Research On Parallel Architecture Of Large-Scale Blockchain Transactions For Complex Assets

With the continuous promotion of Digital China Strategy and Trusted Society System,more and more fields have started to introduce blockchain technology,and the parallel demand for large-scale transactions of complex assets is getting stronger.However,the existing blockchain system cannot efficiently support this scenario,mainly because the granularity of the existing transaction conflict detection method stays at the deeply coupled "Account-AssetContract" level,which cannot effectively determine the conflict between transactions involving complex assets,resulting in the inefficient operation or even failure of many existing transaction sorting mechanisms,and the throughput performance cannot meet the demand of realistic scenarios.The main work of this paper is as follows:(1)A fine-grained transaction conflict detection method is proposed.To solve the problem that existing transaction conflict detection methods are not applicable to transactions involving complex assets,this paper firstly constructs a transaction model based on account separation,which makes asset storage not dependent on contracts,and secondly constructs a finite state machine model CA-FSM for complex assets,which describes and restricts the state and state transitions of complex assets,and proposes a conflict detection method based on this model that refines the granularity of conflict detection to the asset The conflict detection method is proposed to refine the granularity of conflict detection to the asset level.At the same time,to reduce the extra storage overhead brought by the CA-FSM model,a storage optimization strategy based on state pruning is proposed in this paper,and finally,the effectiveness of the CA-FSM-based transaction conflict detection method and storage optimization strategy is demonstrated experimentally.(2)A hierarchical large-scale parallel architecture based on consensus groups and execution groups is proposed.To solve the problem that the existing parallelism mechanism is not applicable to complex asset scenarios,this paper firstly improves the transaction dependency graph,constructs a transaction dependency graph based on asset states,and proposes a conflict-minimizing transaction dependency graph partitioning strategy to reduce the transaction conflict rate between blocks.At the same time,this paper improves the transaction parallel execution capability of blockchain systems from two aspects.On the one hand,through the asynchronous pipeline based execution group parallel mechanism and the node function decoupling based consensus group execution group parallel mechanism,the nonblocking parallel execution of transaction execution processes is achieved.On the other hand,through the state integration based intra node parallel mechanism,the transaction parallel capability of a single node is improved.Finally,it is experimentally demonstrated that the throughput performance of the cascaded parallel architecture in this paper is better than that of existing blockchain parallel architectures.(3)A blockchain prototype system for complex assets is designed and implemented.In order to make the users of blockchain system better understand the ideas and solutions in this paper,this system implements a blockchain prototype system supporting fine-grained transaction conflict detection and cascading parallelism for complex assets.This system visualizes the processes of transaction dependency graph construction,segmentation and transaction parallel execution,and also implements several functional modules such as transaction initiation and account management,which are of practical significance for the practical application of blockchain to complex asset scenarios.