TOMSPIN-a Tool for Modelling with Stochastic Petri Nets

The basic objective of TOMSPIN (Tool for Modelling with Stochastic Petri Nets) is to provide industrial users with a tool that can analyse fairly big systems using GSPN (Generalized Stochastic Petri Nets) models with some slight extensions. The front-end of TOMSPIN consists of a hierarchically structured net description language, with an optional graphical editor. After transforming the net into a Markov chain, it can be evaluated by steady-state or transient analysis. For the steady-state analysis several algorithms can be used: Besides the well-known Gau/spl beta/-Seidel etc. we have developed a very efficient algorithm called Method 100. For the systematic variation of architectural parameters, a series of experiments can be defined, which vary all relevant net parameters. TOMSPIN has been used to evaluate a broad range of complex commercial computing system, particularly considering performance and reliability aspects.