zeugma wrote:
This is a real question that I am curious about as an interested layman (I am not trying to be snarky or sarcastic).
Good questions all! First, the reason there is money in VoIP is because you want PSTN interconnection. If you don't want PSTN interconnection, you could use Skype. Skype is free (although Skype does make money). The reason that there aren't that many people doing raw SIP on their PCs is that, conversely, there is no money in it, but if they did, you wouldn't want to use it, because it would be full of spam (and some other nastiness like denial of service attacks). While there has been lots of interesting research on preventing spam in decentralized SIP environments, I think its far easier to do if you have a central "VoIP Centrex" which handles everyone's calls and can see suspicious call patterns from a central position and block them. Therefore I don't think you are ever going to see raw SIP on a broad basis. This perspective is my own and is very unpopular with both the IETF and several people I work with. I am, nevertheless, convinced that I am correct about it. is it really that expensive to set up a STUN server to traverse the NAT?
Yes, the purpose of STUN (really ICE) is to do what you're saying and the people who are promoting it believe that people want to do raw SIP and need ICE to do it because NAT devices just aren't going to behave. My perspective is that people want to control what is going on with their networks, and there is nothing wrong with that. The problem isn't NAT devices, the problem is the fact that SIP causes unpredictable UDP back connections on random ports. ICE layers sin upon sin, by taking a protocol that is hard to accommodate and making it even more unruly in hopes of successfully piercing holes in firewalls. You know something is not right when it has become hard to tell the difference between IETF standards and malware. Also, if the move to IPv6 ever happens, will the need for NAT traversal pretty much disappear?
Another belief that is popular with the IETF and which I reject. There are three problems with this perspective. The first problem is the idea that NAT is going to go away. The IETF seems to believe that the world is teeming with people who hate NAT and are praying for the day when someone parachutes into their networks and liberates them from its oppression. IPv6 will not be welcomed as a liberator. The fact is that people like NAT. They NAT because they don't want to have to ask an external third party permission in order to change their internal network architecture. If you NAT, you don't have to bother your ISP with a request for more IP addresses and you don't have to wait for them to decide whether or not they approve of your actions before your proceed. Furthermore, some people NAT because it makes them f... [ Read More (0.2k in body) ] | 
