4 is WebRTC’s Magic Number

16/09/2013

Looking for barrier reduction in video conferencing? Here’s what I noticed.

Magic number 4

I was involved in a video conferencing client project in a past life. In it, we had to develop something from scratch to interwork with known VoIP protocols and deliver HD video. When we planned the project, we ended up with around 10 developers involved (not including testers). It was split to 5 developers on the media part, 4 more on signaling and logic and an additional UI developer.

With time, the team only grew. There was a need for 1-2 voice engineers – people capable of dealing with nasty things like echo and device drivers. Codecs had to be purchased (with royalties attached), and then integrated. The media engine that was developed required constant care.

Signaling wasn’t fun either. These protocols interoperate only in theory. When it comes to practice, they require modifications at all times – standards changes, requirements from customers and the occasional interoperability bug with yet-another-third-party-vendor’s-product.

During the Atlanta WebRTC Conference and Expo a while back, I asked several of the vendors there what’s the size of their WebRTC team – the one that developed their product.

The baseline number?

4.

Not 10. 4.

You need 4 developers to build a WebRTC service. This includes both client and server.

4.

Sure. Once a service gets traction, more developers will be required. More features need to be added. Whatever. But as a baseline magic number, 4 is pretty low.

When the number is shrunk this far down, what does it mean for the industry? There’s a drastic reduction here in the cost of building a service from scratch – both in labor (=developers) as well as capital (=licensed software). This s what brings us to a point where many of the vendors introducing new services is so large, and the use cases are so varied.

Want to take it to the extreme? Nicholas Buchanan single handedly developed OpenVRI. And he is hard of hearing – do you envision anyone hard of hearing being able to develop a serious VoIP product without WebRTC? Think again.

Looking to get these 4 developers to grok WebRTC? Try my online WebRTC training courses.

Responses

blank
song zheng says:
September 16, 2013

Hi, great article!
An alternative to your suggestion of using 4 developers to build a webrtc solution is to use an api. Shameless plug, I built a sample WebRTC demo using TokBox API and I’m very happy with the result. You can check it out at opentokrtc.com, and all code is open sourced and available on github.

Reply

Comment