Data Consistency Detection And Consensus Mechanism Of Blockchain System

There is data inconsistency in blockchain systems,which is caused by transactions.This master dissertation focuses on work and its results of data consistency detection and its guarantee mechanism which are as follows.(1)Aiming at data preprocessing of data consistency detection,a "partition and data slicing algorithm" is proposed to realize the data partition and slicing of consistency detection,in accordance with consistency constraints and equivalence relationships being determined,so as to obtain an appropriate data consistency detection granularity.Aiming at imbalance of node computing power in blockchain systems,a "centralized data consistency detection algorithm based on task concentration",namely "detection algorithm A",is proposed.On the basis of a response time cost model of data consistency detection being established and its task allocation process being optimized,a "task allocation optimization strategy" and its corresponding "task allocation optimization algorithm" are proposed.A "data consistency optimization detection algorithm" based on "task allocation optimization strategy",namely "detection algorithm B",is proposed,according to balance of computing power of blockchain system nodes.According to recall rate,precision rate and detection response time cost being selected as evaluation indexes of data consistency detection algorithms,the recall rate,precision rate and detection response time cost of the above mentioned "detection algorithm A" and "detection algorithm B" are measured respectively under different data sets.The experimental results show that the "data consistency optimization detection algorithm" is better than the "centralized data consistency detection algorithm".The data consistency detection effect of blockchain systems can be further improved by means of balancing computing load of their nodes.(2)Aiming at low consensus success rate,low consensus efficiency and poor scalability of the "PBFT algorithm",a "blockchain system global consensus mechanism based on PBFT " is proposed,on the basis of its consensus formation algorithm,which includes node cluster being grouped,node credit score being evaluated,and participant node being selected in the above mentioned global consensus mechanism based on its node credit score in its group.On this basis,a "blockchain system global consensus algorithm based on PBFT",namely "I-PBFT algorithm",is proposed to realize the above mentioned "blockchain system global consensus mechanism based on PBFT".Under the condition of the consistency,effectiveness and integrity of the "PBFT consensus mechanism" being meted,the number of nodes participating in the consensus in each stage is reduced,so as to resist risk of malicious nodes' interfering with blockchain systems,and to for throughput and system efficiency of consensus network systems be improved.The consensus delay,throughput and reliability of consensus nodes of the network system on the "I-PBFT algorithm" and "PBFT algorithm",are tested respectively.The experimental results show that network system performance indexes on the "I-PBFT algorithm" are better than ones on the "PBFT algorithm".Thus an optimization tool for data consistency guarantee of blockchain systems is provided by the "I-PBFT algorithm".In summary,the data consistency detection algorithm and the global consensus algorithm proposed in this master dissertation can be applied to data maintenance,management and analysis of blockchain systems.