Victor F. Nicola
ABSTRACT
The CLOCK algorithm is a popular buffer replacement algorithm because of its simplicity and its ability to approximate the performance of the Least Recently Used (LRU) replacement policy. The Generalized Clock (GCLOCK) buffer replacement policy uses a circular buffer and a weight associated with each page brought in buffer to decide on which page to replace. We develop an approximate analysis for the GCLOCK policy under the Independent Reference Model (IRM) that applies to many database transaction processing workloads. We validate the analysis for various workloads with data access skew. Comparison with simulations shows that in all cases examined the error is extremely small (less than 1%). To show the usefulness of the model we apply it to a Transaction Processing Council benchmark A (TPC-A) like workload. If knowledge of the different data partitions in this workload is assumed, the analysis shows that, with appropriate choice of weights, the performance of the GCLOCK algorithm can be better than the LRU policy. Performance very close to that for optimal (static) buffer allocation can be achieved by assigning sufficiently high weights, and can be implemented with a reasonably low overhead. Finally, we outline how the model can be extended to capture the effect of page invalidation in a multinode system.
- 1.Belady, L. A., "A Study of Replacement Algorithms for a Virtual Storage Computer", IBM Systems Journal, Vol.5, No.2, 1966, pp. 78-101.Google Scholar
Digital Library
- 2.Cart, R. W., and J. L. Hennessy, "WSClock - A Simple and Effective Algorithm for Virtual Memory Management", A CM SIGMETRICS, 1981, pp. 87-95. Google ScholarDigital Library
- 3.Casas, R. I., and K. C. Sevcik, "A Buffer Management Model for Use in Predicting Overall Database System Performance", 5th International Conference on Data Engineering, Los Angeles, CA, Feb. 1989, pp. 463-469. Google ScholarDigital Library
- 4.Chou, H. T., and D. J. Dewitt, "An Evaluation of Buffer Management Strategies for Relational Database Systems", 11th International Conference on Very Large Databases, Stockholm, Sweden, 1985, pp. 127-141.Google Scholar
- 5.Corbato, F. J., "A Paging Experiment with the Multics System", MIT Pro3ect MAC Report MAC-M-384, May 1968.Google Scholar
- 6.Dan, A., and D. Towsley, "An Approximate Analysis of the LRU and FIFO Buffer Replacement Schemes," A CM SIGMETRICS, Denver, CO, May 1990, pp. 143- 152. Google ScholarDigital Library
- 7.Dan, A., D. M. Dins, and P. S. Yu, "The Effect of Skewed Data Access on Buffer Hits and Data Contention in a Data Sharing Environment," 16th International Conference on Very Large Databases, Brisbane, Australia, Aug. 1990. Google ScholarDigital Library
- 8.Dan, A., D. M. Dins, and P. S. Yu, "Modelling a Hierarchical Buffer for the Data Sharing Environment", ACM SIGMETRICS, San Diego, CA, May 1991. Google ScholarDigital Library
- 9.Dan, A., and P. S. Yu, "Performance Comparisons of Buffer Coherency Policies," 11th International Con- }erence on Distributed Computing Systems, Arlington, TX, May 1991.Google ScholarCross Ref
- 10.Dan, A., P. S. Yu and J. Y. Chung, "Characterization of Database Access Skew of a Transaction Processing Environment", IBM Research Report RC 17436, Yorktown Heights, NY, Sept. 1991.Google Scholar
- 11.Effelsberg, W. and T. Haerder, "Principles of Database Buffer Management", A CM Trans. Database Systems, Vol. 9, No. 4, Dec. 1984, pp. 560-595. Google ScholarDigital Library
- 12.Kearns, J. P., and S. Defazio, "Diversity in Database Reference Behavior", Performance Evaluatwn Review, Vol. 17, No. 1, pp. 11-19. Google ScholarDigital Library
- 13.King, W. F., "Analysis of Paging Algorithms", Proc. IFIP Congress, Ljublanjana, Yugoslavia, Aug. 1971, pp. 485-490.Google Scholar
- 14.Nicola, V. F., A. Dan, and D. M. Dins, "Analysis of the Generahzed Clock Buffer Replacement Scheme for Database Transaction Processing", IBM Research Report RC 17225, Yorktown Heights, NY, Sept. 1991.Google Scholar
- 15.Reiter, A., "A Study of Buffer Management Pohcies for Data Management Systems", Technical Summary Report 1619, Mathematics Research Center, Univ. of Wisconsin, Madison, March 1976.Google Scholar
- 16.Sacco, G. M., and M. Sckolnick, "Buffer Management in Relational Database Systems", A CM Trans. Database Systems, Vol. 11, No. 4, Dec. 1986, pp. 473-498. Google ScholarDigital Library
- 17.Smith, A. J., "Sequentiahty and Prefetching in Database Systems", A CM Trans. Database Systems, Vol. 3, No. 3, Sept. 1978, pp. 223-247. Google ScholarDigital Library
- 18."TPC Benchmark A Standard Specification", Transaction Processing Council, 1989.Google Scholar
Index Terms
- Analysis of the generalized clock buffer replacement scheme for database transaction processing
Recommendations
Analysis of the generalized clock buffer replacement scheme for database transaction processing
The CLOCK algorithm is a popular buffer replacement algorithm because of its simplicity and its ability to approximate the performance of the Least Recently Used (LRU) replacement policy. The Generalized Clock (GCLOCK) buffer replacement policy uses a ...
Counter-Based Cache Replacement and Bypassing Algorithms
Recent studies have shown that in highly associative caches, the performance gap between the Least Recently Used (LRU) and the theoretical optimal replacement algorithms is large, motivating the design of alternative replacement algorithms to improve ...
Dueling CLOCK: adaptive cache replacement policy based on the CLOCK algorithm
DATE '10: Proceedings of the Conference on Design, Automation and Test in EuropeWe consider the problem of on-chip L2 cache management and replacement policies. We propose a new adaptive cache replacement policy, called Dueling CLOCK (DC), that has several advantages over the Least Recently Used (LRU) cache replacement policy.
LRU'...
Comments