| | This Article | |
| |
| |
| | Share | |
| |
| |
| | Bibliographic References | |
| |
| |
| | Add to: | |
| |
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
| |
| | Search | |
| |
| |
| | |
How Useful Is Old Information?
January 2000 (vol. 11 no. 1)
pp. 6-20
Abstract—We consider the problem of load balancing in dynamic distributed systems in cases where new incoming tasks can make use of old information. For example, consider a multiprocessor system where incoming tasks with exponentially distributed service requirements arrive as a Poisson process, the tasks must choose a processor for service, and a task knows when making this choice the processor queue lengths from T seconds ago. What is a good strategy for choosing a processor in order for tasks to minimize their expected time in the system? Such models can also be used to describe settings where there is a transfer delay between the time a task enters a system and the time it reaches a processor for service. Our models are based on considering the behavior of limiting systems where the number of processors goes to infinity. The limiting systems can be shown to accurately describe the behavior of sufficiently large systems and simulations demonstrate that they are reasonably accurate even for systems with a small number of processors. Our studies of specific models demonstrate the importance of using randomness to break symmetry in these systems and yield important rules of thumb for system design. The most significant result is that only small amounts of queue length information can be extremely useful in these settings; for example, having incoming tasks choose the least loaded of two randomly chosen processors is extremely effective over a large range of possible system parameters. In contrast, using global information can actually degrade performance unless used carefully; for example, unlike most settings where the load information is current, having tasks go to the apparently least loaded server can significantly hurt performance.
[1] M. Alanyali and B. Hajek, “Analysis of Simple Algorithms for Dynamic Load Balancing,” Math. Operations Research, vol. 22, no. 4, pp. 840-871, 1997.
[2] E. Altman and P. Nain, “Closed Loop Control with Delayed Information,” Proc. ACM Sigmetrics Conf. Measurement and Modeling of Computer Systems, pp. 193-204, Newport, R.I., June 1992.
[3] E. Altman and S. Stidham, “Optimality of Monotonic Policies for Two-Action Markovian Decision Processes, with Applications to Control of Queues with Delayed Information,” Queueing Systems: Theory and Applications, vol. 21, nos. III-IV, pp. 267-291, 1995.
[4] Y. Azar, A.Z. Broder, A.R. Karlin, and E. Upfal, Balanced Allocations Proc. 26th ACM Symp. Theory of Computing, pp. 593-602, 1994.
[5] D.R. Cox and W.L. Smith, Queues. Wiley, 1961.
[6] D.L. Eager, E.D. Lazowska, and J. Zahorjan, "Adaptive Load Sharing in Homogeneous Distributed Systems," IEEE Trans. Software Eng., vol. 12, no. 5, pp. 662-675, May 1986.
[7] D.L. Eager, E.D. Lazowska, and J. Zahorjan, "A Comparison of Receiver-Initiated and Sender-Initiated Adaptive Load Sharing," Performance Evaluation, Vol. 6, Mar. 1986, pp. 53-68.
[8] S.N. Ethier and T.G. Kurtz, Markov Processes: Characterization and Convergence. John Wiley&Sons, 1986.
[9] A. Fox, Private communication.
[10] A. Fox, S.D. Gribble, Y. Chawathe, E.A. Brewer, and P. Gauthier, “Cluster-Based Scalable Network Services,” Proc. 16th Symp. Operating System Principles, pp. 78-91, Oct. 1997.
[11] R.M. Karp, M. Luby, and F. Meyer auf der Heide, Efficient PRAM Simulation on a Distributed Memory Machine Proc. 24th ACM Symp. Theory of Computing, pp. 318-326, May 1992.
[12] L. Kleinrock, Queueing System: Volume 1: Theory. New York: John Wiley&Sons, 1975.
[13] J. Kuri and A. Kumar, “Optimal Control of Arrivals to Queues with Delayed Queue Length Information,” IEEE Trans. Automatic Control, vol. 40, pp. 1,444-1,450, 1995.
[14] T.G. Kurtz, Approximation of Population Processes. SIAM, 1981.
[15] R. Mirchandaney, D. Towsley, and J.A. Stankovic, "Analysis of the Effect of Delays on Load Sharing," IEEE Trans. Computers, vol. 38, no. 11, pp. 1,513-1,525, Nov. 1989.
[16] R. Mirchandaney, D. Towsley, and J.A. Stankovic, "Adaptive Load Sharing in Heterogeneous Distributed Systems," J. Parallel and Distributed Computing, Vol. 9, 1990, pp. 331-346.
[17] M. Mitzenmacher, “Constant Time per Edge Is Optimal on Rooted Tree Networks,” Proc. Eighth ACM Symp. Parallel Algorithms and Architectures, pp. 162-169, 1996.
[18] M. Mitzenmacher, “Load Balancing and Density Dependent Jump Markov Processes,” Proc. 37th IEEE Symp. Foundations of Computer Science, pp. 213-222, 1996.
[19] M. Mitzenmacher, The Power of Two Choices in Randomized Load Balancing PhD thesis, Univ. of California Berkeley, 1996.
[20] M. Mitzenmacher, “On the Analysis of Randomized Load Balancing Schemes,” Theory of Computing Systems, vol. 32, pp. 361-386, 1999.
[21] M. Mitzenmacher, “Tight Thresholds for the Pure Literal Rule,” Technical Report 1997-011, DEC/SRC, 1997.
[22] A. Shwartz and A. Weiss, Large Deviations for Performance Analysis. Chapman&Hall, 1995.
[23] G.D. Stamoulis and J.N. Tsitsiklis, “The Efficiency of Greedy Routing in Hypercubes and Butterflies,” IEEE Tran. Comm., vol. 42, no. 11, pp. 3,051-3,061, 1994.
[24] D. Towsley and R. Mirchandaney, “The Effect of Communication Delays on the Performance of Load Balancing Policies in Distributed Systems,” Proc. Second Int'l MCPR Workshop, pp. 213-226, 1988.
[25] R. Weber, “On the Optimal Assignment of Customers to Parallel Servers,” J. Applied Probability, vol 15, pp. 406-413, 1978.
[26] W. Whitt, “Deciding Which Queue to Join: Some Counterexamples,” Operations Research, vol 34, pp. 55-62, 1986.
[27] W. Winston, “Optimality of the Shortest Line Discipline,” J. Applied Probability, vol 14, pp. 181-189, 1977.
[28] N.C. Wormald, “Differential Equations for Random Processes and Random Graphs,” Annals Applied Probability, vol 5, pp. 1,217-1,235, 1995.
[29] N.D. Vvedenskaya, R.L. Dobrushin, and F.I. Karpelevich, “Queueing System with Selection of the Shortest of Two Queues: An Asymptotic Approach,” Problems of Information Transmission, vol. 32, pp. 15-27, 1996.
Index Terms:
Load balancing, stale information, old information, queuing theory, large deviations.
Citation:
Michael Mitzenmacher, "How Useful Is Old Information?," IEEE Transactions on Parallel and Distributed Systems, vol. 11, no. 1, pp. 6-20, Jan. 2000, doi:10.1109/71.824633