Chain-Oriented Load Balancing in Microservice System

摘要

With the continuous development of cloud computing and container technology, microservice architecture has received more and more attention. Due to the convenience in the development and maintenance of splitting large applications into fine-grained single services, lots of large applications have evolved from monolithic architecture to microservice architecture. In a microservice system, a request is often processed by a chain composed of multiple different microservices, and these chains often share microservices [2]. Hence, there is competition for resources between the microservices chain. Furthermore, the latency of a long chain that traverses more microservices is often longer than a short chain, and it consumes user patience seriously. Existing research often ignores or simplifies the competition between microservice chains. This paper proposed a multi-priority queue based chain-oriented load balancing method to reduce the latency of the long chain. Our evaluations on docker containers in 6 cloud servers shows that the method proposed in this paper could reduce the latency of long chains without having much impact on the short chains.