基于密码累加器的无状态区块链性能优化

Performance Optimization of Stateless Blockchain Based on Cryptographic Accumulator

区块链由于其去中心化的特点,被认为是近年来最具颠覆性和革命性的技术创新之一。然而,目前大多数公有链由于数据高度冗余使得区块链系统数据迅速增长,从而造成节点存储资源消耗严重,此外它还存在区块共识成本较高导致系统吞吐量低的问题。针对区块链的可扩展性问题,提出一种适用于智能合约的无状态区块链性能优化方案STiPChain。STiPChain基于密码累加器与可验证计算完成无状态设计,采用RSA累加器生成智能合约的有效性证明实现节点状态数据压缩,同时通过分布式设计将智能合约的有效性证明更新效率优化至常数,极大降低节点对磁盘和内存的需求。在此基础上,STiPChain将区块共识与区块运算解耦,基于可验证计算技术提出新的智能合约执行逻辑和交易验证方案,解决无状态条件下的交易验证与执行问题,有效提高系统吞吐量。实验结果表明,与Ethereum相比,STiPChain无状态区块链性能优化方案将共识节点的存储需求降低了99%,吞吐量提高了1.6~2 500倍。

Blockchains are considered among the most disruptive and revolutionary technological innovations in recent years because of their decentralization characteristics.However,most public chains are faced with the significant consumption of node storage resources owing to the high data redundancy,making the blockchain system data grow rapidly.In addition,they exhibit low system throughput caused by the high cost of block consensus.High storage and low throughput have become urgent problems to be solved in the development of blockchain.A stateless blockchain storage optimization scheme,STiPChain,suitable for the smart contract,is proposed to solve the scalability problem of blockchains.STiPChain completes a stateless design based on the accumulator and verifiable computing.The RSA accumulator is used to generate the validity proof of the smart contract and realize the state compression.It also generates state commitment using a new cryptographic accumulator technology,optimizing the update efficiency of the state accumulator to a constant value and significantly reducing the disk and memory requirements of the node.In addition,this study decouples block consensus and block operation and proposes a new smart contract execution logic and transaction verification scheme,which solves the smart contract execution under stateless conditions and effectively improves the system throughput.The experimental results show that compared with Ethereum,the STiPChain storage expansion scheme reduces the storage requirements of consensus nodes by 99% and increases the throughput by 1.6-2 500 times.