Clustering Based Efficient MEC Server Placement and Association in 5G Networks

Abstract

Efficient placement of Mobile Edge Computing (MEC) servers is a great challenge in a 5G cellular networks, as the number of Radio Access Networks (RANs) increases drastically. With the increase in RAN, the number of MEC servers to be deployed will also increase, leading to an increase in the overall cost for any telecom operator. In order to address this issue and minimize the overall cost, many researches have already been carried out. But the research outcomes mostly present different techniques for minimizing the propagation delay between the MEC servers and their associated RANs. In most of these works, the same service capacity MEC servers are deployed throughout the networks, irrespective of the total workload, leading to overutilization and underutilization of the resources. To address this issue, we model our problem as a cost minimization problem in which minimum number of MEC servers are used. In this paper, we propose a clustering-based efficient approach to locate the MEC servers properly as well as to associate the RANs with the MEC servers appropriately. We propose a greedy-based algorithm called MEC Placement and Association (MEC_PA) algorithm that determines the MEC server location as well as its associated RANs. The results show that our proposed algorithm outperforms the existing works.