The client requests for a page from my application. A servlet will be invoked at this point of time. The servlet creates an instance of a dataholder class,look up for the Application server through JNDI Naming ( The server IP is
taken from parameter file). Then it creates a session bean using the Home interface and then invoke a method in the session bean through the remote interface and dataholder class will be passed on to the App server. Now at
the app server, the session bean, does the processing ( it may require to use some entity/session beans for this). Put the values on to the dataholder class and pass the class back to the web server. Now webserver displays a JSP page to the client by filling the details from the Data
holder class.I wanted to know in this scenario what will be impact of weblogic clustering.

I guess it depends if you want your app to be effected by clustering.
I presume you are using some sort of a proxy server (to receive requests from clients)..assuming you are doing that.
1)You could have the cluster machines load balanced and access the app servers based on some known algorithms (round robin,weight based and so on...)

2)If you are using Statefull session bean , you need not look up the name in JNDI since the instance is replicated by the cluster..

-N R
PS:What do you want from the cluster apart from load balancing and failover ?