Abstract: Abstract: In view of the huge amount of data and high concurrency access requirements in power scenarios, and in view of the limitations of the Raft consensus algorithm used in distributed storage systems in serial submission, this paper proposes a performance optimization algorithm for distributed power storage systems, the Parallel Raft algorithm (PRaft), which integrates the leader election strategy in the Paxos algorithm and the Raft algorithm, which not only ensures the consistency of logs. Moreover, it effectively reduces the delay of the client and the probability of conflict in the consensus process. In addition, the PRaft algorithm realizes the parallel execution of log entries by finely managing the dependencies between logs, which significantly enhances the parallel processing capability of the system and greatly improves the execution efficiency. Experimental comparison and analysis show that the overall throughput of the proposed algorithm is increased by 20%~100% and the throughput of log submission is increased by 52.8% compared with Raft.。

Key words: power scenarios, distributed storage, Raft algorithm

摘要： 摘 要: 面向电力场景庞大的数据量和高并发的访问需求，针对分布式存储系统采用的Raft共识算法在串行提交方面的限制，本文提出了一种针对分布式电力存储系统的性能优化算法—并行优化的Raft算法（Parallel Raft，简称PRaft），该算法整合了Paxos算法与Raft算法中的领导者选举策略，这一结合不仅保障了日志的一致性，而且有效降低了客户端延迟和共识过程中的冲突概率。此外，PRaft算法通过精细化管理日志间的依赖关系，实现了日志项的并行执行，从而显著增强了系统的并行处理能力，并极大提升了执行效率。实验对比分析表明，所提算法相比Raft的总体吞吐量提升了20%~100%，日志提交吞吐量提升了52.8%。

关键词: 关键词: 电力场景, 分布式存储, Raft算法