Abstract
We consider a distributed server system and ask which policy should be used for assigning jobs (tasks) to hosts. In our server, jobs are not preemptible. Also, the job's service demand is not known a priori. We are particularly concerned with the case where the workload is heavy-tailed, as is characteristic of many empirically measured computer workloads. We analyze several natural task assignment policies and propose a new one TAGS (Task Assignment based on Guessing Size). The TAGS algorithm is counterintuitive in many respects, including load unbalancing, non-work-conserving, and fairness. We find that under heavy-tailed workloads, TAGS can outperform all task assignment policies known to us by several orders of magnitude with respect to both mean response time and mean slowdown, provided the system load is not too high. We also introduce a new practical performance metric for distributed servers called server expansion. Under the server expansion metric, TAGS significantly outperforms all other task assignment policies, regardless of system load.
- Bestavros, A. 1997. Load profiling: A methodology for scheduling real-time tasks in a distributed system. In Proceedings of ICDCS '97 (May).]] Google Scholar
- Crovella, M. E., and Bestavros, A. 1997. Self-similarity in World Wide Web traffic: Evidence and possible causes. IEEE/ACM Trans. Netw. 5, 6 (Dec.), 835--846.]] Google Scholar
- Crovella, M. E., Harchol-Balter, M., and Murta, C. 1998a. Task assignment in a distributed system: Improving performance by unbalancing load. In Proceeding of ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems Poster Session. ACM, New York.]] Google Scholar
- Crovella, M. E., Taqqu, M. S., and Bestavros, A. 1998b. Heavy-tailed probability distributions in the World Wide Web. In A Practical Guide to Heavy Tails. Chapman & Hall, New York, chap. 1, pp. 1--23.]] Google Scholar
- Downey, A. B. 1997. A parallel workload model and its implications for processor allocation. In Proceedings of High Performance Distributed Computing (Aug.). pp. 112--123.]] Google Scholar
- Ephremides, A., Varaiya, P., and Walrand, J. 1980. A simple dynamic routing problem. IEEE Trans. Automat. Cont. AC-25, 4, 690--693.]]Google Scholar
- Feitelson, D., and Jette, M. A. 1997. Improved utilization and responsiveness with gang scheduling. In Proceedings of IPPS/SPDP '97 Workshop (Apr.). Lecture Notes in Computer Science, vol. 1291. Springer-Verlag, New York, pp. 238--261.]] Google Scholar
- Feitelson, D., Rudolph, L., Schwiegelshohn, U., Sevcik, K., and Wong, P. 1997. Theory and practice in parallel job scheduling. In Proceedings of IPPS/SPDP '97 Workshop (Apr.), Lecture Notes in Computer Science, vol. 1291. Springer-Verlag, New York, pp. 1--34.]] Google Scholar
- Harchol-Balter, M., Crovella, M., and Murta, C. 1999. On choosing a task assignment policy for a distributed server system. J. Paral. Distr. Comput. 59, 204--228.]] Google Scholar
- Harchol-Balter, M., and Downey, A. 1997. Exploiting process lifetime distributions for dynamic load balancing. ACM Trans. Comput. Syst. 15, 3.]] Google Scholar
- Irlam, G. 1994. Unix file size survey---1993. Available at http://www.base.com/gordoni/ ufs93.html.]]Google Scholar
- Khinchin, A. Y. 1932. Mathematical theory of stationary queues. Mat. Sbornik 39, 73--84.]]Google Scholar
- Koole, G., Sparaggis, P., and Towsley, D. 1999. Minimizing response times and queue lengths in systems of parallel queues. J. Appl. Prob. 36, 1185--1193.]]Google Scholar
- Leiserson, C. 1998a. The Pleiades alpha cluster at M.I.T. Documentation at: http://bonanza.lcs. mit.edu.]]Google Scholar
- Leiserson, C. 1998b. The Xolas supercomputing project at M.I.T. Documentation available at: http://xolas.lcs.mit.edu.]]Google Scholar
