A Model for Pipelined Query Execution
Annita N. Wilschut, Stephan A. van Gils · University of Twente Research Information · 1993
This paper develops an analytical model for pipelined execution of complex queries. The goal of this research is gaining insight in potential performance gain from the concurrent execution of pipelined relational operations. First, the model is developed in general terms, and after that it is elaborated for selection, unique, and join operations. Finally, it is shown how the model is used to understand the behavior of multi-operation queries. It is shown, how the individual characteristics of operations influence their cooperation in a pipeline. The model increases the understanding of dataflow query execution and can form a basis to generate query optimization heuristics for a parallel DBMS. 1 Introduction During the last years much attention has been paid to the development of parallel DBMSs. Teradata [Ter83], GAMMA [DGS90], Bubba [BAC90], HC16-186 [BrG89] and PRISMA/DB [Ame91,ABF92, WFA92] are examples of parallel DBMSs that actually were implemented. Each of these systems exploits...