Optimization Scheme For Hyperledger Fabric Consensus Mechanism Based On Verifiable Random Function

The blockchain technology originated in Bitcoin has attracted extensive attention in the fields of digital finance,data traceability,and property rights protection because of its decentralization,data tamper-proof,and traceability.Blockchain is divided into public blockchain,consortium blockchain and private blockchain according to the degree of centralization.With the exploration of blockchain technology,the consortium blockchain with privacy protection and identity management functions has become the mainstream choice for enterprise-level blockchain applications.At present,most of the consortium blockchain building blockchain use the open source blockchain platform Hyperledger Fabric.Hyperledger Fabric has the advantages of modularity,high security,and fast transaction speed.However,the endorsing peers in Hyperledger Fabric are highly centralized and vulnerable to attacks,which can lead to security risks.In addition,a single endorsing peer processes all transactions across the entire network,causing performance bottlenecks.In view of the problems in the Hyperledger Fabric consensus mechanism,this paper proposes an optimization solution from the perspective of security and transaction performance,which solves the pain points of Hyperledger Fabric,and also provides reference for the optimization of other consortium blockchain consensus mechanisms.The main work of this paper is as follows:First,this paper designed a verifiable random function algorithm based on elliptic curves.The algorithm satisfies verifiability,uniqueness,pseudorandomness and collision resistance.This algorithm is implemented with Go language.The results show that the algorithm's key generation time,random number generation time,and random number verification time satisfy the transaction requirements in Hyperledger Fabric.Second,based on the original "endorsement-sorting-verification" consensus model,this paper designed a candidate endorsing peers set mechanism,and the original fixed endorsing peer was changed to a set of candidate endorsing peers,and then a select algorithm based on verifiable random function was used to determining the endorsing peer and makes it difficult for the adversary to determine the target of the attack,thereby increasing the difficulty of the adversary's attack and improving the security of the consensus mechanism.Third,this paper designed and implemented a multi-group endorsing peer mechanism.A select algorithm is used to transfer each transaction to a different set of endorsing peers for parallel processing.Relying on the pseudorandomness of the select algorithm,all transactions can be distributed to different endorsing peers as a whole,reducing the number of transactions for each endorsing peer,thereby shortening the transaction processing time of the system and increasing the consensus speed.Based on the Go programing language,the consensus process in Hyperledger Fabric is modified to achieve non-interactive random selection of endorsing peers and parallel processing of endorsing processes under verifiable conditions.Analysis and testing show that the optimized consensus mechanism has higher security,lower transaction delay and faster transaction processing speed under the premise that the communication cost is basically unchanged.