Adaptive Replication for Mobile Edge Computing

Abstract

Mobile edge computing (MEC) has been proposed to complement the limitations of cloud computing. Allocating replicas in MEC servers can shorten transmission latency and improve the quality of service (QoS). However, cloud-based replication schemes may not be suitable for MEC servers due to their limited resources and the varied traffic in RANs. In this paper, we analyze the revenue, cost, and profit for the replicas in MEC servers. Accordingly, we present algorithms of adaptive replication for MEC. Our algorithms dynamically allocate replicas based on the number of read/write operations in each time interval. We also employ an algorithm that adaptively forwards requests to replicas in the neighboring MEC servers to balance the load of MEC servers. The simulation results show that our scheme can effectively shorten the average response time and improve the QoS of packets for better profit.