The View from Mexico City

By: Scott Bradner

I've just returned from Mexico City and the "Simposium de Telecomunicacionses." I was a bit disappointed but not all that surprised by a few of the things I heard there. Some people there have the same view of the future of national and international data communications as some people here do. That view is of discrete sand piles.

One of the presentations spoke glowingly of the data network services model exemplified by the French Minitel network. This is low speed (2400 baud) dial-up network of 5 million users and over 12,000 on-line services that in 1990 logged over 85 million hours of connect time. Billboards in the Paris subways display Minitel numbers for more information and you can order concert tickets and much more through its services.

This is rather impressive in comparison to the Internet and might actually gives us a hint as to the potential user population of good, easy-to-use data services. One out of 11 French citizens has a Minitel terminal and the usage is more than 1.5 hours per citizen per year. Projecting this rate of acceptance and usage onto the U.S. population would mean over 20 million users and 375 million hours of connect time.

The problem with Minitel is that it is basically a closed network. Sort of like someone buying a TV set that only receives HBO. Sure it's useful but more open system that is a fully integrated part of a global Internet would be much more useful.

I saw a tendency for this same sort of thinking during the Information Superhighway Summit last month in San Jose. There was an undercurrent at the meeting which kept returning to the question of what was going to become the information superhighway. The options were the cable TV network, the telephone network and the Internet. I gave a presentation that claimed that, in fact, the proof of concept of the Information Infrastructure was already here and is called the Internet.

One of the powers of the Internet is that it is not selective in the transports it requires. It already runs over telephone links and will run over just about anything else including the cable TV nets. Someone even claimed on one of the mailing lists that TCP/IP could run over carpet static.

I saw one problem during the conference, namely that too many of the people involved in these efforts seem to think that they should build their own highways - highways that may have limited interconnections with each other and the existing Internet.

This type of feeling is not limited to people trying to plan the National Information Infrastructure (NII) but also seems to have contaminated many of those companies trying to enter this data networking business. All too many of them are setting up their own private networks that offer some set of functionality rather than figuring out how to offer the same functionality over the Internet and thus make the infrastructure richer and more attractive.

I expect that many of these ventures will be quite successful, as many have been in the past. Compuserve, from what I hear, is doing fine. I don't doubt that more companies will also do well. However I do not think that this sort of fragmentation is a good long term strategy for any of the players. The building of this type of private sandbox is probably inevitable but none the less lamentable.

PS - My admiration of the inventiveness of the engineers who work for the airlines was reinforced during my trip to Mexico City. I took vertebrate biology when I was in college and even though it was quite a while ago I do not remember any part of a chicken that quite resembles what we were served as a food substitute. It is truly amazing what can be accomplished by inventive minds.

Disclaimer: Above opinions != Harvard's.