ADAPTIVE SELECTION OF A DISTRIBUTION METHOD DURING EXECUTION OF PARALLEL JOIN OPERATIONS

摘要

A method, system, and computer program product for high-performance database systems. The method commences by estimating, during a compile phase, a statically-calculated system cost of performing parallel join operations over two relations, then selecting, during the compile phase, a first distribution method to perform the parallel join operations. Instructions (e.g., table partitioning) are communicated to two or more execution units for execution of the parallel join operations, and during execution, various measurements are taken pertaining to the performance of a first portion of the parallel join operations using the selected first distribution method. Then, responsive to the measurements a second distribution method is selected and corresponding instructions are communicated to the two or more execution units. The parallel join operations continue, now using the second distribution method.

主权项

1. A computer implemented method for adaptive selection of distribution methods responsive to measurements taken during execution of parallel join operations over two relations, the method comprising:
estimating, during a compile phase, a system cost of performing the parallel join operations over the two relations; selecting, during the compile phase, a first distribution method to perform the parallel join operations; communicating first distribution method instructions to two or more execution units, the first distribution method instructions according to the selected first distribution method; collecting, from at least one of the two or more execution units, at least some measurements pertaining to performing a first portion of the parallel join operations using the selected first distribution method; selecting, based at least in part on the measurements, a second distribution method; and communicating second distribution method instructions to the two or more execution units.