- Leland, W. E., and Ott, T. J. 1986. Load-balancing heuristics and process behavior. In Proceedings of Performance and ACM Sigmetrics. ACM, New York, pp. 54--69.]] Google Scholar
- Nelson, R. D., and Philips, T. K. 1989. An approximation to the response time for shortest queue routing. Perf. Eval. Rev. 7, 1, 181--189.]] Google Scholar
- Nelson, R. D., and Philips, T. K. 1993. An approximation for the mean response time for shortest queue routing with general interarrival and service times. Perf. Eval. 17, 123--139.]] Google Scholar
- Parsons, E. W., and Sevcik, K. C. 1997. Implementing multiprocessor scheduling disciplines. In Proceedings of IPPS/SPDP '97 Workshop (Apr.), Lecture Notes in Computer Science, vol. 1459. Springer-Verlag, New York, pp. 166--182.]] Google Scholar
- Paxson, V., and Floyd, S. 1995. Wide-area traffic: The failure of Poisson modeling. IEEE/ACM Trans. Netw. (June), 226--244.]] Google Scholar
- Peterson, D. L., and Adams, D. R. 1996. Fractal patterns in DASD I/O traffic. In CMG Proc.]]Google Scholar
- Pollaczek, F. 1930. Uber eine aufgabe dev wahrscheinlichkeitstheorie. I-II Math. Zeitschrift. 32, 64--100.]]Google Scholar
- The PSC's Cray J90's. 1998. http://www.psc.edu/machine/cray/j90/j90.html.]]Google Scholar
- . Supercomputing at the NAS facility. 1998. http://www.nas.nasa.gov/Technology/Supercomputing/.]]Google Scholar
- Rudolph, L., and Smith, P. H. 2000. Valuation of ultra-scale computing systems. In Proceedings of the 6th Workshop on Job Scheduling Strategies for Parallel Processing (Cancun, Mexico, May). Lecture Notes in Computer Science, vol. 1911. Springer-Verlag, New York, http:/www.cs.huji.ac.il/~feit/ parsched/parsched00.html.]] Google Scholar
- Schroeder, B., and Harchol-Balter, M. 2000. Evaluation of task assignment policies for supercomputing servers: The case for load unbalancing and fairness. In Proceedings of the 9th IEEE Symposium on High Performance Distributed Computing (Aug.) IEEE Computer Society Press, Los Alamitos, Calif.]] Google Scholar
- Shaikh, A., Rexford, J., and Shin, K. G. 1999. Load-sensitive routing of long-lived ip flows. In Proceedings of SIGCOMM (Sept.). ACM, New York.]] Google Scholar
- Sozaki, S., and Ross, R. 1978. Approximations in finite capacity multiserver queues with poisson arrivals. J. Appl. Prob. 13, 826--834.]]Google Scholar
- Whitt, W. 1986. Deciding which queue to join: Some counterexamples. Oper. Res. 34, 1 (Jan.). 226--244.]] Google Scholar
- Winston, W. 1977. Optimality of the shortest line discipline. J. Appl. Prob. 14, 181--189.]]Google Scholar
- Wolff, R. W. 1989. Stochastic Modeling and the Theory of Queues. Prentice-Hall, Englewood Cliffs, N.J.]]Google Scholar
Index Terms
Task assignment with unknown duration
Recommendations
A least flow-time first load sharing approach for distributed server farm
The most critical property exhibited by a heavy-tailed workload distribution (found in many WWW workloads) is that a very small fraction of tasks make up a large fraction of the workload, making the load very difficult to distribute in a distributed ...
On Choosing a Task Assignment Policy for a Distributed Server System
Special issue on software support for distributed computingWe consider a distributed server system in which each host processes tasks in First-Come-First-Served order and each task's service demand is known immediately upon task arrival. We consider four task assignment policies commonly proposed for such ...
Task assignment based on prioritising traffic flows
OPODIS'04: Proceedings of the 8th international conference on Principles of Distributed SystemsWe consider the issue of task assignment in a distributed system under heavy-tailed (ie. highly variable) workloads. A new adaptable approach called TAPTF (Task Assignment based on Prioritising Traffic Flows) is proposed, which improves performance ...
Comments