Extension of the Business Process Execution Language (BPEL) with probabilistic time management of choreographies and time constraints in workflow systems

Abstract

WS-BPEL (=Web Service Business Process Execution Language) provides a language for describing the behavior of a business process based on interactions between the process and its partners. The interactions of partners occurs through web service interfaces which ensure interoperability between applications by using web standards.<br />Time management and temporal conformance is an important criteria for business processes. It ensures that activities are performed in a timely manner and that the right information is delivered to the right activities at the right time, such that overall temporal restrictions are satisfied and no deadline is violated. Deadline violations increase the execution time and cost of business processes because they require some type of exception handling. The contribution of this diploma thesis will be the implementation of a tool that enables users to augment WS-BPEL-processes with temporal information in design time and a run time component for checking the temporal behavior of the processes at run time. In design time a valid temporal interval for (activities of) each process is calculated with consideration for the structure of each process and the interactions between different processes and then checked if the model is temporally feasible, i.e. there is a solution satisfying all temporal constraints. Calculating the temporal intervals and checking the temporal conformance will be performed for two scenarios. The interval-based scenario, which allows a variable duration of activities within an interval. The stochastic scenario, that allows the definition of different activity durations with different probabilities and also takes the conditional branching behavior into account. In run time, it is checked if the activities are executed within the valid calculated intervals at design time. Therefore the time points of all processes are mapped to a calendar with current dates and times. Then, the temporal status of each process instance can be monitored.<br />The prototype of this diploma thesis can be downloaded at http://www.ifs.tuwien.ac.at/oesterle.<br />