ITPub博客

首页 > Linux操作系统 > Linux操作系统 > 测量、基线和性能优之一:测量和性能

测量、基线和性能优之一:测量和性能

原创 Linux操作系统 作者:sunsapollos 时间:2013-10-22 19:59:24 0 删除 编辑
       要让Oracle性能优化从艺术变成科学,首先必须建立目标系统性能指标的可以测量。一个无法测量的系统或者说一个只能依赖于人的眼睛,耳朵等器官来进行感知的系统是无法进行性能优化的,这个时候就只能依赖于直觉,性能优化就变成了艺术或者所谓的超能力了。
      为了完成性能优化,我们需要完成大量的可测量性工作。幸运的是,Oracle对于可测量的性能一直付出了艰辛的努力,使其性能相关的测量指标远远的超出了其他数据库。
      可测量的性能优化目标:吞吐量和响应时间。
      我们可以从以下几个方面来建立基于Oracle的测量体系:
     (1)、基于终极目标性能的测量
     (2)、基于Oracle构成部件运行是否良好的测量
     (3)、基于物理资源供应运行是否良好的测量
     (4)、基于并发性资源运行是否良好的测量
     (5)、基于流程分解的响应时间的测量
     
      从另一个角度考虑,我们建立起终极目标性能指标:吞吐量和响应时间的因果分析测量体系和相关性分析测量体系。基于因果关系确认的艰难性,我们更多的基于相关性进行分析测量。
     我们来看下面的简单例子:
     我们这里从简单业务出发来考虑问题,比如全部业务为tpcc测试的订单业务。

    吞吐量: 每分钟完成的订单数。

    吞吐量:= ( 60/每个订单处理的响应时间 ) * 并行的处理会话
   
    可以认为以下两个结论是正确的: 在并行处理会话确定的前提下,降低每个订单处理的响应时间可以提高吞吐量。
                                                                 在每个订单处理的响应时间确定的情况,增加并行的处理会话可以提高吞吐量。
    具有足够经验的DBA都知道,上面的结论完全是理论推导,在实际的环境中,所谓的吞吐量下降的场景,每个订单处理的响应时间延长,几乎总是可以观察并行的处理会话数量增加。我们甚至可以如此认为,在业务量变化不大的前提下,并行的处理会话增加几乎必然意味着吞吐量的下跌,每个订单处理的响应时间延长导致并行的处理会话增加。
    从以上分析,我们可以具有以下结论,在业务变化不大的前提下:
    订单处理的响应时间延长导致吞吐量的下跌。
    订单处理的响应时间延长导致并行的处理会话增加。
    并行的处理会话增加和吞吐量降低具有相关关系。
  
    我们为什么需要这样来描述?原因很简单,每个订单处理的响应时间是相对难以被测量的指标,而并行的处理会话极其容易被测量。
   
    订单的响应处理时间:=  订单的处理时间  + 订单的等待时间
    订单的处理时间和订单的等待时间,Oracle都在系统级别做出了很好的测量。我们再次来看吞吐量曲线:
         
 
data:image/png;base64,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

    假设吞吐量曲线图中蓝点和黄点是响应时间和吞吐量的最佳平衡点,在这个平衡点具有服务时间和排队时间。当我们测量的排队时间超过这里的排队时间,以为着响应时间出现问题。
   
    订单的处理时间:=  处理次数 * 每次处理时间+处理次数*每次处理时间+.....
    订单的排队时间:= 排队次数*每次排队时间 + 排队次数*每次排队时间+.....

    这里可以注意一下:当处理次数发生明显变化,意味着业务特征或者访问特征发生了变化。对于任何性能优化DBA来说,这都是一个性能的直接要素。每次处理时间,在Oracle数据库的描述,服务处理是由CPU来执行,正常情况下应该保持稳定,一旦其发生明显变化,意味着CPU无法提供足够的能力。
    排队次数同处理次数一样,排队次数同样表示着业务特征或者访问特征,排队时间表示访问能力。
    作为Oracle数据库系统的性能优化工作者极其幸运,响应时间的分解几乎都可以直接或者间接通过测量获得,从而使Oracle数据库系统完全是一个优化就绪的数据库系统。

  
    
            

来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/92650/viewspace-774836/,如需转载,请注明出处,否则将追究法律责任。

请登录后发表评论 登录
全部评论
专注于Oracle,BI,Security,DR &^BCP,Performance tuning

注册时间:2013-10-15

  • 博文量
    68
  • 访问量
    726256