City Research Online

Transaction behaviour in large database environments: A methodological approach

Youssef, M. W. A. F. (1993). Transaction behaviour in large database environments: A methodological approach. (Unpublished Doctoral thesis, City, University of London)

Abstract

This thesis presents the CITY benchmark, a database benchmark that fairly represents On-Line Transaction Processing (OLTP) environments. It analyses the most widely used benchmarks in general putting more emphasis on the Wisconsin benchmark and the Transaction Processing Council (TPC) benchmarks (TPC-A, TPC- B and TPC-C) in particular. It also presents an empirical approach to examine the workload of those benchmarks and discovered several technical limitations in their scripts. The thesis also presents an investigation of on-line transactions in large database environments. The tested environments were three of the largest organisations in the UK, those organisations were different in objectives and activities. The investigation identified on-line transaction behaviour and defined the salient characteristics of databases in high-volume transaction environments. The findings from those studies established the basis of a transaction and set of tables that are representative of them. The CITY benchmark design is directly driven from the findings from the empirical studies. The benchmark design took into consideration all the critiques directed towards the TPC benchmarks A, B and C. It is the first benchmark that is designed as a result of studying the behaviour of on-line transactions and databases in large database environments. The CITY benchmark is mainly designed to test and compare database systems performance in high-volume transaction environments (OLTP).

The work revealed the salient characteristics of large database environments and identified a typical behaviour of on-line transaction in OLTP environments. This research has clearly shown that the TPC benchmarks are not representative to the domain of high-volume transactions environments (OLTP) and it explained why they could be misleading if used to test database management systems in this domain. Additionally, this thesis presents a database performance evaluation methodology that is based on in-depth studies in large database environments.

Publication Type: Thesis (Doctoral)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments: School of Science & Technology > Computer Science
School of Science & Technology > School of Science & Technology Doctoral Theses
Doctoral Theses
[img]
Preview
Text - Accepted Version
Download (21MB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login