Μελέτη Της Επίδοσης Των Ad-hoc Decision-support Ερωτημάτων

Abstract

Decision-support queries are submitted by higher management executives who need to retrieve data from a database server in order to draw a pattern of the company financial results and facilitate their decision making process. These queries are highly complex, handle large amounts of data and are usually submitted unpredictably, or else ad-hoc. Therefore, it is inevitable that they take a long time to execute and it becomes crucial to find ways to optimize their performance. The first step would be to choose the most efficient database management system; then, tune it appropriately; finally, identify any ways to facilitate the task of query optimization, as optimization becomes extremely important at such levels of query complexity. A useful tool that could be used to experiment before making all the above choices is the TPC-H benchmark for ad-hoc decision-support queries.This diploma thesis constitutes an attempt to examine the TPC-H database benchmark in detail and use it to compare the Microsoft SQL Server 2008 and MySQL 5.1 database systems. Furthermore, we will try to explain the performance differences by observing the systems behaviour under different configurations, as well as by examining the logic behind query optimization decisions in each system.