Goal-Based Workflow Adaptation for Role-Based Resources in the IoT

Ubiquitous Systems (UbiSys) are highly runtime dynamic and context-sen-sitive systems, integrating a large number of heterogeneous components as well as users. Sensors, actuators, WebServices and software components create a closed loop between sensing the physical world, reasoning in the cyber world and subsequently acting on the physical world. Common application scenarios for UbiSys are smart environments like smart homes and the factory of the future, where ubiquitous system components collaborate in order to reach user-specified goals. Processes and workflows represent the most prominent means for formalising such goals and repeating tasks. However, the dynamic structure and context-sensitivity of ubiquitous systems represent major challenges for process modelling and execution, as a large variety of heterogeneous resources and subsystems as well as sources of information and context factors have to be integrated. In addition, the dynamic removal of process resources has to be supported at runtime. These factors represent serious obstacles for the successful employment of ubiquitous systems for process execution nowadays and in future.

Existing approaches apply property-based or goal-oriented under-specification of dynamic processes not considering context-sensitive behaviour. Our research aims at a goal-oriented process execution approach for ubiquitous systems that combines goal-based process model specification at design time with role-based resource allocation at runtime. This approach increases flexibility and runtime-adaptivity of processes as required by ubiquitous systems. It allows for per activity process model adaptation according to available resource roles. The application of roles for resource abstraction and allocation enables a context-sensitive selection of component compositions. Goals simplify the process specification and yield more flexible activities and processes.