USING SINGULAR VALUE DECOMPOSITION TO IMPROVE THE SECURITY OF DISTRIBUTED BLOCKCHAIN ORACLES

Kornienko Anatoly Adamovich, Beksaev Nikolai Sergeevich

This paper proposes and experimentally validates a method for improving the reliability of distributed blockchain oracles using singular value decomposition (SVD). This method enables efficient detection of anomalies in data received from multiple oracles, which is particularly important for protecting against coordinated attacks (so-called "collusion attacks"). The paper analyzes the architecture and classification of oracles, examines existing verification mechanisms, and describes in detail the application of SVD analysis. Experimental results on simulated data from a decentralized oracle operating on a parametric insurance blockchain confirm that the proposed approach successfully identifies compromised sources, increasing trust in blockchain applications in finance and insurance.

Keywords:Blockchain, smart contract, blockchain oracles, decentralized blockchain oracles, coordinated attack, singular value decomposition, data security, anomaly detection, consensus, decentralized applications (dApps).