This week I am attending a standards meeting at ITU here in Geneva. The topic of discussion is Next Generation Network (NGN). Last time I attended an ITU meeting (at that time it was known as CCITT) was some 20 years back while working on ISDN. Many things have changed but some remain the same. Technological progress is very much visible; seemingly pedantic hair splitting is still in vogue. Of course, the human dynamics hasn’t changed at all.
All the meeting rooms are equipped with Ethernet and WiFi connectivity. One can download the copies of contributions through the network. This means one does not have to lug paper copies to take back home. Indeed my colleagues who wanted to hear only a handful of contributions could hear them through an ad-hoc voice chat session. The meetings are not streamed, at least not yet. But not surprisingly, the human political dynamics is still evident. During one of the dead times, through Google I tried to locate the heavyweights from the past. I came across a positive reminiscence from Bob Amy that was penned in 1992. That is the topic for today.
In that article, Bob describes the forces that drove the development of ISDN and also postulates many exciting advances. I do not have such a romantic memory, not because the promise never materialized. I fault the prospective service providers and their vendors for not being true to the objectives and capabilities of ISDN. It is my opinion that this is one of the main reasons why ISDN failed as a business. At least in US, BRI ISDN turned out to be an expensive screen based phone and PRI ISDN turned out to be a CTI application. There is a lesson for us, lest we repeat the mistakes in VoIP as well. With that in mind, let me discuss some of the mistakes made by ISDN developers.
Bob says, “The concept of a personal computer was still a dream and the on premises speeds of current local area networks was not yet developed.” It is true that in early eighties, computers were not that widespread. But the fact that ISDN was going to integrate voice and data suggests that the anticipation was that data communication devices will be widespread. Still, ISDN planners were visualizing data devices will be connected to the network through the phone. (Take a look at an ISDN phone – it has an RS-232 connector.) Instead, we should have considered the scenario where the phone is connected to the network through a computer. So, ISDN architects should have believed in their own slogan: ISDN was going to “interconnect the network switches to intelligent customer terminal equipment”. If they had done so, they would have developed a switch which executes only a handful of basic functions and the terminal realizes features at the ends. This would have simplified the switch development and reduced interoperability problems. And there would have been no need to “indefinitely” keep developing standardized procedures for ISDN supplementary services. For example, Busy line and Call waiting does not make sense in ISDN (because it has an independent signaling channel, unlike the tip and ring line). I am to a large extent disappointed that both H.323 and SIP communities are repeating the same mistakes.
So a tap on the shoulders of VoIP architects and H.323/SIP proponents. Be true to the technology. Don’t artificially insert a service provider when the end users can realize the features by themselves.
Posted by aswath at February 6, 2004 03:40 AM