Decentralized service discovery and selection in Internet of Things applications based on artificial potential fields

The Internet of Things (IoT) vision involves a future Internet integrated with real-world objects that can commonly offer their functionality trough services. In such pervasive environments of IoT networks, locating and invoking suitable services is quite challenging and traditional service discovery and selection approaches have been proven inadequate. In this paper, taking inspiration from natural metaphors, a decentralized service discovery and selection model is proposed. The model is based on artificial potential fields (APFs) which are formed upon each user service request and become active at points where services can be provided. Such points are termed as service provision nodes (SPNs). The strength of each APF depends on the percentage of requested services that can be provided by the respective SPN, as well as on SPN service load and availability with the aim to balance service load among SPNs. Service discovery and selection is then driven by artificial forces applied among user service requests and SPNs. Simulation results indicate that the proposed approach maintains satisfactory performance and scalability as the number of SPNs in an IoT network increase and efficient load balancing of the requested services among the SPNs in comparison with other approaches.