Realization of real-time updating of Web site cache data

The implementation of a Web site will inevitably need to deal with a large number of memory data, which are mostly due to the performance of the site and solve I/O bottlenecks, such as the Javaeye site's home page ranking data and blog channel rankings and other data, have been cached processing, and is not real-time, In particular, the first page of the ranking data is almost a day will be updated, similarly, QQ and other portal sites also exist a large number of static memory data, so a Web site cache processing is critical, and good architecture needs to do code without intrusion, that is, memory updates, log records and other operations should be "cut" processing, In my later web site development, I generally define a caching class, as follows:

Java code

public　static　LinkedList<Login>　regRecentUser=new　LinkedList();　//最近注册人员
　public　static　LinkedList<Login>　regLoginedUser=new　LinkedList();//最近登录人员
　public　static　List<Topic>　hotTopic=new　ArrayList();//最热发表话题

And will implement the object of the sorting interface, such as the hottest topic, may be subject to a certain number of attributes of the weight distribution to achieve, similar code is as follows:

Java code

public　class　TopicCompare　implements　Comparator{
　　　　//热门的话题与顶的数目和评论的数目有关，权重分配比例为4：6　
　public　int　compare(Object　arg0,　Object　arg1)　{
　if((arg0　instanceof　Topic)　&&　(arg1　instanceof　Topic)){
　　double　y=((Topic)arg0).getAgreeit()*0.4　+　((Topic)arg0).getCmtnum()*0.6;
　　double　h=((Topic)arg1).getAgreeit()*0.4　+　((Topic)arg1).getCmtnum()*0.6;
　　if(y==h)　return　0;
　　else　if(y>h)　return　1;
　　else　return　-1;
　}
　return　0;
　}

}

Cached slice updates generally I like to use spring AOP, simple and practical, general requirements can be met, aspect syntax is more complex, no time to see, hehe

Java code

import　java.lang.reflect.Method;

import　org.springframework.aop.AfterReturningAdvice;

import　com.common.cache.WebCacheUtils;
import　com.cxlh.dao.hibernate.Login;

public　class　LoginAdvice　implements　AfterReturningAdvice　{
　public　void　afterReturning(Object　returnValue,　Method　method,　Object[]　args,
　　Object　target)　throws　Throwable　{
　if(method.getName().equalsIgnoreCase("checkUserLogin")　&&　!returnValue.toString().equalsIgnoreCase("null")){
　　//最近登录用户缓存更新
　　System.out.println("=======update　login　cache");
　　WebCacheUtils.addRegLoginedUser((Login)returnValue);
　}else　if(method.getName().equalsIgnoreCase("addUser")){
　　//更新最新注册用户缓存
　　System.out.println("=======update　add　user　cache");
　　WebCacheUtils.addRegRecentUser((Login)args[0]);
　}
　}
}

In this way, the cached data of the most recently logged in and newly registered users are updated in real time by the custom collation, and it is not known whether the real-time cache processing can meet the requirements for large concurrent web sites and has not been tested for performance! Careful with!!!