Take a look at the new version of this page: [ beta version ]. Tell us what you think.

ABSTRACT

It is generally accepted that On-Line Transaction Processing (OLTP) systems benefit from large database memory buffers. As enterprise database systems become larger and more complex, hardware vendors are building increasingly large systems capable of supporting huge memory configurations. Database vendors in turn are developing buffer schemes to exploit this physical memory. How much will these developments benefit OLTP workloads? Through empirical studies on databases sized comparably to those seen in the real-world, this paper presents the characteristics of an industry-standard OLTP benchmark as memory buffer size changes. We design the experiments to investigate how the database size, the buffer size and the number of CPUs impact performance, in particular the throughput and the buffer hit rate on Symmetric Multiprocessor Systems. The relationships of these major database attributes are plotted and key observations are summarized. We discuss how these relationships change as the number of CPUs changes. We further quantify the relationships: 1) between database buffer data hit rate, buffer size and database size, 2) between throughput, buffer data hit rate and database size and 3) between throughput and number of CPUs. Algorithms, rules-of-thumb and examples are presented for predicting performance, sizing memory and making trade-offs between adding more memory and increasing the number of CPUs.

REFERENCES

• Amda67: Amdahl, G.,"Validity of the Single Processor Approach to Achieving Large Scale Computing Capabilities", AFIPS Conf. Proceedings, 1967
• Arti91: Artis, H.,"Quantifying Multiprocessor Overheads", Proceedings of CMG Conference, 1991.
• ChDe85: Chou, H. and DeWitt, D., "An Evaluation of Buffer Management Strategies for Relational Database Systems", Prigs of VLDB, 1985.
• DB2Adm: "IBM DB2 Administration Guide for Common Sewers", IBM Corp. 1995.
• DB2Info: "IBM DB2 Information and Concepts Guide for Common Sewers", IBM Corp. 1995
• DaTo90: Asit Dan , Don Towsley, An approximate analysis of the LRU and FIFO buffer replacement schemes, Proceedings of the 1990 ACM SIGMETRICS conference on Measurement and modeling of computer systems, p.143-152, April 1990, Univ. of Colorado, Boulder, Colorado, United States
• DaYC93: Asit Dan , Philip S. Yu , Jen-Yao Chung, Database Access Characterization for Buffer Hit Prediction, Proceedings of the Ninth International Conference on Data Engineering, p.134-143, April 19-23, 1993
• DataD: "DamDesk Statistic Guide", Data Description, Inc.
• DrSm81: Draper, N. and Smith, H., "Applied Regression Analysis", seexmd edition, John Wtley& Sons, 1981.
• EfWH84: Wolfgang Effelsberg , Theo Haerder, Principles of database buffer management, ACM Transactions on Database Systems (TODS), v.9 n.4, p.560-595, Dec. 1984 [doi>10.1145/1994.2022]
• Gray93: Jim Gray, Benchmark Handbook: For Database and Transaction Processing Systems, Morgan Kaufmann Publishers Inc., San Francisco, CA, 1992
• Gunt93: Gunther, N., "A Simple Capacity Model for Massivdy Parallel Transaction Systems", proceedings of CMG, 1993.
• Gunt96: Gunther, N., "Understanding The MP Effect: Multiprocessing in Pictures", to appear in Proee, edings of CMG, 1996.
• LZGS84: Edward D. Lazowska , John Zahorjan , G. Scott Graham , Kenneth C. Sevcik, Quantitative system performance: computer system analysis using queueing network models, Prentice-Hall, Inc., Upper Saddle River, NJ, 1984
• LeDi93: Scott T. Leutenegger , Daniel Dias, A modeling study of the TPC-C benchmark, Proceedings of the 1993 ACM SIGMOD international conference on Management of data, p.22-31, May 25-28, 1993, Washington, D.C., United States
• McGa95: McGalliard, J'., "Case Study of Table-Top Sizing with Workload-Specific Estimates of the Multiprocessor Effect", proceeAings of CMG Conference, 1995.
• MoHa92: C. Mohan , Don Haderle , Bruce Lindsay , Hamid Pirahesh , Peter Schwarz, ARIES: a transaction recovery method supporting fine-granularity locking and partial rollbacks using write-ahead logging, ACM Transactions on Database Systems (TODS), v.17 n.1, p.94-162, March 1992 [doi>10.1145/128765.128770]
• Ston81: Michael Stonebraker, Operating system support for database management, Communications of the ACM, v.24 n.7, p.412-418, July 1981 [doi>10.1145/358699.358703]
• TPC95: TIC Benchmark C Standard Specification Revision 3.0", Transaction Processing Performance Council, February 15, 1995.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.2 DATABASE MANAGEMENT

Additional Classification:
  G. Mathematics of Computing
  G.4 MATHEMATICAL SOFTWARE
      Subjects: Algorithm design and analysis

  H. Information Systems
  H.2 DATABASE MANAGEMENT
      H.2.4 Systems
          Subjects: Transaction processing
  H.3 INFORMATION STORAGE AND RETRIEVAL

  K. Computing Milieux
  K.6 MANAGEMENT OF COMPUTING AND INFORMATION SYSTEMS
      K.6.2 Installation Management
          Subjects: Benchmarks

General Terms:
Algorithms, Design, Experimentation, Management, Measurement, Performance, Standardization, Theory