When I think of distributed jobs, I think of JMS and MDB. Rather than using SOAP internally, simply publish a JMS message to the cluster. Since JMS messages are processed asynchronously, unlike SOAP, your SOAP client doesn't need to block for completion. JMS has a notion of transactions and supports things such as automatic retries. So, if the system goes down, your jobs aren't lost. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4087020#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...