Cross-domain reputation-based trust management for beyond 5G scenarios

Abstract

The rapid expansion of interconnected services and devices has led to a surge in the variety and number of relationships between entities. With the advent of 5G networks, it iss increasingly common for entities from different administrative domains to establish connections. This growth in interconnectivity often places a strain on resources, pushing end-users to seek assistance from network infrastructure or service providers to enhance their capabilities. Moreover, the plethora of options available for deploying services and resources sometimes leads to an inherent trust. Depending on their needs, end-users might opt for solutions like Infrastructure-as-a-Service, which allows them to tailor virtualized computing resources. Alternatively, they might gravitate towards marketplaces, which offer the convenience of using third-party resources without the hassle of setting up and maintaining them. Marketplaces, in particular, are gaining traction due to their adaptability in dynamic environments. However, trust remains a cornerstone in both scenarios. It's essential when deciding on a business partnership or choosing a service. While marketplaces often lack trust-based filtering options, they do offer advanced filters based on performance, hardware, or location. It is crucial to note that traditional trust models might not be suitable for modern 5G networks, given the ever-evolving challenges. Hence, there is a need for trust models that consider the unique attributes and cutting-edge KPIs of these networks. Under the compendium modality, the three chapters composing this PhD dissertation and their principal objectives are outlined as follows. Firstly, a comprehensive literature review on trust and reputation models in 5G settings was conducted; the first chapter reports on the essential properties and features of trust models, eight key enablers along with a review and comparison of the most recent trust-related papers, and trends and challenges that set the direction for this thesis. Secondly, a pre-standardization strategy for reputation-based trust models beyond 5G was developed; the second chapter reports on examining standardization papers, research initiatives, and regulatory bodies, providing a list of requirements and vital KPIs, and offering preliminary recommendations to address the lack of standardized trust models in post-5G networks. Thirdly, an analysis of the reputation-enabled trust framework when suffering different trust-related attack bursts was examined; the third chapter develops the proposed framework, composed of four modules, and provides a dynamic and automated solution, developed under the 5GZORRO European project, for on-demand service and resource provisioning in decentralized marketplaces. In this article, an adapted PeerTrust model was created to compute trust scores based on statistical information inferred from product offers, network providers, and recommenders. Additionally, an SLA-driven reward and punishment mechanism was designed to continuously adapt trust scores of an ongoing trust relationship when SLA violation, breach prediction, or breach detection appeared in real time. Experiments demonstrated that our reputation framework was resilient to bad-mouthing and on-off attacks. In summary, the chapters composing this PhD dissertation promote cohesive research exploring, analysing and, ultimately, addressing reputation-based trust solutions for beyond 5G scenarios. Nevertheless, some questions mooted by this research remain unsolved, so they still require more effort. Prime among them is whether it will be feasible to conduct an official standardization of reputation-based trust models to homogenise solutions and rowing in the same direction