Detection and Mitigation of Insider Attacks in Financial Systems

In the realm of financial applications relying on relational databases, the susceptibility to insider attacks poses a significant threat, with users holding administrative privileges capable of manipulating crucial entries. In response to this challenge, this project proposes a smart-contractless blockchain approach to detect insider attacks by harnessing the tamper- resistant attributes of blockchain. This approach functions as a data-agnostic framework, enabling the utilization of blockchain network for storing fixed-length fingerprints of tuples extracted from an SQL database, eliminating the necessity for a full-scale data migration and thereby enhancing the security. The project employs a formalism to intercept SQL queries and their results to ensure the integrity of respective tuples by cross-referencing with the fingerprints stored on the blockchain, which serves to identify and detect insider attacks. The enhanced hashing algorithm deployed in the technique of fingerprinting makes uses of parallel processing approach which creates an overlay- like structure inside the blockchain, which helps in reducing the computation overhead thereby detecting the attacks faster. The system automatically directs towards mitigation of the attack when an attack is detected and the tampered data gets restored to its previous consistent state upon mitigation.