On the application of markovian object nets to integrated qualitative and quantitative software analysis

In this paper a Petri net based methodology is outlined for an integrated qualitative and quantitative analysis of parallel software systems which is based on different (Petri) net representations of die software system under consideration. The software validation methodology starts from the source text of a set of communicating processes which specify the system under development. From this source text skeleton, a Petri net representation of the general control structure of the system is generated. On this basis a set of reduction steps is defined for transforming the net into different intermediate representations all owing the validation of both qualitative and qualitative properties. Particular attention is paid to the validation of quantitative properties which is performed using a transformation into locally Markovian Object Nets (MONs). Models of this new class of modular Petri nets are obtained by property-preserving structural compression and by enhancement with quantitative i nformation This information is obtained in the form of frequency and delay parameters by monitoring testing the software on the basis of prototypes. Compared to a previous paper on the same topic /Heiner94/ the methodology as well presentation have been improved. In particular, a new, more comprehensive running example been selected and the object net approach is explained in much more detail.