This is about very interesting scenario I came across and I think many of you 
might have also faced the smiliar situation.Scenario : I have one distributed topic in my Integration scenario, and I am subscribing to that Distributed topic (Clustered Environment ofcourse) with a proxy service (Oracle service Bus 10g) and storing the same message again into another distributed Queue. So whenever message is getting published to the Distributed Topic, Message is stored in both the clusters and as a result I was ending up with two same messages in single Distributed Topic, each message in one cluster. Because of this, my proxy service was getting called twice for the same message and same two messages were getting stored in Distributed Queue.
Well so issue here is duplicate messages in distributed topic, but I guess 
writing a word "issue" is not correct, because Distributed Topics are made to behave like this. This is what we call expected behaviour. But my requirement was how to remove this duplication from Distributed Topic.Here is the trick, Create two seperate Topic (Not distributed Topic) and targeting both the clusters one by one. Now we have something called "Local JNDI Name" in Topic. Set the same value for this field for both the topic and refer to this common name while refering to Topic. This will give very much similar functionaliy like Distributed Topic and yeah ofcourse without message duplication.
Limitation : All limitations applied to local JNDI name will be implicit to this solution.
Sun Cluster Data Service for Weblogic Server Guide for Solaris OS
Oracle Application Server 10g: J2EE Deployment and Administration
