MEC-Assisted Admission Control Based on Convergence of Communication and Computation.

Yanli Qi,Lin Tian,Yiqing Zhou,Jinhong Yuan,Jinglin Shi,Xiaohu Ge

DOI: https://doi.org/10.1109/icc.2019.8761843

2019-01-01

Abstract:As an important component of resource management, admission control is vital to prevent the wireless network from congestion and ensure the quality of service (QoS). Mobile edge computing (MEC), which provides computing resources at the edge of radio access networks (RAN), is able to better support new mobile services. Therefore, enhanced admission control policies should be designed for mobile cellular networks based on MEC. In this paper, a novel MEC-assisted admission control mechanism is proposed from the perspective of convergence of communication and computation. In this mechanism, MEC computing resources are leveraged to pre-process the transmission content. The purpose is to reduce the consumption of wireless bandwidth and increase the number of accepted services. Next, the admission control process is modeled as a Markov decision process (MDP) with objective to maximize the long-term expected average effective throughput. In consideration of the large state space, a simulation-based optimization algorithm of MDP is adopted to obtain the optimal policy. Simulation results show that our proposed admission control mechanism achieves higher effective throughput than that without MEC computing resources. And the probability of accepted services can also be improved significantly. Furthermore, the optimal amount of MEC computing resources can be acquired according to the system traffic statistics.