Exploiting process lifetime distributions for dynamic load balancing

We measure the distribution of lifetimes for UNIX processes and propose a functional form that fits this distribution well. We use this functional form to derive a policy for preemptive migration, and then use a trace-driven simulator to compare our proposed policy with other preemptive migration policies, and with a non-preemptive load-balancing strategy. We find that, contrary to previous reports, the performance benefits of preemptive migration are significantly greater than those of non-preemptive migration, even when the memory-transfer cost is high. Using a model of migration costs representative of current systems, we find that preemptive migration reduces the mean delay (queueing and migration) by 35% -- 50%, compared to non-preemptive migration.

[1]  Yongbing Zhang,et al.  A performance comparison of adaptive and static load balancing in heterogeneous distributed systems , 1995, Proceedings of Simulation Symposium.

[2]  Edward R. Zayas,et al.  Attacking the process migration bottleneck , 1987, SOSP '87.

[3]  Yung-Terng Wang,et al.  Load Sharing in Distributed Systems , 1985, IEEE Transactions on Computers.

[4]  Fred Douglis,et al.  Transparent process migration: Design alternatives and the sprite implementation , 1991, Softw. Pract. Exp..

[5]  Thomas Kunz,et al.  The Influence of Different Workload Descriptions on a Heuristic Load Balancing Scheme , 1991, IEEE Trans. Software Eng..

[6]  Shyamal K. Chowdhury.  The Greedy Load Sharing Algorithm , 1990, J. Parallel Distributed Comput..

[7]  C. G. Rommen.  The probability of load balancing success in a homogeneous network , 1991 .

[8]  R. Larsen.  An introduction to mathematical statistics and its applications / Richard J. Larsen, Morris L. Marx , 1986 .

[9]  Teunis J. Ott,et al.  Load-balancing heuristics and process behavior , 1986, SIGMETRICS '86/PERFORMANCE '86.

[10]  Arif Ghafoor,et al.  Performance prediction of distributed load balancing on multicomputer systems , 1991, Proceedings of the 1991 ACM/IEEE Conference on Supercomputing (Supercomputing '91).

[11]  Erol Gelenbe,et al.  Dynamic load balancing in distributed systems , 1994, Proceedings of International Workshop on Modeling, Analysis and Simulation of Computer and Telecommunication Systems.

[12]  Donald F. Towsley,et al.  Imbedding gradient estimators in load balancing algorithms , 1988, [1988] Proceedings. The 8th International Conference on Distributed.

[13]  Dejan S. Milojicic.  Load distribution - implementation for the Mach microkernel , 1994 .

[14]  Donald F. Towsley,et al.  Adaptive Load Sharing in Heterogeneous Distributed Systems , 1990, J. Parallel Distributed Comput..

[15]  Phillip Krueger,et al.  A comparison of preemptive and non-preemptive load distributing , 1988, [1988] Proceedings. The 8th International Conference on Distributed.

[16]  Robert F. Rosin.  Determining a computing center environment , 1965, CACM.

[17]  Keith A. Lantz,et al.  Preemptable remote execution facilities for the V-system , 1985, SOSP 1985.

[18]  David A. Patterson,et al.  Computer Architecture: A Quantitative Approach , 1969 .

[19]  Songnian Zhou.  Performance Studies of Dynamic Load Balancing in Distributed Systems , 1987 .

[20]  Anna Hác,et al.  Dynamic Load Balancing in a Distributed System Using a Sender-Initiated Algorithm , 1987, SIGMETRICS Perform. Evaluation Rev..

[21]  Dick H. J. Epema,et al.  An analysis of decay-usage scheduling in multiprocessors , 1995, SIGMETRICS '95/PERFORMANCE '95.

[22]  Cauligi S. Raghavendra,et al.  A state-aggregation method for analyzing dynamic load-balancing policies , 1993, [1993] Proceedings. The 13th International Conference on Distributed Computing Systems.

[23]  David J. Evans,et al.  Dynamic Load Balancing Using Task-Transfer Probabilities , 1993, Parallel Comput..

