Adaptive Trust Evaluation under Dynamic Service Conditions Using Distributed Verification

Abstract

The dynamic service conditions, the lack of centralised control in the distributed and federated services, and the growing sophistication of the malicious behaviours are the key challenges to trust management in the distributed and federated services. Standard trust models, based on fixed credentials, central authorities, or aggregation of reputation over the whole world, are no longer suitable to serve high-rate changing contexts in services, and have very high communication and coordination costs. In an effort to curb these issues, this paper puts forward a proposal of adaptive trust evaluation framework that has distributed verification in highly dynamic service-oriented architectures. The suggested model represents trust as a context-based, multi-dimensional digit that conservatively adapts to the context changes in service conduct, workload, and environmental state. To satisfy the decentralized nature of trust updates, a lightweight peer-based verification system is presented and does not need any centralized sources of trust but instead, trust updates are validated through decentralized means without depending upon a full blockchain consensus system. The framework constitutes adaptive weighting of trust, decay of trust and enforcement policies to reliably detect malicious or unreliable services in changing situations. Between the two widely used approaches, the widespread performance analysis of the suggested approach demonstrates that it has a better accuracy in trust, faster in detecting malicious service and with a much lower communication overhead than state of art centralised, reputation-based and ledger-driven approaches to trust. The findings affirm the viability, scalability, as well as viability of the suggested solution to secure trust execution in the next-generation distributed cloud and edge service surroundings.