2011 Sixth International Symposium on Parallel Computing in Electrical Engineering (2011)
Luton, United Kingdom
Apr. 3, 2011 to Apr. 7, 2011
The Info bright database engine stores data in column wise chunks in a highly compressed form. Access to data is slowed down by the necessity of fetching required chunks from disk and decompressing them. A pref etching mechanism is meant to perform these tasks shortly before a running query needs the data for processing. The pref etching implementation in Info bright is described and evaluated. An average speedup of 1.7 is reported. The speedup reaches 2.4 for some queries and generally is higher for long lasting queries for which the data fetching and decompression cost dominated the execution time. The limitations of the pref etching are presented on both system level and SQL query level.
prefetching, compression, database engine, columnar store
J. Borkowski, "Increasing Query Speed with Multithreaded Data Prefetching," 2011 Sixth International Symposium on Parallel Computing in Electrical Engineering(PARELEC), Luton, United Kingdom, 2011, pp. 117-122.