okay im writing my second networked game, and my first one turned out pretty nasty. I would send a message followed by any variables to the equal class over the network. This required a class being registered when created and getting a unique id from the server. Then this class could then shuffle messages to its network counterparts. The problem was that i ended up with a lot of messages and this was not very flexible. Im thinking of trying some sort of RMI (remote method invocation) which doesnt return values so it doesnt need to block. I'm also trying to work out a good way to structure my network architecture to make it easy to add it new objects in to my game that work via network. my next game im using a messaging system to communicate between objects so i have: class MessageReceiver { bool HandleMessage(Message* msg); } im thinking i could simply make the Message class and its children replicatable, but then i would need messages just do any sync'ing which seems excessive. any one have any suggestions on this?

hmm its not really clear to me what kind of help you want.

Try taking a look at raknet, on the main page there is a doxygen manual which makes it easier to view the overall structure of it.