Li Chen, Pooja Jain, Kingsum Chow, and Colin Cunningham, Intel Corporation
There exist multitudes of cloud performance metrics, including workload performance, application placement, software/hardware optimization, scalability, capacity, reliability, agility and so on. Various software applications run in the cloud. In this paper, we consider jointly optimize the performance of the various software applications in the cloud. The challenges lie in bringing a diversity of raw data into tidy data format, unifying performance data from multiple systems based on timestamps, and assessing the quality of the processed performance data. To address the above issues, we proposed an innovative procedure to verify the quality of cloud performance data [1].
Target Audience: Introductory
Li Chen, Pooja Jain, Kingsum Chow, and Colin Cunningham, 2016 Technical Paper, Abstract, Paper, Slides, Notes, Video.
As a continuation of our effort based on [1], here we propose a data-driven framework of applying mathematical programming and machine learning techniques to the cloud performance data, which includes time series data of system activity data or garbage collection logs from multiple systems and software performance data, and obtaining a scientifically optimized solution for cloud computing quality. As a result, we use data-driven analytical methods to cultivate the software performance in the cloud computing environment. We present a case study to demonstrate the effectiveness of our modeling framework and identify several interesting future research directions.
The key contributions of this paper are: identifying and addressing challenges of applying analytics to cloud performance data; formulating mathematical optimization for jointly optimizing performance of various software running in the cloud; proposing a data-driven analytical framework for assessing software performance quality in the cloud.
References: [1] Chen, Li, et al. “Brewing Analytics Quality for Cloud Performance.” arXiv preprint arXiv:1509.00095 (2015). The Proceedings of 33rd Annual Pacific Northwest Software Quality Conference (PNSQC) 2015.
