Research Of Paxos Consensus Algorithm In Token-less Blockchain

In recent years,major information technology powers in the world are speeding up the development of blockchain technology.The application of blockchain technology has extended to digital finance,mobile communication,Internet of things,supply chain management,digital asset trading and other fields.The research and application of various blockchain technologies,such as public chain,private chain and alliance chain,have attracted the attention of government departments,research institutions and commercial companies.Xi Jinping,President of People's Republic of China,emphasized that the integrated application of blockchain technology played an important role in new technological innovation and industrial transformation,we should take blockchain as an important breakthrough in independent innovation of core technologies,increase investment,strive to tackle key core technologies,and accelerate the development of blockchain technology and industrial innovation.Token-less blockchain refers to a blockchain that does not rely on token for value exchange.As a kind of blockchain technology,its main function is to ensure the consistency and non-tamperability of data records in decentralized distributed system.The key technologies of token-less blockchain include consensus mechanism,smart contract and encryption algorithm.In token-less blockchain without Byzantine problem,consensus algorithm is basically equivalent to distributed consistency algorithm(hereinafter referred to as: consistency algorithm).At the same time,the consistency mechanism still faces various problems and challenges.First,some consensus mechanisms have the problems of slow consensus building speed and high communication load.Second,there is a lack of multi-dimensional security analysis for consistency algorithm.Third,the consistency algorithm involving leader election has the hidden trouble of manipulating the election process and cannot ensuring decentralization.This paper will be required from the above three aspects.Firstly,by improving the classical Paxos consistency algorithm,which has the widest influence,it speeds up the time of reaching consensus,reduces the probability of "livelock",cuts down the communication cost.Then,by using game theory it analyzes the security and reliability of Paxos algorithm,and puts forward the problem of possible computing resource dilemma.It analyzes that Paxos algorithm can run continuously and reliably only when it meets certain security conditions,and proves the correctness of Paxos mechanism design by using game theory.Finally,analysis shows that the consistency algorithm involving Leader election has the problem of destroying decentralization by manipulating Leader election,and proposes a new election mechanism that can effectively reduce such security issues.The main research contents and contributions of this paper include the following three aspects:(1)Improving Basic Paxos algorithm by optimizing the role behavior of the algorithmIn order to improve the running efficiency of the consistency algorithm,this paper analyzes the behavior of each role in each phase of Basic Paxos algorithm,and proposes four improvement measures to optimize the behavior of Proposer and acceptor roles in the original algorithm on the premise of inheriting the basic mechanism design of the original algorithm.Firstly,the specific behavior of each role in each phase of the original algorithm is analyzed.Then,the original algorithm is optimized by distance association waiting mechanism,sending the approval proposal to acceptors in stages,broadcast reduction and acceptors self-learning.The correctness and security of the improved algorithm are proved theoretically,and the number of message passing between the two algorithms in normal execution is compared and analyzed.Finally,the fault tolerance,running time,number of message passing,number of proposal submission failures and average number of proposal submission per second of the two algorithms are analyzed and compared from the experiments.Compared with Basic Paxos,Adv Paxos proposed in this study can effectively improve the speed of reaching consensus among server clusters on the basis of the same fault tolerance under the same cluster scale,and has lower communication load and higher operating efficiency.Adv Paxos inherits the mechanism design of Basic Paxos,it is a supporting algorithm that needs to be called during the operation of other Paxos like engineering implementation algorithms,and has high scalability,so the optimized algorithm will also play a good role in improving the execution efficiency of other Paxos like engineering implementation algorithms.(2)Analyzing the security of the distributed system using Paxos algorithm by game theoryFrom the perspective of the safe operation of the distributed system using Paxos algorithm,this paper analyzes the game relationship among the nodes in the typical system,and finds that each node will emerge a selfish behavior when the system is running:To get data updates,but do not to be elected leader and excessive consumption of their own computing resources.Because of this selfish behavior,the system will form a kind of game which is called computing resource dilemma game in this paper.Then,by analyzing the payoff matrix and establishing the utility function model,it is found that Pareto optimal Nash equilibrium in the game needs to meet certain security conditions.If the security conditions are not met,each game node will make the uncooperative choice,and the uncooperative choice may cause chain reaction,which will cause the system to work abnormal eventually.Finally,by solving the perfect Nash equilibrium solution of the subgame in the game,the security conditions for achieving the purpose(nodes no not deviate from the set trigger strategy or two-phase strategy)of Paxos mechanism design are obtained.And the effectiveness of the security conditions is proved by experiments.This study is helpful for the designers who adopt Paxos algorithm alliance chain to establish the system access conditions,as well as the enterprise who want to join the system to weigh the actual benefits and consumption costs.(3)Analyzing and researching the more secure election mechanism of minority win in the consistency algorithmThe purpose of this work is to analyze the security of the leader node generation mechanism,find out the hidden danger of losing the decentralized attribute of the system due to the control of voting in the traditional majority win election,and propose the more secure leader node election mechanism of phase out the majority with the same choice(minority win).First,by studying the attack methods,attack effects and the possibility of attack success that minority win election mechanism may face,such as absolute control attack,team attack,fraud attack.Then,it is concluded that there will be a phenomenon of low control rate and the regularity of low control rate when the number of nodes is certain.And the implementation approches of the minority win election mechanism in the leader election based consistency algorithm are described.Some difficult problems such as liveness,fault tolerance and final are solved.Finally,through theoretical analysis and experiments,it is further verified that this kind of election mechanism has better security in the face of absolute control attack,compared with the traditional majority win election mechanism,without significantly increaseing the election time.In the common blockchain distributed system(more than 4 nodes),in order to complete the absolute control attack,adversariesneed to control more than 80%of the nodes.Compared with the control rate of 50% of majority win,minority win can increase the attack difficulty by at least 60%.Moreover,by enumeration analysis,we can see that the difficulty of absolute control attack in minority win elections increases with the number of participating nodes.