This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Semi-Distributed Load Balancing for Massively Parallel Multicomputer Systems
October 1991 (vol. 17 no. 10)
pp. 987-1004

A semidistributed approach is given for load balancing in large parallel and distributed systems which is different from the conventional centralized and fully distributed approaches. The proposed strategy uses a two-level hierarchical control by partitioning the interconnection structure of a distributed or multiprocessor system into independent symmetric regions (spheres) centered at some control points. The central points, called schedulers, optimally schedule tasks within their spheres and maintain state information with low overhead. The authors consider interconnection structures belonging to a number of families of distance transitive graphs for evaluation, and, using their algebraic characteristics, show that identification of spheres and their scheduling points is in general an NP-complete problem. An efficient solution for this problem is presented by making exclusive use of a combinatorial structure known as the Hadamard matrix. The performance of the proposed strategy has been evaluated and compared with an efficient fully distributed strategy through an extensive simulation study. The proposed strategy yielded much better results.

[1] I. Ahmad, A. Ghafoor, and K. Mehrotra, "Performance prediction for distributed load balancing in multicomputer systems," School of Comput. and Inform. Sci., Syracuse Univ., Tech. Rep. no. SU-CIS-91- 12, Apr. 1991.
[2] W. C. Athas and C. L. Seitz, "Multicomputers: message-passing concurret computers,"IEEE Computer, pp. 9-24, Aug. 1988.
[3] E. Bannai and T. Ito,Algebraic Combinatorics and Association Schemes. Menlo Park, CA: Benjamin-Cummings, 1984.
[4] A. Barak and A. Shiloh, "A distributed load-balancing policy for a multicomputer,"SPAE, vol. 15, pp. 901-913, 1985.
[5] K. M. Baumgartner, R. Kling, and B. Wah, "A global load balancing strategy for a distributed system," inProc. of Int. Conf. on Future Trends in Distributed Comput. Syst. (Hong Kong), 1988, pp. 93-102.
[6] S. H. Bokhari, "Dual processor scheduling with dynamic reassignments,"IEEE Trans. Software Eng., vol. SE-5, pp. 341-349, July 1979.
[7] N. S. Bowen, C. Nikolaou, and A. Ghafoor, "On the assignment problem of arbitrary process systems to heterogeneous distributed computer systems,"IEEE Trans. Computers, to be published.
[8] R. M. Bryant and R. A. Finkel, "A stable distributed scheduling algorithm," inProc. 2nd Int. Conf. on Distributed Comput. Syst., 1981, pp. 314-323.
[9] T. L. Casavant and J. G. Kuhl, "Analysis of three dynamic load-balancing strategies with varying global information requirements," inProc. 7th Int. Conf. on Distributed Comput. Syst., 1987, pp. 185-192.
[10] T. F. Chan, "Hierarchical algorithms and architectures for parallel scientific computing," inProc. Int. Conf. on Supercomput., 1990, pp. 318-329.
[11] Y.-C. Chow and W. H. Kohler, "Models for dynamic load balancing in homogeneous multiple processor systems,"IEEE Trans. Computers, vol. C-36, pp. 667-679, May 1982.
[12] S. Chowdhury, "The Greedy load sharing algorithm,"J. Parallel and Distributed Comput., No. 9, pp. 93-99, May 1990.
[13] "Strategic computing: new generation computing technology," Defence Advance Res. Project Agency, Arlington, VA, Oct. 1983.
[14] D. Eager, E. Lazowska, and J. Zahorjan, "Adaptive load sharing in homogeneous distributed systems,"IEEE Trans. Software Eng., vol. SE-12, no. 5, pp. 662-675, May 1986.
[15] D. L. Eager, E. Lazowska, and J. Zahorjan, "A comparison of receiver-initiated and sender-initiated adaptive loading,"Perform. Eval., vol. 6, 1986.
[16] K. Efe and B. Groselj, "Minimizing control overhead in adaptive load sharing," inProc. 9th Int. Conf. on Distributed Comput. Syst., 1989, pp. 307-315.
[17] A. K. Ezzat, R. D. Bergerson, and J. L. Pokoski, "Task allocation heuristics for distributed computing systems," inProc. 6th Int. Conf. on Distributed Comput. Syst., 1986, pp. 337-346.
[18] D. Ferguson, Y. Yemini, and C. Nickolaou, "Microeconomic algorithms for load balancing in distributed computer systems," inProc. 8th Int. Conf. on Distributed Comput. Syst., 1988, pp. 491-499.
[19] G. C. Fox, A. Kolawa, and R. Williams, "The implementation of a dynamic load balancer," inProc. SIAM Hypercube Multiprocess. Conf., 1987, pp. 114-121.
[20] M. Foxet al., Solving Problems on Concurrent Processors, vol. 1. Englewood Cliffs, NJ: Prentice-Hall, 1988.
[21] A. Ghafoor and T. Bashkow, "A study of odd graphs as fault-tolerant interconnection networks,"IEEE Trans. Computers, vol. 40, no. 2, pp. 225-232, Feb. 1991.
