Multi-Context-based Trust Management Framework and Simulator for Social Internet of Things

The rise of the Social Internet of Things (SIoT) brings forth new challenges in trust management within interconnected device networks. This thesis endeavours to address these challenges by developing and validating a novel Multi-Context-based Trust Management (MCTM-SIoT) framework tailored specifically for SIoT environments. The framework aims to enhance trust assessment by incorporating diverse contextual information into the final evaluation of each node within the SIoT network. This approach enables the selection of the most trustworthy Service Provider (SP) even in the absence of prior behavioural history from the node.

The research objectives encompass a thorough investigation of various research areas, including IoT, SIoT, and trust management in SIoT environments. These investigations pave the way for the development of an MCTM-SIoT framework and model specifically tailored for SIoT. Additionally, a novel SIoT simulator is developed to generate diverse SIoT scenarios and produce realistic datasets, facilitating the evaluation of the proposed framework and model's performance and scalability across different scenarios.

The contributions of this research are manifold. Firstly, it introduces a new MCTMSIoT framework that elucidates the relationship between different SIoT components and trust management. Additionally, the framework incorporates multiple contextual information into the final trust score of each node in the SIoT network to facilitate trustworthy inference, thereby enhancing the overall security and reliability of the system by selecting the most reliable Service Provider. Secondly, a scalable MCTMSIoT model is developed, to identify the most trustworthy service provider based on a set of trust contextual metrics, namely user context trust metrics (UCT), device context trust metrics (DCT), environmental context trust metrics (ECT), and task context trust metrics (TCT). Thirdly, a novel simulation tool is designed to simulate diverse SIoT scenarios, generating realistic datasets crucial for testing and evaluating the proposed framework and model. Finally, a proof of concept is developed to demonstrate the efficacy of the MCTM-SIoT framework and model in SIoT environments using the generated datasets and employing machine learning techniques. Testing of the framework demonstrates that the impact of context on SIoT trustworthiness grows with the level of dynamism and complexity of the SIoT environment, highlighting the importance of considering contextual factors in trust management strategies for SIoT.

In summary, this research contributes significantly to the field of SIoT by providing a comprehensive framework for trust management that addresses the dynamic nature of contextual information. The developed framework and model offer promising solutions to the challenges posed by trust assessment in SIoT environments, paving the way for enhanced security and reliability within interconnected device networks.