Log data is an essential tool to identify the footprint of unauthorized activities executed in a network system. Hence, a compact storage mechanism is required for the massive volume of log data to protect them from malicious tampering attacks. In this regard, Blockchain (BC) has been used to design tamper-proof storage of log records. However, the existing BC-based solutions cannot efficiently handle continuously growing massive log data, creating tremendous storage overhead on the participating BC nodes. Although some works address the storage scalability issue through separate off-chain storage, these works cannot support log data confidentiality and essential query mechanisms to manage log data are missing. Moreover, due to inadequate analysis of the real-time implementation, the performance gain obtained by these schemes is not clearly understood. To handle these deficiencies, we propose a BC-based network log data storage and management scheme that uses an InterPlanetary File System (IPFS) to outsource most of the log data to external off-chain storage. In addition, the proposed scheme performs query and audit operations to manage plaintext and encrypted log records efficiently. Besides, we present a theoretical analysis to show our scheme’s scalability in storage gain. Extensive experiments on the prototype implementation of the proposed system show that storage gain increases exponentially with increasing log records per transaction. Moreover, our scheme attains nearly 93% storage reduction in supporting per day storage demand of log records. The experimental results also demonstrate that the proposed system can be realized with a low computational overhead.
Read full abstract