Mobility-aware dynamic service placement in D2D-assisted MEC environments

Abstract

Mobile Edge Computing (MEC) has emerged as a promising networking paradigm that provides delay-sensitive service for mobile users at the edge of core networks, where mobile users can offload their computing-intensive tasks to MEC networks for processing on no time. Furthermore, with the advance of communication and fabrication technologies, mobile devices now have adequate computing and storage processing capabilities. The device-to-device (D2D) offloading as a new offloading technique enables mobile users to offload their tasks to other mobile devices (referred as helper mobile devices) for processing, thereby alleviating the processing burden on servers in MEC. However, fully utilizing the D2D technique in an MEC network for task offloading service is challenging. Particularly, the mobility of both mobile users and their helper mobile devices makes efficient offloading service placement become difficult. In this paper, we study a novel Mobility-aware Dynamic Offloading Service Placement (MDOSP) problem in a D2D-assisted MEC environment with the aim to minimize the total cost of offloading task services that consists of the computing cost, communication delay cost and migration cost, without the knowledge of future mobility information of mobile users and helper mobile devices. We first formulate an Integer Nonlinear Programming (INP) for the offline setting of the problem. We then prove the NP-hardness and develop an online algorithm with a provable competitive ratio for the problem. We finally evaluate the performance of the proposed algorithms through experimental simulations. Experimental results demonstrate that the proposed algorithms are promising, compared with existing baseline algorithms.