Achieving High Efficiency and High Throughput in Erasure Code-Based Distributed Storage for Blockchain

Abstract

A blockchain is a decentralized peer-to-peer network in which all nodes store data in copies, ensuring data integrity, as transactions cannot be changed or deleted. This can lead to duplicate data storage, resulting in high storage overhead, especially in storage-constrained environments, such as the Internet of Things (IoT) or sensor systems. Distributed storage techniques utilizing erasure code (EC) have been investigated to address this issue. Although EC-based blockchain storage increases storage efficiency, encoded chunks distributed across multiple nodes must be received to restore and access the original blocks. However, studies on increasing the data transmission efficiency of EC-based blockchain storage are limited. In this study, we propose a data transmission technique called trigger-based automatic repeat request (ARQ), enabling stable data recovery while ensuring low latency and high-throughput performance, even with frequent node failure. This technique increased the throughput efficiency by 8% while maintaining the decentralization of the blockchain. Furthermore, it maximized the storage efficiency of EC-based distributed blockchain storage by >99.8%, while solving the recovery overhead problem due to data transmission. Using the trigger-based ARQ scheme with an EC-based distribution technique, bl