On Trust Optimisation for Decentralised M2M Services

Abstract

The service provision in traditional Machine-to-Machine Communication (M2M) ecosystems relies on commercial service providers. Their services are implemented using centralised infrastructures and focus on specific business processes. Recently, the role of end-users and their personal environments became interesting in terms of service provision. Intelligent devices located in the end-user's environment can be used to create M2M services. These services enhance the competition in the M2M marketplace and provide more service variety. The end-user integration creates a decentralised network of service providers, coming with benefits such as energy efficiency or platform independence. However, inexperienced end-users and missing authorities monitoring the service creation workflow can lead to unreliable M2M services. Decentralised environments also increase the risk of uncontrollable behaviour and untrustworthy activities. In such scenarios, trust relationships are essential to protect these networks and render them a perspective. Trust positively affects the overall security of decentralised systems and should be maintained using suitable trust management systems. The aim of this thesis is to optimise the overall trustworthiness in decentralised M2M communities. This research initially identifies limitations related to three topics: decentralised M2M service provision, service testing, and trust approaches. Decentralised M2M service platforms suffer from a lack of trust. Existing M2M service testing concepts are not end-user-friendly, use centralised architectures, and do not follow end-user-based service provision aspects. Current trust management systems are using incomplete trust models not covering several attributes of an entity. They are also not considering the trust scores of new M2M services or service providers. This aspect allows malicious nodes to join the community and harm other members. Another drawback is that existing trust approaches do not provide a secure storage system for all the computed trust data. Overall, the review highlights the need for a new framework addressing these issues. This thesis proposes a novel framework for trust evaluation and corresponding functional testing of decentralised M2M services. It presents a comprehensive trust approach that considers several aspects such as the trust status of new M2M entities, secure trust data storage, and reliable decision-making among the nodes. In order to overcome the lack of trust information about new M2M entities, service testing is incorporated within the trust evaluation process. Additionally, this research utilises blockchain technology to increase data integrity and optimise different parts of the framework. In terms of blockchain, this thesis also identifies drawbacks to existing consensus protocols and proposes a novel trust-based one. Furthermore, all M2M community members maintain the various trust activities using a community-based approach that relies on blockchain principles. Finally, this thesis also presents a prototypical proof of concept evaluation of the proposed framework components. The evaluation results confirm the benefits of this approach and its resilience against attacks in comparison to others. Overall, the combination of blockchain, service testing, and trust optimises the security in decentralised M2M communities.