Robust, Secure, and Adaptive Trust-Oriented Service Selection in IoT-Based Smart Buildings

Abstract

Internet of Things (IoT) has become an integral part of a smart community that connects computing devices, smart objects, and mechanical and digital machines, having unique identifiers, to communicate with each, without human intervention. This smart connectivity within a building assists the users in real time to provide an enormous range of connected applications and facilitate the optimized use of multiple resources. These smart building applications can make our lives more comfortable and provide a more sustainable, healthy, and safe workplace. A key challenge for IoT toward smart buildings is to ensure that the service has been taken from a trustworthy service provider. It requires a reliable, robust, and adaptive context-oriented trust mechanism that conforms to the specific requirements of the end users. To enhance the security of smart buildings in IoT, this research proposes an adaptive context-based trust evaluation system for smart building (CTES-SB) applications. The trust score for service is calculated based on the client’s previous interaction and recommendation from context-similar clients. Using CTES-SB, the client selects the best service provider based on the previous and current trust scores for the next interaction. The model also helps to filter out malicious nodes through an indirect trust calculation process. This process dynamically assigns weights based on direct interactions and trustworthy recommendations for detecting and avoiding malicious interactions. We have demonstrated the effectiveness of CTES-SB by simulating multiple smart building scenarios under malicious attacks. The proposed architecture of CTES-SB has been experimentally evaluated to benchmark its performance for best service selection and resiliency against malicious nodes. The CTES-SB is proved to be efficient by having a comparison with the state-of-the-art algorithms. The comparison is in terms of filtering the malicious nodes from the network and the result shows that the trust converges quickly toward the ground-truth value.