CPI2: CPU performance isolation for shared compute clusters
CPI2: CPU performance isolation for shared compute clusters
复制标题
DOI：
10.1145/2465351.2465388
复制
发表时间：
2013-04
期刊：
影响因子：
--
通讯作者：
Xiao Zhang;Eric Tune;R. Hagmann;Rohit Jnagal;Vrigo Gokhale;J. Wilkes
中科院分区：
文献类型：
--
作者：
Xiao Zhang;Eric Tune;R. Hagmann;Rohit Jnagal;Vrigo Gokhale;J. Wilkes
性能隔离是云计算中的一个关键挑战。不幸的是，Linux对共享资源（如处理器缓存和内存总线）中的性能干扰几乎没有防御措施，因此云中的应用程序可能会遇到由其他程序的行为引起的不可预测的性能。我们的解决方案CPI2使用硬件性能计数器获得的每指令周期（CPI）数据来识别问题，选择可能的肇事者，然后选择性地对其进行节流，以便受害者可以返回到其预期的行为。它通过聚合来自同一工作中多个任务的数据来自动学习正常和异常行为。我们已经在Google的所有共享计算集群中推出了CPI2。本文介绍的分析，导致我们的结果，包括案例研究和大规模的评估，其解决真实的生产问题的能力。
Performance isolation is a key challenge in cloud computing. Unfortunately, Linux has few defenses against performance interference in shared resources such as processor caches and memory buses, so applications in a cloud can experience unpredictable performance caused by other programs' behavior. Our solution, CPI2, uses cycles-per-instruction (CPI) data obtained by hardware performance counters to identify problems, select the likely perpetrators, and then optionally throttle them so that the victims can return to their expected behavior. It automatically learns normal and anomalous behaviors by aggregating data from multiple tasks in the same job. We have rolled out CPI2 to all of Google's shared compute clusters. The paper presents the analysis that lead us to that outcome, including both case studies and a large-scale evaluation of its ability to solve real production issues.