Decentralized Reputation Model and Trust Framework Blockchain and Smart contracts

Sammanfattning: Blockchain technology is being researched in diverse domains for its ability to provide distributed, decentralized and time-stamped transactions. It is attributed to by its fault-tolerant and zero- downtime characteristics with methods to ensure records of immutable data such that its modification is computationally infeasible. Trust frameworks and reputation models of an online interaction system are responsible for providing enough information (e.g., in the form of trust score) to infer the trustworthiness of interacting entities. The risk of failure or probability of success when interacting with an entity relies on the information provided by the reputation system. Thus, it is crucial to have an accurate, reliable and immutable trust score assigned by the reputation system. The centralized nature of current trust systems, however, leaves the valuable information as such prone to both external and internal attacks. This master's thesis project, therefore, studies the use of blockchain technology as an infrastructure for an online interaction system that can guarantee a reliable and immutable trust score. It proposes a system of smart contracts that specify the logic for interactions and models trust among pseudonymous identities of the system. The contract is deployed on a blockchain network where the trust score of entities are computed, stored and updated. The proposed method and the trust metrics used is evaluated by simulating an interaction graph using an existing dataset. The obtained results then illustrate that the proposed method is resilient against several threat models and distinguishes between honest or malicious participation of entities on the network.