Dynamic consensus algorithm and performance enhancement of decentralized data asset registration system

Aiming at the problem of insufficient performance and dynamic adaptability of existing consensus algorithms in decentralized data asset registration system, this paper proposes a consensus algorithm based on dynamic reputation weight and its performance enhancement scheme. By constructing a dynamic weight model that integrates behavior reputation and resource contribution, the adaptive election of consensus Committee is realized, and witch attacks and "hitchhiking" behaviors are effectively suppressed; The hierarchical consensus architecture is designed, which divides the global consensus into two stages: local fast consensus and global final confirmation, so that the communication complexity is reduced from O(n²) to O(n), and the system throughput is up to thousands of TPS by combining parallel sharding technology. At the same time, zero knowledge proof (ZKP) is introduced to realize the verification of offline calculation, which can protect data privacy and reduce the burden of online calculation. Intelligent routing and dynamic partition recovery strategy are adopted to ensure the high availability of the system when nodes join/exit dynamically. The experimental results show that the throughput of the scheme is stable at 4,900 TPS, the average delay is controlled in the order of 100 milliseconds, and malicious nodes can be quickly identified and isolated, which is significantly superior to traditional algorithms such as PBFT and PoS, and provides an efficient, safe and scalable solution for large-scale data asset registration.