A Mobile Device Multitasking Model with Multiple Mobile Edge Computing Servers

摘要

Mobile Edge Computing (MEC) is an emerging computing framework that meets the growing computing needs of mobile applications. However, when the user scale is too large, the MEC server may be overloaded. In order to provide satisfactory computing power, it is of great significance that how to break the performance bottleneck of the MEC server. In this paper, we propose a Mobile Device Multitasking Model in Multiple MEC Servers Scenario (MDMS). This model combines collaborative computing with MEC. Each idle mobile device is treated as a virtual MEC server, which shares computing resources with other devices in the network, thereby relieving the pressure on real MEC servers. Each mobile device can perform tasks to the maximum. Based on these, we consider the multi-task interference problem, which further improves the rationality of the model. Meanwhile, this model can run in multiple MEC server scenarios. Furthermore, taking into account that the offloading decision and the computing resource allocation are tightly coupled, a two-layer optimization method is adopted to minimize the total energy consumption of MEC system. Experimental results show the superiority of the model. Moreover, we analyse the relationship between the number of MEC servers and the energy consumption.