the users cannot collect results of tasks and users have to proceed the tasks again. The connection loss due to mobility of the users is a big issue. Therefore, this paper predicts user mobility with Kalman filter for estimation of the connectivity. Using mobility prediction, users can select the stable ES during task request and task collection. By using this process, it has been shown that the success rate of collecting results improves.