摘要
为了规避数据库集群中部分节点过载风险,提升分布式数据库的可靠性,分析了HBase数据库的数据格式及总体架构。通过综合对比各类实现方法,选择成本低易实现的软件算法实现负载均衡。针对原有负载策略的缺陷设计了基于热点访问的改进算法,综合考虑region迁移成本、region本地化成本与集群负载情况后选择迁移策略。在VMware搭建环境,利用shell读写数据进行分析,验证结果表明改进算法无论在region count还是request方面均衡效果均更佳,为HBase分布式数据库的负载均衡策略提供了可参考方案。
In order to avoid the overload risk of some nodes in the database cluster,and improve the reliability of the distributed database,the data format and overall architecture of the HBase database are analyzed.Through comprehensive comparison of various implementation methods,the software algorithm with low cost and easy implementation is selected to achieve load balancing.Aiming at the defects of the original load strategy,an improved algorithm based on hot spot access is designed,and the migration strategy is selected after considering the migration cost of region,localization cost of region and cluster load situation.An environment is built in VMware,and the shell is used to read and write data for analysis.The verification results show that the improved algorithm has better balancing effect in both region count and request,which provides a reference scheme for load balancing strategy of HBase distributed database.
作者
李峻屹
LI Junyi(Department of Information Technology,Shaanxi Police College,Xi’an 710021,China)
出处
《微型电脑应用》
2022年第5期138-141,共4页
Microcomputer Applications
