Surprising results on task assignment in server farms with high-variability workloads

SIGMETRICS/Performance'09 - Proceedings of the 11th International Joint Conference on Measurement and Modeling of Computer Systems

Volumn 37, Issue 1, 2009, Pages 287-298

  Harchol Balter, Mor ^a   Scheller Wolf, Alan ^b   Young, Andrew ^c  

^a Carnegie Mellon University   (United States)

^b CARNEGIE MELLON UNIVERSITY   (United States)

^c NONE

Author keywords

Algorithms; C.1.4 processor architectures : Parallel architectures Distributed architectures; C.4 performance of systems : Design studies; D.4.8 operating systems : Performance modeling and prediction; Design; Performance

Indexed keywords

C.1.4 [PROCESSOR ARCHITECTURES]: PARALLEL ARCHITECTURES - DISTRIBUTED ARCHITECTURES; DESIGN PERFORMANCE; DESIGN STUDIES; DISTRIBUTED ARCHITECTURE; OPERATING SYSTEMS; PERFORMANCE OF SYSTEMS; PROCESSOR ARCHITECTURES;

COMPUTER OPERATING SYSTEMS; PARALLEL ALGORITHMS; PARALLEL ARCHITECTURES; SIZE DISTRIBUTION;

DESIGN;

EID: 70449641216     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: 10.1145/1555349.1555383     Document Type: Conference Paper

References (30)

1. Full paper (with appendix) available at http://www.cs.cmu.edu/~harchol/ Papers/Sigmetrics09sita.pdf.
2. Paul Barford and Mark Crovella. Generating representative Web workloads for network and server performance evaluation. In Proceedings of the ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, pages 151-160, July 1998.
3. Onno Boxma, Q. Deng, and Albertus Zwart. Waiting-time asymptotics for the M/G/2 queue with heterogeneous servers. Queueing Systems, 40(1):5-31, 2002.
4. James Broberg, Zahir Tari, and Panlop Zeephongsekul. Task assignment with work-conserving migration. Parallel Computing, 32:808-830, 2006.
5. John Buzacott and George Shanthikumar. Stochastic Models in Manufacturing Systems. Prentice Hall, 1993.
6. Valeria Cardellini, Emiliano Casalicchio, Michele Colajanni, and Philip Yu. The state of the art in locally distributed web-server systems. Technical report, 2001.
7. Gianfranco Ciardo, Alma Riska, and Evgenia Smirni. Equiload: a load balancing policy for clustered web servers. Performance Evaluation, 46:101-124, 2001.
8. Mark Crovella and Azer Bestavros. Self-similarity in World Wide Web traffic: Evidence and possible causes. pages 160-169, May 1996.
9. Muhammad El-Taha and Bacel Maddah. Allocation of service time in a multiserver system. Management Science, 52(4):623-637, 2006.
10. Hanhua Feng, Vishal Misra, and Dan Rubenstein. Optimal state-free, size-aware dispatching for heterogeneous M/G-type systems. Performance Evaluation, 62:475-492, 2005.
11. Serguei Foss and Dmitry Korshunov. Heavy tails in multi-server queue. Queueing Systems, 52:31-48, 2006.
12. Bin Fu, James Broberg, and Zahir Tari. Task assignment strategy for overloaded systems. In Proceedings of the Eighth IEEE International Symposium on Computers and Communications, 2003.
13. Mor Harchol-Balter. Task assignment with unknown duration. Journal of the ACM, 49(2):260-288, March 2002.
14. Mor Harchol-Balter, Mark Crovella, and Cristina Murta. On choosing a task assignment policy for a distributed server system. IEEE Journal of Parallel and Distributed Computing, 59:204-228, 1999.
15. Mor Harchol-Balter and Allen Downey. Exploiting process lifetime distributions for dynamic load balancing. In Proceedings of ACM SIGMETRICS, pages 13-24, Philadelphia, PA, May 1996.
16. Wallace Hopp and Mark Spearman. Factory Physics. McGraw Hill/Irwin, 2 edition, 2000.
17. Steven Hotovy, David Schneider, and Timothy O'Donnell. Analysis of the early workload on the Cornell Theory Center IBM SP2. 1996.
18. Leonard Kleinrock. Queueing Systems, volume I. Theory. John Wiley & Sons, 1975.
19. Guy Latouche and Vaidyanathan Ramaswami. Introduction to Matrix Analytic Methods in Stochastic Modeling. Society for Industrial Mathematics, 1987.
20. Kazumasa Oida and Kazumasa Shinjo. Characteristics of deterministic optimal routing for a simple traffic control problem. In IPCCC, February 1999.
21. Konstantinos Psounis, Pablo Molinero-Fernández, Balaji Prabhakar, and Fragkiskos Papadopoulos. Systems with multiple servers under heavy-tailed workloads. Performance Evaluation, 52:456-474, 2005.
22. Alan Scheller-Wolf. Further delay moment results for FIFO multiserver queues. Queueing Systems, 34:387-400, 2000.
23. Alan Scheller-Wolf and Karl Sigman. Delay moments for FIFO GI/GI/s queues. Queueing Systems, 25:77-95, 1997.
24. Alan Scheller-Wolf and Karl Sigman. New bounds for expected delay in FIFO GI/GI/c queues. Queueing Systems, 26:169-186, 1997.
25. Alan Scheller-Wolf and Rein Vesilo. Structural interpretation and derivation of necessary and sufficient conditions for delay moments in FIFO multiserver queues. Queueing Systems, 54:221-232, 2006.
26. Bianca Schroeder and Mor Harchol-Balter. Evaluation of task assignment policies for supercomputing servers: The case for load unbalancing and fairness. Cluster Computing: The journal of Networks, Software Tools, and Apps, 7(2):151-161, April 2004.
27. Anees Shaikh, Jennifer Rexford, and Kang Shin. Load-sensitive routing of long-lived IP flows. In Proceedings of ACM SIGCOMM, September 1999.
28. Zahir Tari, James Broberg, Albert Zomaya, and Roberto Baldoni. A least flow-time first load sharing approach for distributed server farm. Journal of Parallel and Distributed Computing, 65:832-842, 2005.
29. Nigel Thomas. Comparing job allocation schemes where service demand is unknown. Journal of Computer and System Sciences, 74:1067-1081, 2008.
30. Ward Whitt. The impact of a heavy-tailed service-time distribution upon the M/GI/s waiting-time distribution. Queueing Systems, 36:71-87, 2000.