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.
Was Bob a year too early?
For the last year or so Bob Metcalfe and I have been trading barbs over his rather well reported predictions that the Internet was going to collapse during 1996. It is now 1997 and Bob, in his last Info World column of 1996 sort of, but not quite, ate his words.
Clearly at one level the predictions, at least as they have been understood, were quite silly. It is very hard indeed to imagine the Internet itself collapsing. The Internet is not one thing, it is a collection of things (ISPs). One or more of the components might (and undoubtedly will) have serious problems but to say that this means the whole 'Net is down is quite a stretch. But at the level of the perception of the users Bob's predictions have some validity. It matters little to a user trying to download Bob's latest pearls of "wisdom" where a breakage or slowdown is in the path, the user still sees poor performance. But overall, any prediction of the collapse of The Internet seems more that a bit simplistic.
Come to think of it though, I wonder if Bob was just a year or so too early in the gist of his predictions. This Internet thing continues to confound the traditional (read telephone company and mainframe thinking) pundits. These people think that the Internet can not possibly continue to grow with out some "killer app", the establishment of settlements between ISPs, major commercial content providers, usage based pricing, or some other favorite fixation. One does sort of wonder why observers, such as the trade journals, continue to pay any attention to these prognosticators with their lengthily records, unblemished by accuracy.
The Internet continues to grow in users, traffic, sites, and public mind share at an astonishing rate. And this is before new initiatives such as the administration push to get the Internet into all schools and libraries have really gotten started. When this sort of thing happens, there will be still more pressure from the kids for mom & pop to subscribe to an Internet service so they can have the same access at home as they do in school (while having to stand in line to use it). This coupled with new trends like webTV will tend to accelerate the growth curves.
At this point it is a bit hard to see how the ISPs can continue to keep ahead of the demands. Continued unbridled growth is a hard argument to make economically and the availability of the very high performance routers and switches that will be required is currently questionable. Coupled with the growing shortage of fiber in the ground (and under the water) this means that Bob's crunch may actually be on its way for real.
This is not to say that The Internet will sink into the mud next week but trying to reach an undersized server over a too small link thorough an underfunded ISP will not produce the best of results. I'm sure that this will bring out the pundits in droves to say that all will be fine when the Bell-ets decide to help and when traditional telephone-like traffic-based pricing is instituted to install discipline in the users. In some universe that might be true, but I don't expect the telcos to ride to the rescue in this one. The ISPs will have to do it on their own and we may have some fun times while they figure out how to do so.
disclsimer: With its own fiber in the ground Harvard is above these (and some say many) real world problems but the above are my own trepidations.