Business Process Decomposition & Distribution for Adaptive Internet of Things

Andres Kiik
The Internet of Things (IoT) offers a great potential in many different application areas such as health care and home automation. However, in order to realize this potential, significant hurdles still have to be overcome. One of the hurdles is how to integrate IoT systems into business processes in the context of using standards such as BPMN2.0. In this thesis I present the results of a state of the art synthesis of Business Process Driven Internet of Things approaches. Based on the results, a business process modelling standard was chosen and used to overcome challenges which are related with IoT and home automation. The first challenge is the adaptiveness of the current IoT systems. For example, while new and more capable sensors enable complex applications such as smart door locks with face recognition that require significant computing resources, home controllers are often resource-constrained devices and therefore considered as a bottleneck for these complex systems. The second challenge is the interoperability of the current home automation systems because systems provided by different vendors cannot be controlled by the same application. To address these issues, this thesis proposes a software framework that enhances the adaptiveness and performance of IoT solutions. This is achieved through a novel approach where the process model is decomposed and the decomposed subparts are offloaded to an external entity. A prototype of this framework has been implemented using Camunda workflow engine. The framework supports IoT by integrating with the OpenHAB smart home system. The performance and scalability of the system is evaluated through a series of experimental case studies. Results showed that offloading can help in case of compute intensive tasks, a face recognition task performed three times faster for example.
Graduation Thesis language
Graduation Thesis type
Master - Software Engineering
Jakob Mass
Defence year