Client/Server Performance Testing
First published 06/11/2009
This paper, written by Paul Gerrard was presented to the EuroSTAR Conference in London, November 1995 and won the prize for 'Best Presentation'.
Client/Server (C/S) technology is being taken up at an incredible rate. Almost every development organisation has incorporated C/S as part of their IT strategy. It appears that C/S will be the dominant architecture taking IT into the next millennium. Although C/S technology is gaining acceptance rapidly and development organisations get better at building such systems, performance issues remain as an outstanding risk when a system meets its functional requirements.
This paper sets out the reasons why system performance is a risk to the success of C/S projects. A process has been outlined which the authors have used to plan, prepare and execute automated performance tests. The principles involved in organising a performance test have been set out and an overview of the tools and techniques that can be used for testing two and three-tier C/S systems presented.
In planning, preparing and executing performance tests, there are several aspects of the task which can cause difficulties. The problems that are encountered most often relate to the stability of the software and the test environment. Unfortunately, testers are often required to work with a shared environment with software that is imperfect or unfinished. These issues are discussed and some practical guidelines are proposed.
Tags: #performancetesting #client/server
Paul Gerrard My linkedin profile is here My Mastodon Account