This Article 
 Bibliographic References 
 Add to: 
The NUMA with Clusters of Processors for Parallel Join
July-August 1997 (vol. 9 no. 4)
pp. 653-660

Abstract—Recently, a number of hybrid systems have been proposed to combine the advantages of shared nothing and shared everything concepts for computing relational join operations. Most of these proposed systems, however, presented a few analytical results and have produced limited or no implementations on actual multiprocessors. In this paper, we present a parallel join algorithm with load-balancing for a hybrid system that combines both shared-nothing and shared-everything architectures. We derive an analytical model for the join algorithm on this architecture and validate it using both hardware/software simulations and actual experimentations. We study the performance of the join on the hybrid system for a wide range of system parameter values. We conclude that the hybrid system outperforms both shared-nothing and shared-everything architectures.

[1] D.J. DeWitt and R. Gerber, "Multiprocessor Hash-Based Join Algorithms," Proc. 11th Int'l Conf. Very Large Data Bases, pp. 151-164, 1985.
[2] D.J. DeWitt, S. Ghandeharizadeh, and D. Scheider, "A Performance Analysis of the GAMMA Database Machine," Proc. ACM Special Interest Group Management Data, pp. 350-360, 1988.
[3] C.K. Baru and O. Frieder,“Database operations in a cube-connected multiprocessor system,” IEEE Trans. on Computers, vol. 38, no. 6, pp. 920-927, June 1989.
[4] D.J. DeWitt,S. Ghandeharizadeh,D.A. Schneider,A. Bricker,H.I. Hsiao,, and R. Rasmussen,“The gamma database machine project,” IEEE Trans. on Knowledge and Data Engineering, vol. 2, no. 1, pp. 44-62, Mar. 1990.
[5] M.C. Taylor, "Parallel Multi-Join Algorithms for Main Memory Databases," Proc. Int'l Conf. Parallel Processing, 1989.
[6] H. Lu,K.L. Tan,, and M.-C. Shan,“Hash-based join algorithms for multiprocessor computers with shared memory,” Proc. 16th Int’l Conf. Very Large Data Bases, pp. 198-209, Aug. 1990.
[7] V. Deshpande and P.A. Larson, "The Design and Implementation of a Parallel Join Algorithm for Nested Relations on Shared-Memory Multiprocessors," IEEE Trans. Knowledge and Data Eng., pp. 68-77, 1992.
[8] M. Kitsuregawa, S. Tsudaka, and M. Nakano, "Parallel GRACE Hash Join on Shared-Everything Multiprocessor: Implementation and Performance Evaluation on Symmetry S81," IEEE Trans. Knowledge and Data Eng., pp. 256-264, 1992.
[9] D. DeWitt and J. Gray, “Parallel Database Systems: The Future of High-Performance Database Systems,” Comm. ACM, Vol. 35, No. 6, June 1992, pp. 85-98.
[10] P. Valduriez, "Parallel Database Systems: The Case for Shared-Something," Proc. Ninth Int'l Conf. Data Eng., pp. 460-465, 1993.
[11] K.A. Hua, C. Lee, and J.K. Peir, "Interconnecting Shared-Everything Systems for Efficient Parallel Query Processing," Proc. 18th Int'l Conf. Very Large Data Bases, pp. 262-270, 1992.
[12] A. Shatdal and J. Naughton, “Using Shared Virtual Memory for Parallel Join Processing,” Proc. ACM SIGMOD, pp. 119-128, June 1993.
[13] "KSR1 Principles of Operations," Rev. 6.0, tech. report, Kendall Square Research, Waltham, Mass., 1992.
[14] E. Rosti, E. Smirni, T.D. Wagner, A.W. Apon, and L.W. Dowdy, "The KSR1: Experimentation and Modeling of Poststore," tech. report, Dept. of Computer Science, Vanderbilt Univ., Nashville, Tenn., 1993.
[15] R.H. Saavedra, R.S. Gaines, and M. J. Carlton, "Micro Benchmark Analysis of the KSR1," Trans. Computer Systems, pp. 202-213, 1993.
[16] T.H. Dunigan, Kendall Square Multiprocessor: Early Experiences and Performance, Kendall Square Research: Technical Notes, Aug. 1992.
[17] W. Tout and S. Pramanik, "A Distributed Load-Balancing Scheme for Data Parallel Applications," Proc. Int'l Conf. Parallel Processing, pp. II.213-II.216, Aug. 1993.
[18] C. Walton,“Four types of data skew and their effect on parallel join performance,”Univ. of Texas, Tech. Rep. TR-90-12, 1990.
[19] C.B. Walton, A.G. Dale, and R.M. Jenevein, “A Taxonomy and Performance Model of Data Skew Effects in Parallel Joins,” Proc. 17th Conf. Very Large Databases (VLDB), pp. 537-48, Sept. 1991.
[20] D. DeWitt, J. Naughton, D. Schneider, and S. Seshadri,“Practical skew handling in parallel joins,”inProc. 18th Int. Conf. Very Large Databases, Vancouver, B.C., Aug. 1992, pp. 27–40.
[21] G. Zipf, Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology, Addison-Wesley, 1949.
[22] H. Schwetman, "CSIM Users' Guide," tech. report, Micro Electronics and Computer Technology Corp., MCC Tech Report ACT-126-90, Mar. 1990.
[23] D. Bitton,D.J. DeWitt,, and C. Turbyfill,“Benchmarking database systems_A systematic approach,” Proc. Ninth Int’l Conf. Very Large Data Bases, pp. 8-19, Oct. 1983.
[24] C. Larsen, KSR1 Technical Information, Kendall Square Research, personal contact through e-mail, 1993.

Index Terms:
Relational join, parallel processing, shared-nothing architecture, shared-everything architecture, load-balancing.
Sakti Pramanik, Walid R. Tout, "The NUMA with Clusters of Processors for Parallel Join," IEEE Transactions on Knowledge and Data Engineering, vol. 9, no. 4, pp. 653-660, July-Aug. 1997, doi:10.1109/69.617058
Usage of this product signifies your acceptance of the Terms of Use.