An Effective and Balanced Storage Extension Approach for Sharding Blockchain Systems

Sharding technology has become crucial for enhancing the scalability of blockchains owing to the rapid extension of blockchain data. However, data migration and state reconstruction may cause a high overhead when a new shard is added. Existing solutions have high latency when expanding, and the balance of the state data between shards is poor after extension. To this end, this paper proposes an Effective and Balanced Storage Extension (EBSE) approach for sharding blockchain systems. EBSE can reduce the overhead and latency of the system extension, while ensuring a balance between shards after extension. When implementing the EBSE, we address the following three challenges. 1) To design a data structure that incorporates allocation principles, we designed a Jump Merkle Tree (JMT) based on the Merkle Tree prototype, incorporating node migration and orderliness. 2) To design additional rules that ensure the integrity of the state tree during shard addition, we designed a shard addition protocol to coordinate and standardize the behavior of each shard during the extension process. 3) To ensure the sustainability of the system after extension, we designed state tree addition and cleaning algorithms to remove the invalid information after the system extension. Extensive experiments are conducted to evaluate the performance of the proposed approach. The experimental results show that the EBSE outperforms the existing solutions in terms of balance and latency. Compared with the state-of-the-art sharding storage, EBSE effectively reduces the shard addition latency by 60% and achieves 4× superior shard data balance.