Research On Scalability Of Asynchronous Blockchain Based On DAG

In Internet of Everything(Io E),the dynamics and uncertainty of complex and asyn-chronous interactions directly determine the difficulty of data circulation and consensus in data collaboration.In an asynchronous environment,there is a lack of natural mutual trust among cooperative entities.By taking advantage of the open,autonomous,tamper-proof and decentralized characteristics of blockchain technology,a new model of data collaboration can be established at low cost in an untrusted environment.However,asyn-chronous data collaboration still faces the following challenges: First,compared with the synchronous model,the asynchronous model does not rely on time assumptions and is suitable for dynamic and unpredictable network environments.Due to the “FLP Impos-sibility”,it is difficult to reach synchronous and deterministic consensus in asynchronous model system.Second,blockchain with traditional chain structure is challenged by low efficiency,high cost and poor scalability.Current solutions in blockchain scalability,such as improved consensus mechanisms,Directed Acyclic Graph(DAG)technology,have some limitations and need further improvement.Third,it is not enough to improve scalability when applied to real scenarios.Therefore,there is also a need for scientific three-dimensional trade-offs based on requirements,under the famous “the Impossible Triangle” constraints.In view of the above challenges,the main work and contributions are summarized as follows:(1)To solve the trust,cost and efficiency problems of complex and asynchronous data collaboration,we propose Asycome,an asynchronous data collaboration mechanism based on blockchain.The mechanism takes advantage of the new “Gossip about gos-sip” protocol,which is driven by message-transmitting events.Transactions on the same time sequence can proceed to the next transaction without reaching the final confirmation state,constantly maintaining data synchronization.According to the event and commu-nication process through voting and committing of two phase agreement to ensure that the final asynchronous consensus.We design and implement a prototype system,test its performance in terms of key indicators.We qualitatively analyze its liveness and safety.The results show that this mechanism can realize data collaboration efficiently and at low cost in asynchronous environment with no guarantee of message delay,and has certain scalability.(2)Through experiments,we found that Hashgraph,the representative algorithm of DAG technology,was not scalable enough and needed to be further improved.We pro-pose an asynchronous consensus algorithm based on DAG named Asygraph.The algo-rithm adds the weight attribute to each node to distinguish the performance difference of different nodes.According to the cumulative weight,probabilistically selecting neighbors for communication,in order to reach a consensus quickly.Introduce the baseline event and chains in the “virtual voting” for querying historical records.In order to evaluate the performance of the algorithm,the experiment compared throughput and latency of the two algorithms in terms of scalability and qualitatively analyzed the safety and liveness of the algorithm.The results show that the latency and throughput of the algorithm are improved by multiple times under the premise of ensuring a certain degree of security and decentralization.Asygraph has better performance than Hashgraph,improved scalability,and the three-dimensional tradeoff is realized.To sum up,the research of this paper helps to improve the application barriers and performance deficiencies of the existing consensus algorithm of blockchain in the asyn-chronous environment.Moreover,it has rich theoretical research value and wide applica-tion prospect,and can better solve practical problems.It provides key technical support for its final practical application.