A reactive system for specifying and running flexible cloud service business processes based on machine learning

Abstract

In this paper, we present a reactive system for running flexible business process. We define two flexibility patterns based on BPMN (Business Process Model Notation) that deals with changes of resource requirements for business process. The business processes are built on an abstract level, using a BPMN model for the specification of the cloud service business process structure based on flexibility patterns, and the state-chart diagram for the specification of the cloud service business process behaviour. In fact, our approach is based on model driven engineering to facilitate the business process design for developers and free them from the lower cloud details during the running time of such applications. Moreover, the use of flexibility actions ensures the capacity of making a compromise between adapting rapidly and easily business process when running errors occur and keeping the effectiveness of these updated application models. This update is controlled by the real time system which is based on a machine learning algorithm to depict the appropriate cloud service to involve in the business process model using the right flexibility action. Finally, we present some results of our system.