面向微服务架构的云系统负载均衡机制

针对微服务架构的请求响应延迟上升问题，提出一种微服务链感知的请求负载均衡算法。从负载均衡器入手，将微服务链上的平均请求延迟和主机负载情况作为衡量指标，形式化微服务环境和请求延迟，在此基础上研究微服务链调用中存在的共享微服务竞争问题。模拟实验结果表明，与RR算法相比，该算法在复杂的微服务链环境下能够有效降低请求延迟，且在实例分布不均匀的环境中保持较好的负载性能，均衡不同主机之间的负载。在更接近真实应用环境的高频请求测试中，算法能有效降低系统的综合响应时间。

Load Balancing Mechanism for Microservice Architecture in Cloud-based Systems

In order to solve the rising delay in responding to requests in Microservices Architecture(MSA), a load balancing algorithm for requests is proposed based on microservice chain awareness.To formalize the request delay and microservice environment, the algorithm takes the average request delay and the host loads as the evaluation indicators. On this basis, the competition for shared microservices in microservice chain calling is studied.The simulation results show that compared with the RR algorithm, the proposed algorithm can significantly reduce the delay in requests in complex microservice chain environment.In addition, it keeps excellent loading performance in an environment with unevenly distributed instances, and balances the loads of different hosts.In the more realistic tests of high-frequency requests, the proposed algorithm can significantly reduce the overall response time of the system.