Research On Optimization Model Of Consortium Blockchain Capacity

The blockchain technology challenges the traditional trust system based on its decentralization and cryptographic endorsement,changing the way of value sharing and promoting the transformation of the current information Internet to the value Internet.However,there are still some problems with the blockchain technology.The current urgent one is the problem of excessive system storage overhead caused by the characteristics of non-erasable and distributed storage.Based on analyzing the existing optimization schemes of blockchain capacity,the thesis summarizes their advantages and disadvantages,and proposes an optimization model of storage capacity on the consortium blockchain.In the model,the blockchain network is divided into groups by the characteristic of "partial decentralization" of the consortium blockchain.In each group,the full node is selected according to the hardware performance and accumulated reliability evaluation of the nodes.The full node is responsible for storing the complete blockchain data,and other ordinary nodes store the block data according to the consistent hash principle and the hash chain formed by each block header.And the model takes the difference in the performance among ordinary nodes into account and uses the virtual nodes to dynamically allocate the amount of data stored based on the performance of the nodes to achieve load balancing.Finally,the model designs a full node replacement strategy to ensure that the model can resume normal operation when the network is attacked.The thesis simulates the proposed model in the laboratory environment and tests the storage capacity,reliability and transaction verification of the model respectively.According to the experimental results,the following conclusions are obtained.In the aspect of blockchain data storage,the model optimizes the storage structure,improves the storage space compared with the original blockchain system,and the load of nodes is balanced by allocating block data according to hardware performance.In the aspect of system reliability,the model can resume normal state after being attacked and has the anti-attack capability.In the aspect of transaction verification,the verification time of the non-full nodes in the model is slightly longer,but it has little effect on system performance.Therefore,the thesis model is reasonable and feasible.