First Workshop on
Algorithms and Architectures for Self-Managing Systems
http://tesla.hpl.hp.com/self-manage03/
San Diego, California
June 11, 2003
1:00pm - 8:00pm
Held as part of the
Federated Computing Research
Conference (FCRC)
Co-sponsored by the
International Symposium on Computer
Architecture (ISCA)
and the
International Conference
on Measurement and Modeling of Computer Systems (SIGMETRICS)
Description and Goals
The increasing complexity of computing systems is beginning to overwhelm the capabilities of software developers and system administrators to design, evaluate, integrate, and manage these systems. Major software and system vendors are concluding that the only viable long-term solution is to create computer systems that manage themselves, often referred to as autonomic computing systems.The purpose of this one-day workshop, co-sponsored by ISCA 2003 and SIGMETRICS 2003, is to bring together different communities to address the significant algorithmic, methodological, and architectural challenges of self-managing systems. In the last decade, the statistics, probability theory, machine learning and data mining communities (aided by Moore’s law) have developed many exciting new techniques that infer models of system behavior from large volumes of data, and employ these models in a variety of self-management functions, such as problem diagnosis, prediction, and planning. While systems architects and analysts are recognizing that they must go beyond traditional methods of building complex systems and understanding their behavior, incorporating these advances into practice will pose difficult new challenges. What are the best opportunities to use these new techniques in real systems, and what are their limitations? What structural changes are necessary for dependable self-management? What are the most effective ways to embed, distribute, and coordinate analytic capabilities in complex global-scale distributed systems with many interacting elements and control points?
The tentative final program for the workshop is listed below. There will be plenty of time for audience participation. A proceedings document containing papers for each presentation and poster will be provided for registered attendees at the workshop. Individual papers will be made available below.
Tentative Final Program
Welcome
A Human Organization Analogy for Self-* Systems
J. Strunk and G. Ganger
Carnegie Mellon University
Automatic Design of Dependable Data Storage Systems
K. Keeton and J. Wilkes
Hewlett-Packard Laboratories
A Two-Tiered Software Architecture for Automated Tuning of Disk Layouts
B. Salmon, E. Thereska, C. Soules, and G. Ganger
Carnegie Mellon University
Poster Advertisements
Session 2: Storage-related Posters/Break
Application-centric Integrated Storage Management
E. Anderson, M. Kallahalla, K. Keeton, R. Swaminathan and M. Uysal
Hewlett-Packard Laboratories
Automatic Performance Tuning in the Zettabyte File System
V. Henson, M. Ahrens, and J. Bonwick
Sun Microsystems
Managing Overload via Adaptive Scheduling
A. Riska, E. Riedel and S. Iren
Seagate Research
Optimization of Adaptive Distributed Protocols as a Game
V. Marbukh
National Institute of Standards and Technology
Using Data Mining for Discovering Patterns in Autonomic Storage Systems
Z. Li, S. Srinivasan, Z. Chen, Y. Zhou, P. Tzvetkov, X. Yan and J. Han
University of Illinois at Urbana-Champaign
Session
3: Data Center Resource Provisioning
M. Goldszmidt* and B. Sabata+
*Hewlett-Packard Laboratories and +IET, Inc.
Dynamic Surge Protection: An Approach to Handling Unexpected Workload Surges with Resource Actions that have Lead Times
E. Lassettre, D. Coleman, Y. Diao, S. Froelich, J. Hellerstein, L. Hsiung, T. Mummert, M. Raghavachari, G. Parker, L. Russell, M. Surendra, V. Tseng, N. Wadia, and P. Ye
IBM Corporation
Quantifying the Benefits of Resource Multiplexing in On-Demand Data Centers
A. Chandra, P. Goyal* and P. Shenoy
University of Massachusetts at Amherst and *IBM Almaden Research Center
Utility-Directed Allocation
T. Kelly
Hewlett-Packard Laboratories
Poster Advertisements
Session 4: Self-managing Systems Posters/Break
Approach to Policy Execution in Autonomic Manager Toolkit
H. Chan, D. Verma, A. Segal, B. Arnold, J. Giles, D. Agrawal, and D. Olshefski
IBM TJ Watson Research Center
Existential QoS for Storage
V. Henson, J. Bonwick, and M. Ahrens
Sun Microsystems
Profit-driven Service Differentiation in Transient Environments
Q. Zhang, E. Smirni and G. Ciardo
The College of William and Mary
Self-Managed Leasing for Distributed Systems
K. Bowers*, K. Mills, S. Quirolgico and S. Rose
*Renssalaer Polytechnic Institutes and National Institute of Standards and Technology
Admission Control for E-Commerce Web Sites
S. Elnikety, E. Nahum, J. Tracey*, W. Zwaenoepoel
EPFL and *IBM TJ Watson Research Center
Goal Oriented Policy Determination
D. Verma and S. Calo
IBM TJ Watson Research Center
Nonintrusive Remote Healing Using Backdoors Automatic Data Structure Repair for Self-Healing Systems
Participant Feedback and Concluding Remarks
F. Sultan*, A. Bohra*, I Neamtiu+ and L. Iftode+
*Rutgers University and +University of Maryland at College Park
B. Demsky and M. Rinard
Massachusetts Institute of Technology
Organizing Committee
- Jeffrey Chase, Duke University
- Moises Goldszmidt, Hewlett-Packard Laboratories (co-chair)
- Kimberly Keeton, Hewlett-Packard Laboratories
- Jeffrey Kephart, IBM Research (co-chair)
- William Tetzlaff, IBM Research
- self-manage03@harp.hpl.hp.com
Program Committee
- Armando Fox, Stanford University
- David Oppenheimer, UC Berkeley
- Prabhakar Raghavan, Verity
- Bikash Sabata, IET
- Anand Sivasubramaniam, Penn State University
- Mark Squillante, IBM Research
- Yanyong Zhang, Rutgers University
- Organizing Committee members