Blockchain based Secure Log Storage

With the exponential growth of log data in the digital age, the need for effective and secure log management has become essential for organizations. Conventional log storage systems are known to struggle with challenges related to data security and privacy. As log files often contain sensitive data, any leakage may result in security violations and compliance issues. In this article, a Blockchain-based system is proposed, integrating a RoBERTa- based sensitivity classifier with a layered circled blockchain mechanism to ensure secure and efficient log storage. The system begins with the aggregation of logs from local systems, which are then processed using a pre-trained RoBERTa model to determine their sensitivity. Non-sensitive logs are stored in the Ethereum public blockchain to maintain transparency and permanence, while sensitive logs are stored in a Hyperledger private network to ensure confidentiality. A layered circled blockchain mechanism is proposed to dynamically compute the optimal chain length for log storage, facilitating the creation of terminal and progressive blocks. The system is evaluated based on key performance indicators, including chain length, transaction throughput (TPS), and log sensitivity classification accuracy. Experimental results demonstrate that log security is enhanced and privacy-preserving storage is ensured, and overall performance is improved.