Research On PBFT-Based Consensus In Blockchains

Blockchain technology has been increasingly applied in various fields,such as the Internet of Things and supply chain management,due to its key features of decentralization,traceability,and reliability.However,in large-scale public blockchain systems,throughput has become the biggest bottleneck.To improve the throughput of blockchain networks,many studies have focused on applying voting-based consensus algorithms in public blockchains.Nevertheless,there are still some key issues that need to be addressed.One of the challenges in the technical solution of applying voting-based consensus algorithms in public chains is that block rewards are distributed equally among all nodes.As a result,nodes can obtain block rewards without participating in the consensus process,leading to a reduction in the availability of the blockchain network.Another challenge is the need to introduce a reputation system to reward and punish node behavior.The adjustment results of the nodes’ reputation need to be recorded in the global state,which leads to significant storage overhead.Additionally,communication outside of the consensus group also results in significant communication overhead.To address these challenges,this research analyzed the characteristics of existing sharding solutions and the technical features of current consensus algorithms and incentive mechanisms.This led to the design of two technical solutions based on a double-chain architecture for detecting and punishing mute nodes and consensus on observation results based on threshold signatures.These two solutions ensure accurate detection of selfish nodes in the blockchain network,with minimal impact on network performance.The main contributions of this research are twofold.First,to address the issue of network availability caused by mute nodes,a double-chain architecture based mute node detection and punishment technique was proposed.The technique uses a double-chain architecture to handle blockchain transaction consensus and observation result consensus,and an observation result exchange protocol was designed to complete the detection of mute nodes.A reputation system was introduced based on observation result consensus,and an incentive mechanisms was designed to complete the punishment of mute nodes.An analysis of possible attack behaviors showed that if nodes adopt partial or complete mute strategies,or participate in node state consensus dishonestly,their reward will be lower than other nodes.The experimental environment was constructed,and the protocol implementation and sharding simulation were combined to demonstrate that this technique has minimal impact on consensus group performance and has more accurate detection and punishment of mute nodes than existing techniques.Second,to address the issue of storage and communication overhead caused by reputation value recording,a consensus on observation results based on threshold signatures technique was proposed.In this technique,a communication protocol for consensus group nodes to propagate observation result matrices was designed based on the observation result exchange protocol to ensure that the global reputation state can reach consensus.The TS-HBBFT consensus algorithm was designed based on a leaderless consensus scheme and threshold signature technology to reduce the storage and communication overhead caused by high-complexity signatures.Security and complexity analysis showed that this technique can ensure consistency and availability while reducing the protocol message complexity within and outside the consensus group.Experimental analysis demonstrated that this technique reduces redundancy caused by message signatures in consensus and improves the overall performance of the blockchain network in sharding scenarios.In summary,through the key research mentioned above,the challenges of applying voting-based consensus algorithms in public blockchains have been addressed.The proposed techniques have significantly improved the detection and punishment of mute nodes while reducing the storage and communication overhead caused by reputation value recording.These solutions demonstrate the potential for further optimization of the blockchain network’s performance and the importance of addressing the issues mentioned above to promote the wider adoption of blockchain technology in various fields.