tMPT: Reconfiguration across Blockchain Shards via Trimmed Merkle Patricia Trie

Sharding is one of the most promising techniques that can improve the scalability and storage issues of blockchain systems. In the past few years, many sharding protocols have been proposed to contribute to the technique matrix of blockchain sharding such as the coordination mechanisms of blockchain shards, the handling of cross-shard transactions, the security guarantee of blockchain shards, etc. Although those previous solutions are crucial for sharded blockchains, we still have not found any systematic implementation of shard reconfiguration, which determines the security of a sharded blockchain because the shuffling of blockchain nodes could prevent malicious nodes from corrupting a shard. However, the implementation of shard reconfiguration is not easy. When reallocating blockchain nodes to designated shards, typical challenges include the following: i) how to synchronize a large size of state data for a newly arrived node, and ii) how to mitigate the large reconfiguration latency of blockchain shards while keeping the liveness and consistency properties of a blockchain system. To overcome those challenges, we propose a dedicated protocol for shard reconfiguration in blockchain sharding using trimmed Merkle Patricia Trie (tMPT). The proposed tMPT-based protocol is designed to guarantee the high efficiency of the reconfiguration of blockchain shards while ensuring the uninterrupted services of the sharded blockchain. We implement the proposed tMPT-based reconfiguration protocol in a prototype, which enables the functionality of blockchain sharding. We then deploy our prototype in Alibaba Cloud. The experimental results show that the proposed tMPT-based protocol outperforms the existing methods in terms of reconfiguration efficiency. For example, the throughput of the proposed protocol shows 198% higher than Ethereum's full sync method.