BlogJava-java神谕-随笔分类-性能分析http://www.blogjava.net/javaora/category/1887.htmlzh-cnTue, 13 Nov 2007 12:01:53 GMTTue, 13 Nov 2007 12:01:53 GMT60混淆器<>忽悠http://www.blogjava.net/javaora/archive/2007/11/12/159989.htmljava世界畅谈java世界畅谈Mon, 12 Nov 2007 07:57:00 GMThttp://www.blogjava.net/javaora/archive/2007/11/12/159989.htmlhttp://www.blogjava.net/javaora/comments/159989.htmlhttp://www.blogjava.net/javaora/archive/2007/11/12/159989.html#Feedback0http://www.blogjava.net/javaora/comments/commentRss/159989.htmlhttp://www.blogjava.net/javaora/services/trackbacks/159989.html 混淆是好东西，但也不是绝对的安全，而且对于庞大的系统本身而言，需要keep的元素以及与jsp等交互共同的信息量也很大。混淆不是忽悠，忽悠.....

]]>10条通向快速Web 站点现实的步骤http://www.blogjava.net/javaora/archive/2006/02/11/30257.htmljava世界畅谈java世界畅谈Sat, 11 Feb 2006 05:48:00 GMThttp://www.blogjava.net/javaora/archive/2006/02/11/30257.htmlhttp://www.blogjava.net/javaora/comments/30257.htmlhttp://www.blogjava.net/javaora/archive/2006/02/11/30257.html#Feedback0http://www.blogjava.net/javaora/comments/commentRss/30257.htmlhttp://www.blogjava.net/javaora/services/trackbacks/30257.html1.确定瓶颈
 1.1文件大小
   在需要使用这个页面时用户需要加载多少数据。、
   60-100K 页面大小 保持教小的文件
 1.2反应时间
   你向服务器发送请求和数据到达PC的时间间隔
   网络反应时间
2.减少文件大小
  100K以上
  大量的文件当前主要是由巨大的Javascript类库。
  巨大的图片
  老的HTML样式也会产生大文件，尽可能地使用XHTML和CSS
  压缩
3.查找什么导致了很高的反应
  如上我们所提到的，反应时间主要由两个主要的元素
  3.1网络反应时间
  3.2是否花费了太长的时间来产生页面
  3.3性能
4.确定绝缘组件
5.编译缓存
6.查看DB查询
  避免join. 查询缓存
7.发送正确的已修改数据
8.考虑组件缓存
9.减少服务加载
  9.1使用相反的代理
  9.2采用轻量级的HTTP服务器
10. 增加服务器


]]>SQL Profiler Introductionhttp://www.blogjava.net/javaora/archive/2005/06/28/6810.htmljava世界畅谈java世界畅谈Tue, 28 Jun 2005 06:57:00 GMThttp://www.blogjava.net/javaora/archive/2005/06/28/6810.htmlhttp://www.blogjava.net/javaora/comments/6810.htmlhttp://www.blogjava.net/javaora/archive/2005/06/28/6810.html#Feedback1http://www.blogjava.net/javaora/comments/commentRss/6810.htmlhttp://www.blogjava.net/javaora/services/trackbacks/6810.htmlSQL Profiler Introduction

This is a quickly hacked tool to do statistics（统计） on SELECT queries in order to know where it is most efficient to create indexes.  目的是最有效的创建索引.

This small tool, released under an Apache-based license connects to the P6Spy JDBC logger and displays in real time the queries going to the database. It uses an integrated SQL parser to build statistics on the most accessed tables and columns and can generate SQL index creation files. Other information is also gathered and displayed, such as the request time for a single request, for a class of request, and for all the requests. Sorting may be done on these views to detect(察觉) database problems efficiently.


This tool can be very useful when you have a big volume of queries that you need to analyze not one by one一个接一个 (meaning that the specific time isn't that much of interest), but rather(若非) when you want to know what "group" of queries is taking a lot of time, such as queries on the same tables and columns but with different query values. The integrated SQL parser (built with ANTLR) is used to analyze the incoming SELECT queries.

The Swing GUI was based on Apache's Log4J Chainsaw, but all the bugs are mine. Also contributors are welcome to test, make new suggestions, give their opinion and submit patches.



]]>