Hi everyone, Just a quick note on discovery from the RTP MIDI experience: http://www.cs.berkeley.edu/~lazzaro/rtpmidi/index.html The assumption we're making is that existing IETF session management tools -- the Real Time Streaming Protocol (RTSP) for applications that feel like content distribution, and the Session Initiation Protocol (SIP) for applications that feel like telephony -- will be the way RTP MIDI (together with audio and video and text, etc) sessions will get set up, and that the mechanisms these tools use for discovery will be used for RTP MIDI. On SIPPING (the place where new SIP frameworks are born and raised): http://www.ietf.org/html.charters/sipping-charter.html there has been a lot of activity recently about "p2p SIP architectures", which in a LAN context is the sort of discovery people are talking about here (link-local multicast, the backbone of mDNS, is the LAN tool at the bottom). But there's also an interest in making these p2p SIP architectures work outside a LAN context, which is where the "p2p" comes from -- they look for inspiration from the file-sharing world for discovery (Distributed Hash Tables, etc) ... Also, note that basic SIP already has some link-local multicast support -- a new phone coming on line can sent a REGISTER to a well-known local multicast address and port, where any SIP proxies online will be listening. This doesn't solve the bigger question of how you make an architecture where even phone is also a SIP proxy, and act to serve each other ... which is what the SIPPING work is about. --- John Lazzaro http://www.cs.berkeley.edu/~lazzaro lazzaro [at] cs [dot] berkeley [dot] edu ---