Understanding Cloud Workloads Performance in a Production like Environment

Understanding inter-VM interference is of paramount importance to provide a sound knowledge and understand where performance degradation comes from in the current public cloud. With this aim, this paper devises a workload taxonomy that classifies applications according to how the major system resources affect their performance (e.g., tail latency) as a function of the level of load (e.g., QPS). After that, we present three main studies addressing three major concerns to improve the cloud performance: impact of the level of load on performance, impact of hyper-threading on performance, and impact of limiting the major system resources (e.g., last level cache) on performance. In all these studies we identified important findings that we hope help cloud providers improve their system utilization.
Comment: 16 pages, 17 figures. Submitted to Journal of Parallel and Distributed Computing