Some friends and I were recently reminiscing about the old days at the IETF, before the commercial Internet brought the commercial vendors into the mix. In the early days of the Internet – when the goal was to ride the PSTN and not run it – the standards had the principle of taking and masking the issues of transport and delivering an end-to-end controlled experience, letting the modems whistle to get on board, and leaving the routing to be done independently of carrier networks. And router rules were based on efficiency. While that is a bit of romanticism, there is some truth in that vision.
Today the Internet, or at least IP, is replacing the PSTN to the point where the networks are close to being one and the same. We even have terms like Internet Offload, which is a fancy way to say the role of the carriers is to get out of the way and let the routers do their work.
My friends were making the point that the Internet is meant to be global, while the PSTN was meant to be local (national). Country codes are not IP addresses, and anything that attaches them to each other is basically a routing tool that may override efficiency. The use of back-to-back user agents (aka, session controllers) is designed to help the carriers keep their network efficient. SIP and FAX suffered extra RFCs, and they liaison work with other standard bodies that make compromises between blending the local / national needs and the end-to-end principle.
The lack of a signal protocol has served to make it easier for WebRTC to avoid pitfalls. In theory, it works well at bringing back the end-to-end principle, particularly when WebRTC is either in peer-to-peer mode, or associated with specific websites (that may be distributed).
One of the problems with standards bodies is that like U.S. Congressmen they have a tendency to deeply root themselves and never leave. Google, of course, has done a great job navigating the standards, though the company has a history of doing things and then either abandoning them or reaching a point of satisfaction. In the case of the W3C, it has been a good participant up until now, and it is my hope it continues to lead the adoption of the HTML5 standard. To be candid, however, when Google bought GIPS and On2 Technologies I assumed it was done with codec standards, and I think that in the battle over the codecs it was not expecting to suffer. As such, I have my doubts that Google will continue to pay much attention once VP8 is done. Standards bodies at rest, though, get restless.
If I were interested in making WebRTC more like the PSTN, I would be participating in the standards work for the long term. Thus I believe that the Google goal should not be to win in standards, but to have as many implementations in the market as possible. In effect, to make a legacy of WebRTC that is simple and hard to ignore.
Edited by
Rachel Ramsey