[24]  T. Kunz.  The Innuence of Diierent Workload Descriptions on a Heuristic Load Balancing Scheme the Innuence of Diierent Workload Descriptions on a Heuristic Load Balancing Scheme , 2007 .

[25]  Raphael A. Finkel,et al.  Designing a process migration facility: the Charlotte experience , 1989, Computer.

[26]  Raphael A. Finkel,et al.  A Stable Distributed Scheduling Algorithm , 1981, IEEE International Conference on Distributed Computing Systems.

[27]  Mark Nuttall.  Survey of systems providing process or object migration , 1994 .

[28]  K. M. Ahmed,et al.  LSBATCH: A Distributed Load Sharing Batch System , 1993 .

[29]  LivnyMiron,et al.  Load balancing in homogeneous broadcast distributed systems , 1982 .

[30]  Anurag Kumar,et al.  Adaptive Optimal Load Balancing in a Nonhomogeneous Multiserver System with a Central Job Scheduler , 1990, IEEE Trans. Computers.

[31]  Marvin Theimer,et al.  Preemptable remote execution facilities for the V-system , 1985, SOSP '85.

[32]  L EagerDerek,et al.  Adaptive load sharing in homogeneous distributed systems , 1986 .

[33]  Peter B. Galvin,et al.  Operating System Concepts, 4th Ed. , 1993 .

[34]  Anders Svensson.  History, an intelligent load sharing filter , 1990, Proceedings.,10th International Conference on Distributed Computing Systems.

[35]  Miron Livny,et al.  Experience with the Condor distributed batch system , 1990, IEEE Workshop on Experimental Distributed Systems.

[36]  Rakesh Agrawal,et al.  Location Independent Remote Execution in NEST , 1987, IEEE Transactions on Software Engineering.

[37]  Abraham Silberschatz,et al.  Operating System Concepts , 1983 .

[38]  Edward D. Lazowska,et al.  Adaptive load sharing in homogeneous distributed systems , 1986, IEEE Transactions on Software Engineering.

[39]  Miron Livny,et al.  Condor-a hunter of idle workstations , 1988, [1988] Proceedings. The 8th International Conference on Distributed.

[40]  Greg Thiel,et al.  LOCUS operating system, a transparent system , 1991, Comput. Commun..

[41]  Jonathan Walpole,et al.  MPVM: A Migration Transparent Version of PVM , 1995, Comput. Syst..

[42]  Domenico Ferrari,et al.  A Measurement Study of Load Balancing Performance , 1987, IEEE International Conference on Distributed Computing Systems.

[43]  Thomas L. Casavant,et al.  Analysis of Three Dynamic Distributed Load-Balancing Strategies with Varying Global Information Requirements , 1987, ICDCS.

[44]  C. Gary Rommel.  The Probability of Load Balancing Success in a Homogeneous Network , 1991, IEEE Trans. Software Eng..

[45]  Edward D. Lazowska,et al.  The limited performance benefits of migrating active processes for load sharing , 1988, SIGMETRICS '88.

[46]  Amnon Barak,et al.  The MOSIX Distributed Operating System: Load Balancing for UNIX , 1993 .

[47]  Jingwen Wang,et al.  Utopia: A load sharing facility for large, heterogeneous distributed computer systems , 1993, Softw. Pract. Exp..

[48]  Robbert van Renesse,et al.  Experiences with the Amoeba distributed operating system , 1990, CACM.

[49]  Mor Harchol-Balter,et al.  A note on \The Limited Performance Bene ts of Migrating Active Processes for Load Sharing" , 1995 .

[50]  Edward D. Lazowska,et al.  The limited performance benefits of migrating active processes for load sharing , 1988, SIGMETRICS 1988.

[51]  Miron Livny,et al.  Load Balancing in Homogeneous Broadcast Distributed Systems , 1982, SIGMETRICS.

[52]  William E. Weihl,et al.  Lottery scheduling: flexible proportional-share resource management , 1994, OSDI '94.