[22] A. Ghafoor, T. Bashkow, and I. Ghafoor, "Bisectional fault-tolerant communication architecture for supercomputer systems,"IEEE Trans. Computers, vol. 38, pp. 1425-1446, Oct. 1989.
[23] A. Ghafoor and P. B. Berra, "An efficient communication structure for distributed commit protocols,"IEEE J. Selected Areas Commun., vol. 7, pp. 375-389, Apr. 1989.
[24] A. Ghafoor, S. Sheikh, and P. Sol, "Distance-transitive graphs for fault-tolerant multiprocessor systems," inProc. 1989 Int. Conf. on Parallel Process., 1989, pp. I-176-I-179.
[25] A. Ha'c and T. J. Johnson, "Sensitivity study of the load balancing algorithm in a distributed system,"J. Parallel and Distributed Comput., pp. 85-89, Oct. 1989.
[26] M. Hall, Jr.,Combinatorial Theory, 2nd ed. New York: Wiley, 1986.
[27] L. V. Kale, "Comparing the performance of two dynamic load distribution methods," inProc. Int. Conf. on Parallel Process., 1988, pp. 8-12.
[28] F. C. H. Lin and R. M. Keller, "Gradient model: a demand-driven load balancing scheme," inProc. 6th Int. Conf. on Distributed Comput. Systems, 1986, pp. 329-336.
[29] M. Livny and M. Melmen, "Load balancing in homogeneous broadcast distributed systems," inProc. Computer Network Perform. Symp., 1982, pp. 47-55.
[30] A. M. McLoughlin, "The complexity of computing the covering radius of a code,"IEEE Trans. Inform. Theory, vol. IT-30, pp. 800-804, Nov. 1984.
[31] F. J. MacWilliams and N. J. A. Sloane,The Theory of Error-Correcting Codes, vols. I and II. New York: North-Holland, 1977.
[32] R. Mirchandancy, D. Towsly, and J. A. Stankovic, "Analysis of effect of delays on load sharing,"IEEE Trans. Computers, vol. 38, pp. 1513-1525, Nov. 1989.
[33] L. M. Ni and K. Hwang, "Optimal load balancing in a multiple processor system with many job classes,"IEEE Trans. Software Eng., vol. SE-11, pp. 491-496, May 1985.
[34] L. M. Ni, C. Xu, and T. B. Gendreau, "A distributed drafting algorithm for load balancing,"IEEE Trans. Software Eng., vol. SE-11, no. 10, pp. 1153-1161, Oct. 1985.
[35] K. Ramamritham, J. A. Stankovic, and W. Zhao, "Distributed scheduling of tasks with deadlines and resource requirements,"IEEE Trans. Computers, vol. 38, no. 8, pp. 1110-1123, Aug. 1989.
[36] D. A. Reed and R. M. Fujimoto,Multicomputer Networks, Message-Based Parallel Processing. Cambridge, MA: MIT Press, 1987.
[37] K. G. Shin and Y.-C. Chang, "Load sharing in distributed real-time systems with state-change broadcasts,"IEEE Trans. Computers, vol. 38, pp. 1124-1142, Aug. 1989.
[38] P. Solé, A. Ghafoor, and S. Sheikh, "The covering radius of Hadamard code in odd graphs,"Appl. Discrete Math., to be published.
[39] J. A. Stankovic and I. S. Sidhu, "An adaptive bidding algorithm for processes, clusters and distributed groups," inProc. 4th Int. Conf. on Distributed Comput. Syst., 1984, pp. 49-59.
[40] L. J. Stochmeyer and V. V. Vazirani, "NP-completeness of some generalization of the maximum matching problems,"Inform. Proc. Lett., vol. 15, pp. 14-19, 1982.
[41] A. N. Tantawi and D. Towsley, "Optimal static load balancing in distributed computer systems,"J. ACM, pp. 445-465, Apr. 1985.
[42] A. M. Van Tilborg and L. D. Wittie, "Wave scheduling-decentralized task scheduling of task forces in multicomputers,"IEEE Trans. Computers, vol. C-33, pp. 835-844, Sept. 1984.
[43] Y.-T. Wang and R. J. T. Morris, "Load sharing in distributed systems,"IEEE Trans. Computers, vol. C-34, pp. 204-217, Mar. 1985.
[44] J. Xu and K. Hwang, "Heuristic methods for dynamic load balancing in a message-passing supercomputer, " inProc. Supercomput. '90, Nov. 1990, pp. 888-897.
[45] S. Zhou, "Performance studies of dynamic load balancing in distributed systems," Ph.D. dissert., Univ. California, Berkeley, Sept. 1987.

Index Terms:
massively parallel multicomputer systems; semidistributed approach; load balancing; distributed systems; fully distributed approaches; two-level hierarchical control; interconnection structure; multiprocessor system; independent symmetric regions; state information; interconnection structures; distance transitive graphs; scheduling points; NP-complete problem; combinatorial structure; Hadamard matrix; fully distributed strategy; simulation study; computational complexity; multiprocessor interconnection networks; parallel architectures; parallel machines; scheduling
Citation:
I. Ahmad, A. Ghafoor, "Semi-Distributed Load Balancing for Massively Parallel Multicomputer Systems," IEEE Transactions on Software Engineering, vol. 17, no. 10, pp. 987-1004, Oct. 1991, doi:10.1109/32.99188
Usage of this product signifies your acceptance of the Terms of Use.