> > These are really good suggestions - let's keep going. > is the protocol used by the VM sockets documented and intented to provide an entry-point API, so that we can drive a performance by local TCP/IP ? if not what would be good to have IMHO Stef