Scalable parallel query server for decision support applications

Abstract

Decision-support applications require the ability to query against large amounts of detailed historical data. We are exploiting parallel technology to improve query response time through query decomposition, CPU and I/O parallelism, and client/server approach. IBM System/390 Parallel Query Server is built on advanced and low-cost CMOS microprocessors for decision-support applications. In this paper, we will discuss our design, implementation and performance of a scalable parallel query server.