This website requires JavaScript.

An Adaptive Resilience Testing Framework for Microservice Systems

Tianyi YangCheryl LeeJiacheng ShenYuxin SuYongqiang YangMichael R. Lyu
Dec 2022
摘要
Resilience testing, which measures the ability to minimize servicedegradation caused by unexpected failures, is crucial for microservice systems.The current practice for resilience testing relies on manually defining rulesfor different microservice systems. Due to the diverse business logic ofmicroservices, there are no one-size-fits-all microservice resilience testingrules. As the quantity and dynamic of microservices and failures largelyincrease, manual configuration exhibits its scalability and adaptivity issues.To overcome the two issues, we empirically compare the impacts of commonfailures in the resilient and unresilient deployments of a benchmarkmicroservice system. Our study demonstrates that the resilient deployment canblock the propagation of degradation from system performance metrics (e.g.,memory usage) to business metrics (e.g., response latency). In this paper, wepropose AVERT, the first AdaptiVE Resilience Testing framework for microservicesystems. AVERT first injects failures into microservices and collects availablemonitoring metrics. Then AVERT ranks all the monitoring metrics according totheir contributions to the overall service degradation caused by the injectedfailures. Lastly, AVERT produces a resilience index by how much the degradationin system performance metrics propagates to the degradation in businessmetrics. The higher the degradation propagation, the lower the resilience ofthe microservice system. We evaluate AVERT on two open-source benchmarkmicroservice systems. The experimental results show that AVERT can accuratelyand efficiently test the resilience of microservice systems.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答