No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

The Web Scale Dilemma of WebRTC

Look around you at VoIP implementers. The only ones scaling properly these days are OTT vendors, and WebRTC is still nowhere to be seen.

There are challenges associated with WebRTC. I look at the cracks that need to be filled on a daily basis. I think the biggest hurdle in 2014 is going to be Web scale. How do you build a WebRTC-based service and scale it into the realm of tens of millions of users?

The whole notion of on-premises VoIP or even VoIP in the cloud for businesses is usually limited to the lower millions in terms of scale--nothing serious if you ask me. It all fits into a single database server and a couple of PBX servers.

When you talk consumer VoIP--things like WhatsApp, Viber, LINE or WeChat--these don't exactly use VoIP the way it is "intended" to be used. Without knowing, I am betting there's no standard SIP implementation in their infrastructures. It may well be there--but with many tweaks necessary to get it to scale properly.

The way I see it, there are two conflicting paradigms for scaling services with WebRTC:

1. The VoIP way of doing things
2. The Web way of doing things

Dealing a bit with the migration from traditional databases to Big Data technologies feels rather similar to the challenge with VoIP. While VoIP was meant to scale, it is the Web that scales today from millions into billions--and it does so with a smaller number of IT people who need to tend to the system on a daily basis, making sure all of its moving parts are...moving in the right direction.

VoIP (or more specifically SIP) is limited in the way it can scale. The protocol dictates certain entities on the network and certain operations. As with anything else in tech, there are hacks and workarounds--but I don't think it is enough. Not if we want to reach globally en masse.

The Web has multiple ways and approaches. All are basically hacks on top of HTTP, but somehow, a lot more minds were put into finding solutions for it, and these solutions are varied and working today at a large scale.

WebRTC? On one hand, it is a known quantity--it is just media over SRTP with TURN relays--nothing new. On the other hand, it runs on a browser, with huge potential for accessibility to end users, in a scale unknown before now. There are many vendors working with WebRTC and targeting such scales. You see it in the infrastructure they are putting in place, in their use of virtualization and cloud technologies, in the autonomy their service has in firing up new servers around the globe.

The problem? None of it is proven yet. We are still in the scale of millions of minutes a day/a month.

2014 should be the year when a vendor comes out with an announcement that it reached a billion minutes a month. Not yet the size of Skype, but for a 2-year-old technology--that'd be a nice number.