Service Composition “Dynamic” IoT Networks

IoT (Internet-of-things) technologies have advanced significantly in recent years. However, many existing IoT systems are statically built. In these systems, specific IoT devices as well as control and management software are statically selected and configured at the design time to achieve some predefined tasks and to handle some anticipated events. To make further use of available “things” in a dynamic IoT network toward a smarter and brighter physical world, it is necessary to dynamically compose and discover services in order to handle complex tasks.

We introduced a planning algorithm to efficiently compose distributed IoT services and address dynamically arising tasks. We developed a decentralized algorithm to traverse through a logic-based overlay network of services and derive the desired service composition. We then developed semantic-based routing protocols to do IoT service discovery and find required services composed together as the output workflow in the previous step. We first consider how to specify the capabilities of IoT devices so as to facilitate effective matchmaking. Then, we consider the memory constraints of IoT and edge devices and focus on the design of routing tables to achieve space efficiency while preserving the high discovery success rate. 

We analytically and experimentally evaluated our proposed solutions for IoT service discovery and composition. The results demonstrated that our solution yields better performance compared to existing solutions in any approach of centralized or decentralized service discovery and composition in dynamic IoT networks.