They used to say SIP is better than H.323. Simpler. Now it is WebRTC’s turn.
I was an H.323 guy. I knew it inside and out: the ASN.1 definitions, coding rules, signaling nuances, call control, gatekeepers – the works. And then came SIP. This nagging (and stupid) protocol, with the only thing going for it was text. You had these text messages instead of binary, structured ones. And people fell in love with it.
Simple they said. Easy to debug they shouted. But now, look how bloated and impossibly complex it is. Whenever you try to connect 2 SIP systems you need someone to come and configure/debug/fix whatever it is that doesn’t work that well. Along the way, it lost its way.
At the WebRTC Conference in Paris a few months back, it was said that SIP doesn’t work for operators – every feature you want to add – every new use case – required one of two things: you either had to add it to the standards as yet another contribution, or you had to find the right user agent and servers that implemented the exact features you needed. Not an easy task.
WebRTC changes all that. In a way, it is lean, mean and with very few implementations: with 4 you can get complete market coverage (Firefox, Chrome, IE and Safari). Signaling is taken care by whatever you want in your service – heck – you can even use SIP if you want to.
WebRTC picked up all of what we wanted as an industry to achieve from H.323, and then from SIP:
- Interoperability – you get that from relying on a (very) shortlist of implementations to lean on
- Accessibility – Java Script APIs available to web developers beats a written spec for VoIP engineers – it increases the potential developer base by a factor of 100 if not more
- Maintainability – that security issue? Take that to the larger scope of bug fixing and upgrading and you get an easy system to maintain
- Usability – WebRTC shifts RTC from the telecom industry into the realm of the web, making it a feature that is available to a much larger set of use cases
I still believe H.323 lost to SIP due tom marketing and hype – no technology merits.
With WebRTC all the reasons were technological ones. SIP never stood a chance.
And yes, I used the past tense because I think this is already a done deal.