The following text is copyright 1997 by Network World, permission is hearby given for reproduction, as long as attribution is given and this notice is included.

IP under all?

In previous columns I have lamented the fact that a few years ago some of the ATM proponents loudly proclaimed that ATM would become the common network sinew for the globe. It would replace all LAN and WAN network links and LAN and WAN protocols - IP was to be replaced by a seamless global mesh of ATM. There were not many people who felt that way but their influence was out of proportion to their number. All too much of the trade press fell for their 'ATM under all' evangelizing and grossly distorted their readers ability to reasonably evaluate what was the best technology for their own networks. Now it is true that the people most convinced were the people furthest removed from those actually running data networks. The people in the trenches just kept buying Ethernets and converting to TCP/IP. The same hype is now being applied to IP.

At the Next Generation Networks conference a few weeks ago John McQuillan, the conference chair, noted in the closing wrap-up session that "IP has won." He was referring to the general current assumption that IP, and the Internet, are the common bearer service (as the Network Research Council's book Realizing the Information Future put it.) Some members of the audience thought he was referring to some IP vs. ATM contest but he explained that this was not the case. IP can run over ATM just fine. This was more of a winning over of the general networking business, away from SNA, IPX and other networking protocols and a winning of the mind share of the networking professionals.

But many of the sessions in that conference, in a number of other conferences, and in the trade press, show a disturbing trend. Too many people are now starting to see IP in the same way that those few people saw ATM, as the universal underlayment. Voice over IP, video over IP, commerce over the Internet -- suddenly IP is the answer, what was your question? Even the big consulting companies have begun to make a glacial turn away from a pure ATM future to at least a future where ATM shares the stage with IP.

I am worried that too much is being expected of IP and the Internet, just like there were unreasonable expectations being placed on ATM. And I'm worried that when IP and the Internet are not able to meet some of these expectations there will be a backlash which will devalue these technologies for applications for which they are well suited.

I am not at all convinced that a common IP Internet infrastructure is the best or most cost effective way to upgrade the global telephone network, or to bring network TV and video on demand to the home.

It does not follow that just because IP does many things very well it will do all things very well. There seems to be a hunger for single answers to collections of problems; reality is not always ready to feed that hunger, a little analysis and common sense is not out of place here.

disclaimer: Harvard understands glacial movements well but the rest are my own observations.