Chris Kranky

Recent Posts

WebRTC: A new beginning

Chris KoehnckeChris Koehncke

This is Part 1 of a series of articles I’m writing to help explain to the non-techie or semi-techie about WebRTC. WebRTC is a Google initiative to make it easier for web developers to deal with 2-way multimedia content. WebRTC rides on the coat tails of the much larger HMTL5 (which I’ll touch on as well). WebRTC is an abbreviation, by the way, for Web Real Time Communications.

I’ll focus on the impacts of WebRTC to the existing VoIP industry at large. But first, it’s helpful to understand how we got here.

Early Day Telephony
Telephony has a long history and it’s played an important role in the creation of worldwide commerce. Despite the fact that we all are talking less on the phone each month, the reality is voice remains an important component of our communications. The voice industry has a rather large ego and hasn’t wanted to play well with other technologies. It was in the market all by itself for a long time and acceptance of the Internet hasn’t come easy.

If you understand how a voice call is made today and the underlying complexity associated with each call, you would consider it a small miracle that any phone call manages to reach it’s end destination. The only reason this happens with any certainty is that voice networks have been over engineered with a huge amount of complexity and cost. Unfortunately, this concept of how a network should operate is diametrical to how the Internet was conceived, which is in fact, the beginning of the problem.

Enter VoIP
The telephone industry obviously saw the rapid acceptance of IP technologies and decided they didn’t want to be left behind.  The majority of the complexity of a telephone call is signalling and routing. I need to signal I want to make a telephone call, provide information on who I’m trying to reach and the “network” needs to figure out how to do this. Thus there was a need for a signalling system that worked over an IP network.

The industry initially created a new signalling protocol for this known as H.323. But H.323 was extremely complex and cumbersome and required a great deal of knowledge to build, operate and maintain a voice network. It was also mostly designed for point-to-point calling and really didn’t take into account that the Internet is based upon a mesh network design. Nor did H.323 networks scale very well to handle the tremendous amount of calling that happens.  But a phone call could work, but it was ugly to watch and really was only IP in name alone.

At about this time, a group of non-telecom people were looking at this and thought there surely must be a better way. They saw the model being Internet email. Let me backtrack and over simplify. Today, when I send you an email, my email server “calls” your email server asks if “Joe lives here” and if the answer is YES, then says “here’s an email for Joe.” Internet email works well, you and I use it nearly daily and with only scant problems.

So this group figured they would take the principles of Internet email (which looked a helluva of lot like making a phone call) and write specifications for a new VoIP protocol that would be easy to understand and implement. This protocol was SIP or Session Initiation Protocol. The concept this group had was that MY telephone server could call YOUR telephone server and we’d have a phone call, just like email. SIP talks using plain text (for the most part) and if you look at SIP messages, it looks awfully like a email transaction using SMTP. On the face of it, this seem like a great solution.

The telephone industry thought this was a grand idea until they realized that if your telephone server calls my telephone server, exactly what was the telephone company doing to “add” any value? Who makes any money when you send an email? Well .. nobody. SIP quickly started to sound like the end of their business as they knew it. So the telephone industry decided to the only real way to kill this was to to “help” the SIP specification grow. Rather grow in complexity.

Over the course of the last years, the telephone industry has taken SIP, which was rather simple to start with and basically layered in 50 years of telephone requirements. Lacking an international standard organization to mandate what SIP should be, SIP has mutated in multiple flavors that may not be understood by another supposedly SIP compliant device. This, of course, has spawned a whole new industry who make sophisticated protocol converters which convert one type of SIP message to another. All of this in the name of making telephone calls simpler. Obviously, no good deed goes unpunished.

… continued