Adaptive Resource Management in Cloud Infrastructures
Speaker(s) : Rick Schlichting (AT&T Labs)
Shared infrastructures such as those popularized by cloud computing offer the potential for vastly improved resource utilization through a combination of application multi-tenancy and virtualization-enabled resource management techniques. However, to realize this potential requires balancing multiple complex factors such as application performance, resource usage, and the impact of changing resource allocations, all in a highly dynamic environment. This talk will outline two complementary efforts that aim to address these challenges for distributed multi-tier applications using online adaptive system techniques. The first is Mistral, a holistic controller framework that optimizes power consumption, performance benefits, and the transient costs incurred by various adaptations and the controller itself to maximize overall utility. The second is System Tomography, which focuses in on the specific issue of predicting the impact of adaptive actions on the execution and resource consumption of multi-tier applications in a non-intrusive way.
This work is joint with Matti Hiltunen and Kaustubh Joshi of AT&T, Gueyoung Jung and Calton Pu of Georgia Tech, and Shuyi Chen and Bill Sanders of the University of Illinois at Urbana-Champaign.
Rick Schlichting is currently Executive Director of Software Systems Research at AT&T Labs in Florham Park, NJ. He received the B.A. degree in mathematics and history from the College of William and Mary, and the M.S. and Ph.D. degrees in computer science from Cornell University. He was on the faculty at the University of Arizona from 1982-2000, and spent sabbaticals in Japan in 1990 at Tokyo Institute of Technology and in 1996-97 at Hitachi Central Research Lab. Schlichting is an ACM Fellow and an IEEE Fellow, has served on the editorial boards of a number of IEEE magazines and journals, and has been on the technical program committees for over 70 conferences and workshops. He is also the current chair of IFIP Working Group 10.4 on Dependable Computing and Fault-Tolerance, and has been active in the IEEE Computer Society Technical Committee on Dependable Computing and Fault Tolerance, serving as Chair of that organization from 1998-99. His research interests include highly dependable computing, distributed systems, and networks.