Suppose you have two clusters of app servers that have access to the same data in the database. You want to ensure that updates to the data are controlled, i.e. they won't result in data integrity issues.
I think that only one cluster should have the privilege of retrieving and storing the data. The other cluster can gain access to the data by looking up (using JNDI) the corresponding entity beans remotely.
The downside of this is that the performance will likely suffer. Can you guys give me your opinions on this approach? What if one of the clusters is not in Java - presumably a distributed transaction processing manager would be needed.