基于动态视图的数据库性能调优  

DATABASE PERFORMANCE TUNING BASED ON DYNAMIC VIEWS

作　　者：徐增敏[1] 张昆 丁勇[1] 赵汝文[1] 

机构地区：[1]桂林电子科技大学数学与计算科学学院,广西桂林541004 [2]广州从兴电子开发有限公司,广东广州510300

出　　处：《计算机应用与软件》2012年第12期58-60,共3页Computer Applications and Software

基　　金：国家自然科学基金项目(60963024)

摘　　要：针对数据库可能出现负载异常波动影响性能的问题,可使用外部图形化工具监控和优化数据库,但是这些软件在服务器运行将会占用大量资源,而且很多调优方法都相对片面缺乏整体思路。鉴于此,提出一种基于动态视图的数据库性能调优方法。该方法从系统性能参数入手,通过建立全局性能基线,结合动态视图分析资源耗费原因,研究如何降低执行开销。实验选用示例数据库的12万条记录进行实例验证,结果表明该方法可以有效解决大规模数据处理的实际效能问题。For the issue that the abnormal fluctuations of load may affect the performance of database, external graphical tools for monitoring and database optimisation can be employed, but the operation of these software in server would waste a lot of system resource. Moreover, many tuning methods are also relatively one-sided and short of overall view. Therefore, a database performance tuning method based on dynamic views is proposed. This method starts from system performance parameters, by establishing global performance baselines, and in combination with dynamic views to analyse the reasons of resource consumption, it studies how to reduce the cost of execution plan. 120,000 records are chosen from SQL Server 2008 AdventureWorks database to make examples verification, and experimental＂ resuhs show that the method can effectively solve the practical performance problems in large-scale data processing.

关 键 词：数据库管理员 性能基线 动态视图 执行计划 非聚集索引 查询优化器 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]