Research On Sharding And Consensus Technologies For Scalability Of Blockchain

The advent of Bitcoin has attracted widespread attention to the blockchain.Its decentralized and non-tamperable features facilitate multi-party data sharing and value circulation.It is regarded as a supporting technology for building a large-scale Internet of trust,which has attracted many enterprise and organizations in various business fields have explored its application methods,and a large number of Dapps have flourished.However,Bitcoin and Ethereum,the largest blockchain infrastructure at present,have the performance bottlenecks and cannot meet the needs of large-scale Internet applications.Therefore,scalability has always been a research hotspot in the blockchain field.Researchers at home and abroad have proposed many solutions to this problem,from the initial relaxation of Bitcoin parameter restrictions to the improvement of the consensus mechanism,to the layering of the overall blockchain architecture,but there is no perfect solution.The current blockchain scalable solution still has scalability issues,including low system throughput,long transaction confirmation latency,and increasingly high resource costs for consensus.In order to solve the above problems,this thesis conducts research on related technologies of sharding and consensus for blockchain scalability.The main research content includes the following two aspects:1.Aiming at the problems of insufficient system throughput and low decentralization caused by high consensus cost,this thesis designs a Po S-based dynamic validator sharding scheme.The sharding architecture is structured into static sharding based on user groups and storage nodes,and dynamic sharding based on validator nodes.Static sharding splits the global ledger data,and only part of the ledger information is stored inside the shard,thereby reducing the consumption of storage resources.Dynamic sharding allows the validator to rotate randomly among the shards,avoiding some constant validators from controlling a certain shard for a long time.Each shard can process transaction transactions in parallel,which could improve throughput.This thesis designs a cross-chain mechanism for recursive state transfer for dynamic sharding,so that dynamic validator can validate transactions between shards only by the shard status proof which was much smaller than the ledger data,which reduces the threshold for participating in consensus and guarantees decentralization.At the same time,by the introduction of a Po S-based token economy method,encourages the honest behavior of nodes and guarantees the security of the system through token deposits and evil penalties and forfeiture measures.2.Due to the communication message is relatively complex,the traditional Byzantine consensus algorithm does not have scalability,which is not conducive to deployment in largescale blockchain networks,and another problem is the lack of adaptive ability when the algorithm is applied to a blockchain system with complex node components.In order to solve the above problems,this thesis proposes a dynamic byzantine consensus algorithm based on signature aggregation.First,the threshold signature is constructed using the BLS signature aggregation technology.In the consensus process,the threshold signature is used to compress the interactive consensus state and reduce the complexity of the consensus message,which help expand the scope of consensus and enhance scalability.Then a dynamic node evaluation management system is established to enhance the adaptive ability of the algorithm,which including the evaluation mechanism based on the node evaluation table,and the management mechanism based on the view self-adjustment protocol and the node removal protocol.The evaluation mechanism uses the node evaluation table to record the evaluation of the historical consensus behavior of the nodes,and it is dynamically updated in each consensus process.Finally,the management mechanism is used to remove the nodes with low evaluations from the system,which reduces the adverse effects of failed nodes on the system,which optimizes composition structure of consensus nodes and ensure the efficiency and stability of the consensus.