Improving I/O Bandwidth for Data-Intensive Applications
Marcin Żukowski · Centrum Wiskunde & Informatica (CWI), the national research institute for mathematics and computer science in the Netherlands · 2005
Abstract. High disk bandwidth in data-intensive applications is usually achieved with expensive hardware solutions consisting of a large number of disks. In this article we present our current work on software methods for improving disk bandwidth in ColumnBM, a new storage system for MonetDB/X100 query execution engine. Two novel techniques are dis-cussed: superscalar compression for standalone queries and cooperative scans for multi-query optimization. 1