3 ms·
> SIP? SIP is terrible, it's an attempt at replicating the telco model of separate control plane vs talk channel of a traditional telephone company. Good point
by popee 11y ago
> SIP? SIP is terrible, it's an attempt at replicating the telco model of separate control plane vs talk channel of a traditional telephone company.
Good point and I agree, but it solves problem of direct communication between users (NAT traversal). It would be great if SIP had plain and simple GET method (without need for session/dialog).
> This doesn't follow either. If I host a webpage on my personal computer, I can host the js/css in exactly the same way.
How can I get your js/css from outside, if you are behind NAT? STUN? NAT traversal is already part of standard SIP infrastructure.
Sorry on stubbornness, just trying to brainstorm a little.
- pjc50 11y agoSTUN is indeed used by SIP to traverse NAT, as part of a whole pile of techniques called ICE. However, this ultimately requires that you have a central server on which you can "advertise" your SIP presence and exchange SDP data. Yes, the session payload transport (RTP) is peer to peer, but the signalling still requires a central point. > How can I get your js/css from outside, if you are behind NAT? The same way you get the HTML, whatever approach is used. I don't think it's impossible to build a system involving distributing "advertisements" saying how to reach systems on a distributed hash table, then using either ports opened through UPNP on home routers or a "supernode" collection of STUN servers to communicate to home systems. There are three big downsides to this which are probably why we've not seen one yet: - no way to monetise it if it's truly distributed and open source => reduced incentive to build and maintain - no central party to fight spam and abuse => system will drown in spam and abuse - necessarily broadcasts an association between home IP address and person => vulnerable to privacy problems and DOS attacks