ML_Raft算法在mysql分布式集群中的研究  

Research of ML_Raft Algorithm in Mysql Distributed Cluster

作　　者：程书玲 Cheng Shuling(Zhejiang Industry&Trade Vocational College,Wenzhou 325000,China)

机构地区：[1]浙江工贸职业技术学院,浙江温州325000

出　　处：《黑龙江科学》2023年第4期78-79,83,共3页Heilongjiang Science

基　　金：2022年浙江工贸职业技术学院院级课题“ML_Raft算法在mysql分布式集群中的研究”(纵20220005)。

摘　　要：为解决传统的Raft算法存在的数据库同步时间过长、单领导者影响集群性能和稳定性等问题,基于mysql分布式集群,分析了数据一致性问题,提出新的改进算法设计——ML_Raft算法,建立了计算集群与数据同步集群,每个集群功能不同,提高了集群响应能力,降低了每个集群节点的服务压力,提高了集群性能。结合mysql集群的应用环境特点,将计算节点和同步节点分开,提出了多个领导者节点,以保证集群的可靠性,加快了响应速度,提高了用户感知。In order to solve the problems of long time of database synchronization,single leader influence cluster performance and stability,the study analyzes data consistency based on mysql distributed cluster;proposes new improved algorithm design,i.e.ML_Raft algorithm;establishes calculation cluster and data synchrony cluster,with different functions in each cluster,to improve cluster response capacity;reduce the service stress of each group node;and improve cluster performance.Through combining with the application environment characteristics of mysql cluster,the study divides calculation and synchrony nodes,and proposes many leader nodes to guarantee the reliability of cluster,accelerate response speed,and improve users’perception.

关 键 词：日志同步 集群架构 性能瓶颈 数据一致 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]