CI-P: Computer System Failure Data Repository to Enable Data-Driven Dependability Research

点击翻译按钮获取中文摘要

Dependability is a requirement for computer systems; however, researchon dependable systems is hampered by a lack of real and publiclyavailable failure data. This can result in productive paths ofresearch being closed to most researchers and, conversely,unproductive research being performed due to faulty assumptions aboutthe manner in which real systems fail. The goal of this project is toplan a collaborative effort to collect, curate, and provide publicaccess to failure data for large scale computer systems through acommunity repository. One challenge is that failure data isconsidered sensitive by the owners. The ultimate goal of this projectis to collect the data from some of the NSF-funded largecyberinfrastructure projects, such as NEES, LIGO, XSEDE, and NRAO.The specific goal of this planning project is to collect requirementsfrom potential praticipants (both users and contributors of data sets)and seed a prototype repository with data sets from two of the largestand latest clusters at Purdue. The data sets will comprise staticinformation, dynamic information about the workloads, and failureinformation, for both planned and unplanned outages.The broader impact in the project will be achieved through thedissemination of the data sets to a wide variety of researchers, andperhaps even, practitioners. The datasets will let people run theircampus clusters more efficiently, i.e., with fewer failures, at higherutilization and energy efficiency.