FabricETP: A high-throughput blockchain optimization solution for resolving concurrent conflicting transactions

【Author】 Wu, Haibo; Liu, Hui; Li, Jun

【Source】PEER-TO-PEER NETWORKING AND APPLICATIONS

【影响因子】3.488

【Abstract】As one of the emerging technologies that have attracted much attention, blockchain has a wide range of application prospects. However, one prominent issue restricting its development is the limited transaction throughput. Take the well-known open-source system Hyperledger Fabric as an example. In the solution of parallel transaction processing, it follows the processing method of MVCC in the database, which is not compatible with blockchain features, resulting in a sharp drop in throughput and wastes of storage space in high concurrency scenarios. In this paper, we propose a high-throughput optimization scheme FabricETP that can effectively solve concurrency conflicts. According to the different causes of conflicts, FabricETP proposes optimization ideas from two dimensions. For transaction conflicts within a block, FabricETP proposes a scheduling algorithm to minimize the number of conflicting transactions by rearranging the transaction execution order. For transaction conflicts between blocks, FabricETP has established a cache-based conflict transaction avoidance mechanism, so that invalid transactions are aborted early. With the help of general blockchain performance testing tools, we carry out experiments under various workload scenarios. The results show that the throughput of FabricETP can reach up to 9.51 x that of the benchmark blockchain system Fabric and 1.26 x of the optimized version of Fabric + + under high concurrency scenarios. Compared with Fabric + + , the space utilization is increased by 20%.

【Keywords】Blockchain; Hyperledger Fabric; Concurrency conflict; Transaction; Performance optimization

【发表时间】

【收录时间】2023-02-19

【文献类型】实验仿真

【主题类别】

区块链技术-平台项目-Hyperledger Fabric