Blockchain-driven trust evaluation and secure key agreement protocol for rotating savings and credit association

Publications

Blockchain-driven trust evaluation and secure key agreement protocol for rotating savings and credit association

Blockchain-driven trust evaluation and secure key agreement protocol for rotating savings and credit association

Year : 2025

Publisher : Springer Science and Business Media B.V.

Source Title : International Journal of Information Technology (Singapore)

Document Type :

Abstract

Rotating Savings and Credit Associations (ROSCA) is extensively utilized as a vital financial mechanism in both urban and rural communities around the world. This enables individuals to save and borrow money collectively. However, relying on mutual trust, non-transparency, and insecure exchange of sensitive information leads to fraud, data breaches and single point of failure. Therefore, to address these issues, we propose a decentralized trust evaluation mechanism and secure key agreement protocol for ROSCA integrated with blockchain. This framework presents a four-layer, trust-building architecture that evaluates the trustworthiness of ROSCA operators and stores the time-stamped trust index in the blockchain. It leverages Mamdani Fuzzy Inference System(MFIS) to determine the degree of trustworthiness. Consequently, ROSCAs with better trust indexes are evaluated, validated, and stored in blockchain via chaincode. In addition, we propose a secure key agreement protocol for ROSCA-to-customer, thereby enabling a secure communication channel for exchanging sensitive information. Further, to achieve transparency, fairness and to eliminate centralized control, the complete framework is implemented over permissioned blockchain. The formal and informal security verification reveals that proposed system is safe and secure against potential threats. Moreover, performance analysis and comparison with current state-of-the-art reveal that the suggested framework outperforms and potentially reduces chaincode computation time for creating and querying records by 78.32% and 85.42%, respectively.