ACIS recently hosted fellow researchers Jie Xu, PhD and Peter Garraghan, University of Leeds, UK
Jie Xu, PhD and Peter Garraghan visited the ACIS lab recently and gave a talk titled Cloud Analytics: Insight into Real Operational Behavior in Large-scale Cloud Computing Datacenters.
For the past two years researchers at the University of Leeds have been analyzing large-scale Cloud computing datacenters in order to improve energy-efficiency, quantify Cloud computing behavior in terms of resource utilization and system inefficiencies, develop accurate simulation models and provide insight into real Cloud operational environments for the academic community. This presentation presents award-winning work on the analysis of the Google Cloud trace log, which includes over 12,500 servers over an operational period of 29 days. The four main themes of this talk are as follows: 1) challenges in understanding and analyzing large-scale datacenters, 2) workload characterization and simulation, 3) server characteristics and resource inefficiencies and 4) analysis of failures and the impact of failure-related energy in large-scale Cloud environments.
About the Authors:
Professor Jie Xu is Chair of Computer Science at the School of Computing and leads the research at Leeds on Distributed Systems and Internet Computing. He has worked in the field of dependable distributed systems and fault-tolerant computing for over eighteen years. In 1990 he joined the Computing Laboratory of the University of Newcastle upon Tyne as Research Associate and received the PhD degree there in Advanced Fault-Tolerant Software. He was promoted to Senior Researcher in 1995 and was one of the major contributors to several EC-funded research projects on Dependable Distributed Computing Systems. In 1998 he moved to University of Durham where he co-founded the Distributed Systems & Services group and the DPART laboratory supporting distributed and real-time applications.
Peter Garraghan is a PhD candidate within the Distributed Systems and Services Group. He graduated from Staffordshire University, UK with a BSc First Class Honours in Computer Science. In addition, he worked as an IT specialist for Hewlett-Packard in Germany developing and testing an internal database management system for stock shortages as well as a University lecturer in Hangzhou, China. Peter began his PhD at the University of Leeds in November, 2010 funded by the UK EPSRC WRG platform project. His research interests are cloud computing analytics and energy-aware dependability mechanisms.