Optimal plan search in a rule-based query optimizer

  • Ming-Chien Shan
Expert System Approaches To Databases

DOI: 10.1007/3-540-19074-0_49

Part of the Lecture Notes in Computer Science book series (LNCS, volume 303)
Cite this paper as:
Shan MC. (1988) Optimal plan search in a rule-based query optimizer. In: Schmidt J.W., Ceri S., Missikoff M. (eds) Advances in Database Technology—EDBT '88. EDBT 1988. Lecture Notes in Computer Science, vol 303. Springer, Berlin, Heidelberg

Abstract

This paper describes an optimal plan search strategy adopted in a rule-based query optimizer. Instead of attempting to search for the optimal plan directly, an initial plan is first generated based upon a set of heuristic rules. Depending upon the application, the initial plan may be used either as the final plan or as a base in a subsequent search. A new concept — clustering degree of an index — is introduced to better model the I/O costs of index scans. This new statistical information facilitates the formulation of the rules. An exhaustive search based upon the A* algorithm is then invoked to guarantee the optimal property of the plan. A lower bound value is derived and used as the estimation of ”remaining distance” required in the A* algorithm. Noteworthy features of our approach include the capability for dynamic control of exhaustive search for an optimal plan, and on-line performance monitoring/tuning. The preliminary results lead us to believe that the rule-based approach is a promising one to face the new challenges of the optimizer, as created by the requirements of supporting diversified applications.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag 1988

Authors and Affiliations

  • Ming-Chien Shan
    • 1
  1. 1.Hewlett-Packard LaboratoriesPalo Alto

Personalised recommendations