Towards Service Co-evolution in the Internet of Things

Nowadays, in the world of technology, where the new technologies are devolved and replace the old ones every single day, any software product in order to maintain its competitiveness, must be updated continually. Updating a software product results from an evolution process, including adding new features, replacing outdated features, repairing bugs, closing security gaps, and improving performance, which requires a lot of effort and knowledge. Therefore, the Internet of Things (IoT) services are no exception.

An interesting research question is how to handle service changes for service consumers and how to enable and facilitate end-user application updates in case the dependent clients provide services to other clients, especially in IoT environments. These manifold interdependencies make on-the-fly service evolution a particularly difficult and challenging problem because the evolution of one service may incur changes in other dependent services and clients. In analogy to biology, we call this service coevolution. Thus, the thesis aims to develop a comprehensive solution for the coordinated evolution of heterogeneous services in IoT. The main contribution of this dissertation is a set of theoretical models and approaches that facilitate service coevolution. In particular, we developed a solution for coordinated service coevolution through a design technique that equips every service with an intelligent agent, called EVA (Evolution Agent), that performs the service evolution in collaboration with other EVAs. The EVA can control service versions, update local service instances. Furthermore, we proposed a notification management architecture for IoT services. Additionally, an approach to describe and detect changes in IoT services with support for a shared knowledge base is introduced. Last but not least, a method to find out changes in service behaviour is presented by analyzing the data stream between the service client and service provider.