This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
On the Performance of Synchronized Programs in Distributed Networks with Random Processing Times and Transmission Delays
September 1994 (vol. 5 no. 9)
pp. 939-950

A synchronizer is a compiler that transforms a program designed to run in a synchronousnetwork into a program that runs in an asynchronous network. The behavior of a simplesynchronizer, which also represents a basic mechanism for distributed computing and forthe analysis of marked graphs, was studied by S. Even and S. Rajsbaum (1990) under theassumption that message transmission delays and processing times are constant. Westudy the behavior of the simple synchronizer when processing times and transmissiondelays are random. The main performance measure is the rate of a network, i.e., theaverage number of computational steps executed by a processor in the network per unittime. We analyze the effect of the topology and the probability distributions of therandom variables on the behavior of the network. For random variables with exponentialdistribution, we provide tight (i.e., attainable) bounds and study the effect of abottleneck processor on the rate.

[1] B. Awerbuch, "Complexity of network synchronization,"J. ACM, vol. 32, no. 4, pp. 804-823, Oct. 1985.
[2] B. Awerbuch, "Reducing complexities of distributed max-flow and breadth-first-search algorithms by means of network synchronization,"Networks, vol. 15, pp. 425-437, 1985.
[3] O. Raiman, "A circumscribed diagnosis engine," inExpert Systems in Engineering Principles and Applications, Vienna, Austria, Sept. 1990,Lecture Notes in Computer Science, LNAI Subseries 462. New York: Springer-Verlag.
[4] O. J. Boxma, "Sojourn times in cyclic queues: The influence of the slowest server," in G. Iazeolla, P. J. Courtois, and O. J. Boxma, Eds.,Computer Performance and Reliability. New York: Elsevier (North-Holland), 1988.
[5] F. Baccelli, "Ergodic theory of stochastic Petri networks," Raport INRIA 1037, 1989.
[6] V. C. Barbosa and E. Gafni, "Concurrency in heavily loaded neighborhood-constrained systems,"ACM Trans. Programming Languages Syst., vol. 11, pp. 562-584, Oct. 1989.
[7] F. Baccelli, P. Konstantopoulos, "Estimates of cycle times in stochastic Petri nets," in I. Karatzas, Ed.,Proc. Rutgers Conf. Stochastic Analysis. New York: Springer-Verlag, 1991.
[8] F. Baccelli and Z. Liu, "Comparison properties of stochastic decision-free Petri nets," INRIA Res. Raport 1433, May 1991.
[9] F. Baccelli and A. Makowski, "Queueing models for systems with synchronization constraints,"Proc. IEEE, vol. 77, pp. 138-162, Jan. 1989.
[10] F. Baccelli, A. M. Makowski, and A. Shwartz, "The fork-join queue and related systems with synchronization constraints: Stochastic ordering and computable bounds,"Advances in Applied Probability, vol. 21, pp. 629-660, Sept. 1989.
[11] F. Baccelli, W. A. Massey, and D. Towsley, "Acyclic fork/join queueing networks,"J. ACM, vol. 36, pp. 615-642, 1989.
[12] P. Berman and J. Simon, "Investigations of fault-tolerant networks of computers,"Proc. 20th ACM STOC, 1988.
[13] D. P. Bertsekas and J. N. Tsitsiklis,Parallel and Distributed Computations. Englewood Cliffs, NJ: Prentice-Hall, 1989.
[14] K. L. Chung,Markov Chains with Stationary Transition Probabilities, 2nd ed. New York: Springer-Verlag, 1967.
[15] F. Commoner, A. W. Holt, S. Even, and A. Pnueli, "Marked directed graphs,"J. Comput. Syst. Sci., vol. 5, no. 5, Oct. 1971.
[16] K. M. Chandy and L. Lamport, "Distributed snapshots: Determining global states of distributed systems,"ACM Trans. Comput. Syst., vol. 3, no. 1, pp. 63-75, Feb. 1985.
[17] H. A. David,Order Statistics. New York: Wiley, 1970.
