Abstract:Microservices as an emerging architecture are creating new opportunities to enable superior network services in Mobile Edge Computing (MEC). In the presence of huge amounts of user requests, the massive communications among microservices have become notoriously complicated. Due to the intricate data dependencies of the microservices, the overall performance of large-scale MEC applications simultaneously depends on both service deployment and request routing. However, most existing work ignores the interdependencies of microservices and studies the deployment and routing as two isolated problems. In this case, this paper investigates the joint optimization of service deployment and request routing in edge computing. We first formulate a delay minimization problem via mixed integer linear programming and queuing analysis, and then provide a hardness proof on the problem. In addition, this paper presents a 2-approximation algorithm, followed with rigorous mathematical proofs to demonstrate the approximation ratio. The proposed two-phase algorithm consists of rounding based service deployment and adaptive-scaling-based request routing policies, which employ fine grained joint optimization to minimize service response delay. Finally, we illustrate the near-optimal performance of the proposed algorithm via comprehensive experiments.

Joint Optimization of Service Deployment and Request Routing for Microservices in Mobile Edge Computing