Fault Tolerance Framework using Model-Based Diagnosis: Towards Dependable Business Processes

Abstract

Several reports indicate that one of the most important business priorities is the improvement of business and IT management. Management and automation of business processes have become essential tasks within IT organizations. Nowadays, business processes of a organization use external services which are not under our its jurisdiction, and any fault within these processes remain uncontrolled, thereby introducing unexpected faults in execution. Organizations must ensure that their business processes are as dependable as possible before they are automated. Fault tolerance techniques provide certain mechanisms to decrease the risk of possible faults in systems. In this paper, a framework for developing business processes with fault tolerance capabilities is provided. Our framework presents various solutions within the scope of fault tolerance, whereby a practical example has been developed and the results obtained have been compared and discussed. The implemented framework presents innovative mechanisms, based on model-based diagnosis and constraint programming which automate the isolation and identification of faulty components, but it also includes business rules to check the correctness of various parameters obtained in the business process.