ServiceSim: A Modelling and Simulation Toolkit of Microservice Systems in Cloud-Edge Environment

Haomai Shi, Xiang He, Teng Wang, Zhongjie Wang

Abstract: With the utilization of edge servers, cloud-native microservice systems are gradually evolving to the network edge, and large-scale distributed microservice systems in cloud-edge environments are emerging. Due to the limited resources of edge servers and dynamic end-user requests, service providers have to continuously propose optimized resource allocation, scheduling, and microservice system configuration policies to balance cost and quality of services. In the early stages of policy proposal, there is an urgent need for service providers to know how well the policy is working and to use this to rapidly iterate and optimize it. However, policy validation in such large-scale real cloud-edge environment is time-consuming and high resource cost. We propose ServiceSim, a simulation toolkit to simulate microservice systems in large scale cloud-edge environment to support policy validation. By comparing with real microservice system, it is show that ServiceSim can correctly reflect the trend of response time of service chains in a microservice system under dynamic end-user requests. Meanwhile, experiments validating edge collaboration and service deployment policies in traffic scenarios reflecting temporal and spatial preferences further illustrate that ServiceSim can effectively help the analysis of microservice system configuration policies and sensitively perceive cloud-edge network changes and the service structure of microservice system.