[18] S. Even and S. Rajsbaum, "Lack of global clock does not slow down the computation in distributed networks," Tech. Rep. 522, Dept. of Comput. Sci., Technion, Haifa, Israel, Oct. 1988, to appear inMathemat. Syst. Theory. (first part of this paper appears under the title, "Unison in distributed networks" in R. M. Capocelli, Ed.,Sequences: Combinatorica, Compression, Security, and Transmission. New York: Springer-Verlag, 1990.)
[19] S. Even and S. Rajsbaum, "The use of a synchronizer yields maximum rate in distributed networks,"Proc. 22nd ACM STOC, 1990.
[20] J. Garofalakis, S. Rajsbaum, P. Spirakis, and B. Tampakas, "Tentative and definite distributed computations: An optimistic approach to network synchronization," in Proc. 6th Int. Workshop on Distributed Algorithms, Haifa, Israel, Nov. 1992,Lecture Notes in Computer Science 647. New York: Springer-Verlag, pp. 110-119 (to appear inTheoretical Comput. Sci.).
[21] J. F. C. Kingman, "Subadditive ergodic theory,"Ann. Prob., vol. 1, no. 6, pp. 883-909, 1973.
[22] M. A. Marsan, "Stochastic Petri nets: An elementary introduction," inAdvances in Petri Nets 1989, Lecture Notes in Computer Science 424. New York: Springer-Verlag, 1989, pp. 1-29.
[23] J. Malka, S. Moran, and S. Zaks, "Analysis of a distributed scheduler for communication networks," Tech. Rep. 495, Dept. of Comput. Sci., Technion, Haifa, Israel, Feb. 1988 (also inLecture Notes on Computer Science 319. New York: Springer Verlag, 1988, pp. 351-360).
[24] M. K. Molloy, "Performance analysis using stochastic Petri nets,"IEEE Trans. Comput., vol. 31, no. 9, pp. 913-917, Sept. 1982.
[25] M. K. Molloy, "Fast bounds for stochastic Petri nets,"International Workshop on Timed Petri Nets, Torino, Italy, July 1985, pp. 244-249.
[26] Y. Ofek and I. Gopal, "Generating a global clock in a distributed system," IBM Res. Rep., 1987.
[27] D. Peleg and A. A. Schaffer, "Graph spanners,"J. Graph Theory, vol. 13, pp. 99-116, 1989.
[28] D. Peleg and J. D. Ullman, "An optimal synchronizer for the hypercube,"Soc. Indust. and Appl. Math., vol. 18, no. 4, pp. 740-747, Aug. 1989.
[29] N. Pekergin and J.-M. Vincent, "Stochastic bounds on execution times of task graphs," Rep. EHEI, 1989.
[30] A. L. Rosenberg, "Shuffle-oriented interconnection networks," COINS Tech. Rep. 88-84, Univ. of Massachusetts, 1988.
[31] S. Rajsbaum, "Upper and lower bounds for stochastic marked graphs,"Inform. Processing Lett., vol. 49, pp. 291-295, 1994.
[32] Y. Malka and S. Rajsbaum, "Analysis of distributed algorithms based on recurrence relations,"Proc. 5th Workshop on Distributed Algorithms on Graphs (WDAG-5), 1991, pp.
[33] S. M. Ross,Stochastic Processes. New York: Wiley, 1983.
[34] D. Stoyan,Comparison Methods for Queues and Other Stochastic Models, (English translation). New York: Wiley, 1984.
[35] J. G. Shanthikumar and D. D. Yao, "The effect of increasing service rates in a closed queueing network,"J. Appl. Prob., vol. 23, pp. 474-483, 1986.

Index Terms:
Index Termsprogram compilers; performance evaluation; parallel programming; message passing;synchronisation; graph theory; synchronized programs; distributed networks; randomprocessing times; transmission delays; synchronizer; compiler; synchronous network;asynchronous network; distributed computing; marked graphs; message transmissiondelays; processing times; performance measure; computational steps; probabilitydistributions; random variables; exponential distribution; bottleneck processor
Citation:
S. Rajsbaum, M. Sidi, "On the Performance of Synchronized Programs in Distributed Networks with Random Processing Times and Transmission Delays," IEEE Transactions on Parallel and Distributed Systems, vol. 5, no. 9, pp. 939-950, Sept. 1994, doi:10.1109/71.308532
Usage of this product signifies your acceptance of the Terms of Use.