I am working on a project: A server will load messages (XML Strings) into a IBM MQ. Another server will pull these messages and execute the necessary business logic.
I will be responsible for the latter part (pull msgs and apply business logic). This server has bea WebLogic 8.1 and Oracle database.
Here is the issue: I am new to bea Weblogic and EJBs and so I am having difficulty putting all of it together in terms of design or architecture. From all that I have read, I realize that I will be using MDB (message driven beans) to pull the messages, then use SessionBeans (stateless likely) to work on the message (use XMLBeans to extract the message content, and then apply the business logic which eventually create a SQL statement that needs to be executed on the Oracle database).
I know I am still unsure of a lot of things and need to pull this one in matter of 2 weeks. What do you guys recommend and any key points that you can offer that can help me accelerate into understanding how to develop this project?
Some of the questions I have are:
1) How do I configure WebLogic to connect to the MQ to pull the messages and remove it only when I am doing processing it successfully?
2) Should I even bother with SessionBean or use plain java (POJO - plain old java objects) to implement the rest (business logic, xml parsing, sql execution)?
3) I know if I use SessionBeans I should use pooling to improve performance as we expect many messages. How do i configure Weblogic for the pooling or how do I tell MDB how to call a SessionBean from the pool?
4) If I should use XMLBeans to parse the XML strings, do I use Weblogic to pool?
5) Should I use JDBC connection pooling or should I use Weblogic to do the connections? I know I will be executing mostly select, insert and deletes on prolly only one table.
6) Given that the messages do inserts or deletes only on one table, is it even worth having SessionBeans Pool, XMLBeans pool, connection pool?