Well, I had no idea you were going to do this. It was my understanding we were going to keep the current way of doing things in remoting until 1.2.1 when I will make all this moot by replacing this with a NIO based system which will make most of this obselete. Can you explain in more detail how this works? If "The server actually uses that information (the number of in-flight messages) only in a special use case", then what's the point of sending it in other cases? Are these true asynchronous invocations? Or are they remoting style asynch where a normal RPC is done and the result discarded? Have you done any tests to support your statement that the throughput has actually improved? I would really like to be kept in the loop in the future if such major technical changes are going to be made and committed into trunk. Is that reasonable? View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3988259#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...