Computing Science. An Empirical Study Comparing the PEPA Eclipse Plug-In and GPA Tools.

摘要

Performance modellers rely on the accuracy and correctness of tools in order to derive results that can be trusted to give useful insights into system behaviour. However, different tools may be developed from different perspectives and apparently similar analysis may give different results in some situations. In this paper we explore the use of two tools developed to support the analysis of the stochastic process algebra, PEPA. The tools, the PEPA Eclipse Plug-in and the Grouped PEPA Analyser are shown to give consistent results in many situations, but also show some inconsistencies which may not be predictable by the modeller. Hence this paper aims to give some insight as to where the two tools differ in the context of two models of workflow based systems.