A compute-Efficient Bitmap Compression Index for Database Applications

Software · Sun Jan 01 00:00:00 EST 2006

OSTI ID:1230829

Wu, Kesheng; Shoshani, Arie

FastBit: A Compute-Efficient Bitmap Compression Index for Database Applications The Word-Aligned Hybrid (WAH) bitmap compression method and data structure is highly efficient for performing search and retrieval operations on large datasets. The WAH technique is optimized for computational efficiency. The WAH-based bitmap indexing software, called FastBit, is particularly appropriate to infrequently varying databases, including those found in the on-line analytical processing (OLAP) industry. Some commercial database products already include some Version of a bitmap index, which could possibly be replaced by the WAR bitmap compression techniques for potentially large operational speedup. Experimental results show performance improvements by an average factor of 10 over bitmap technology used by industry, as well as increased efficiencies in constructing compressed bitmaps. FastBit can be use as a stand-alone index, or integrated into a database system. ien integrated into a database system, this technique may be particularly useful for real-time business analysis applications. Additional FastRit applications may include efficient real-time exploration of scientific models, such as climate and combustion simulations, to minimize search time for analysis and subsequent data visualization. FastBit was proven theoretically to be time-optimal because it provides a search time proportional to the number of elements selected by the index.

• Short Name / Acronym: FASTBIT; 001887MLTPL00
• Version: 00
• Programming Language(s): Medium: X; OS: GNU/Liux, MS-windows, Unix, Solaris; Compatibility: Multiplatform
• Research Organization: Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC03-76SF00098
• OSTI ID: 1230829
• Country of Origin: United States