| Kurzfassung
|
Today it is hardly possible to find a business process (BP) that does not involve working with an information system (IS). In order to better plan and improve such BPs a lot of research has been done on modeling and analysis of BPs. Given the dependency between BPs and IS such assessment of BPs should take the IS into account. Furthermore, in most assessment of BPs only the functionality, but not the so called non-functional requirements (NFR) are taken into account. This is not adequate, since NFRs influence BPs just as they influence IS. In particular the NFR reliability is interesting for planning of BPs in business environments. Therefore, the presented approach provides an integrated reliability analysis of BPs and IS. The proposed analysis takes humans, device resources and the impact from the IS into account. In order to model reliability information it has to be determined, which metrics will be used for each BP element. Thus a structured literature search on reliability modeling and analysis is conducted in seven resources. Through the structured search 40 papers on modeling and analysis of BP reliability were found. Ten of them were classified as relevant for the topic. The structured search revealed that no approach allows for modeling reliability of activities and resources separate from each other. Moreover, there is no common answer on how to model human resources in BPs. In order to enable such an integrated approach the reliability information of BPs is modeled as an extension of the IntBIIS approach. BP actions get a failure probability and the resources are extended with two reliability related attributes. For device resources the commonly used MTTF and MTTR are added in order to provide reliability information. Roles, that are associated with actor resources, are annotated with MTTF and a newly developed MTTRepl. The next step is a reliability analysis of an BP including the IS. Markov chains and reduction rules are used to analyze the BP reliability. This approach is exemplary implemented with Java in the context of PCM, that already provides analysis for IS. The result of the analysis is the probability of successful execution of the BP including the IS. An evaluation of the implemented analysis presents that it is possible to analyze the reliability of a BP including all resources and the involved IS. The results show that the reliability prediction is more accurate, when BP and IS are assessed through a combined analysis.
|
