Design And Implementation Of High-throughput Blockchain Based On Time Synchronization

With the deepening of the research on blockchain technology,some problems faced by blockchain technology are gradually exposed,such as low transaction throughput and excessive consumption of blockchain capacity.Among them,the most serious constraint on the development of blockchain technology is the problem of too low transaction throughput.The average throughput of Bitcoin is only 7 transactions per second,and the throughput of Ethereum is only 15 transactions per second.Existing methods for solving transaction throughput include methods such as transforming the block structure,offchain transactions,and reducing consensus nodes.However,the transformation of the block structure does not solve the problem well,and at the same time,some new security and verification transaction issues are generated;although off-chain transactions and reducing consensus nodes can solve the problem,the centralization strategy of these two methods will make the blockchain lose its highly decentralized characteristics.At the same time,the existing blockchain has a too high tolerance for the timestamp error of the nodes,making transactions unable to be sequenced according to the actual occurrence time,causing the blockchain to lose its development prospects for real-time applications.This article hopes to design a new consensus algorithm that simultaneously solves the two problems of transaction throughput of the blockchain and the ordering of transactions according to the real occurrence time.Starting from the node timestamp,this thesis designs a proof of time consensus algorithm(PoT).The algorithm is divided into two parts.The first part is to synchronize the time of the nodes in the entire network,and the second part is to design and complete the consensus algorithm by time-synchronized nodes.The node time synchronization part uses experiments to test the algorithm time consuming under different node sizes,which proves that the node time synchronization algorithm designed in this thesis still has high efficiency when the number of nodes increases exponentially.The proof of time consensus algorithm involves the election of validating nodes and the voting process of all nodes.Through the proof of time consensus algorithm,the system time of the entire network node is kept synchronized,so that the transactions can be sorted according to the actual occurrence time,that is,the transactions stored in the block are strictly increased according to time.At the same time,the consensus algorithm has a relatively large throughput improvement over the traditional consensus algorithm.At the scale of 100 nodes,it can process up to 48832 transactions on average in 10 seconds,and the transaction throughput exceeds 4800 transactions per second.Through analysis,we can use super nodes or use blockchain sharding technology to improve node performance and further increase transaction throughput.the amount.In terms of the security of the algorithm,the use of authenticated nodes for consensus can effectively avoid Sybil Attack,and the experimental results of algorithm tolerance show that the consensus algorithm proposed in this thesis generally has a tolerance of more than 60%,which is higher than that of voting consensus and proof consensus.Therefore,the time designed in this thesis proves that the consensus algorithm can improve transaction throughput,and also provides ideas for the development of blockchain in real-time applications.