Evgeniy Khramtsov wrote: > Peter Saint-Andre wrote: > >> At the recent XMPP devcon, I talked a bit with Thiago Camargo about NAT >> traversal and media relays. There are really two separate issues here: >> >> 1. Finding and using STUN servers for NAT traversal. This is discussed >> in XEP-0215. >> >> > New STUNbis doesn't define "STUN servers". Currently STUNbis is a low > level protocol which you can use in other protocols such as ICE and TURN.
Correct. I wonder how long before rfc3489bis is widely implemented. :) >> However, as Rémi Denis-Courmont has pointed out [2] on the >> IETF's BEHAVE list, it is not necessary for the relay to be a TURN >> server. It's great if the relay is a TURN server, but it could be >> something else -- and the important point is that for the purposes of >> media relaying it doesn't really matter to the Jingle client whether the >> relay does TURN or something simpler. >> >> > I can't agree :) Why do we need a zoo of relay servers? What is it for? Who said anything about a zoo? Thiago mentioned that there's no real need for the relay server to be a TURN server, and at least some people on the BEHAVE list (and elsewhere) seem to agree. But I'll let Thiago explain the reasoning more fully. BTW, can you recommend a good open-source a codebase that can be used to run a STUN server and/or TURN relay? /psa
smime.p7s
Description: S/MIME Cryptographic Signature
