W e b w o r d

Vol. 1, No. 2: MARCH-APRIL 1997



Charles Petrie

Stanford Center for Design Research


The Internet was designed to survive a nuclear war. But can it survive capitalism?


The answer depends on what you mean by "the Internet": function or technology?

If you identify the Internet by its historical functions and how well it performs them, then it has already been at least gravely wounded. I'm talking here about the traditional academic services: e-mail, telnet, and FTP.

E-mail, mind you, is better than ever. Last year, for the first time, I could reach almost all of my friends who were not computer scientists by e-mail. But new e-mail is different from the old Internet e-mail. MIME and Web-compliant standards are going to force me off my creaky TOPS-20 MM command-line mailer soon (for those who think computing started with Unix, this was a DEC operating system). And some days I get new messages faster than I can type a response.

All this is better than the bad old days when there was a relatively small community of users exchanging short ASCII messages. And since e-mail is asynchronous, this change has not made network performance an issue.

Unfortunately, the case is not so rosy for telnet and FTP. Since the mid-eighties, I've traveled around the world carrying only a virtual computer, secure in the knowledge that an Internet terminal awaited me at the next stop. I'd just telnet back to my host machine, and pretend I hadn't left. Because I've been in a Unix world most of that time, I could X-host display and have my home monitor's display on the local host. This has sadly changed.

Now, even if I just want a command-line connection via telnet, distance interaction is so slow as to be almost unusable. The original design latency for single-byte packets echoing characters has long been exceeded. Similarly, FTP sessions now take much longer than in the past decade. Telnet, along with FTP, drove the first, original Internet, and those functions are near death.


The wealth and ubiquity of the Web now simply defy description. Never mind what's new on the Web today - everything is there. But along with the new volume of MIME-laden e-mail, this mountainous landscape of information is a new burden for the old underlying Internet technology. Is it buckling?

Robert Metcalfe predicted it would collapse catastrophically in 1996. George Gilder said it wouldn't. Gilder was right and he had the right major reason: People keep adding to the Internet resources, especially now that it's commercial. It's simply a good thing over six million dollars were poured into the Internet every day on the average in 1996. No one now doubts there is a market here, and no one can then doubt that enough investment will be poured in to make the Internet work.

But will the old IP technology support the load even with commercial shoring? Can it survive Web-TV and networked appliances?

Someone told me recently that the Internet slows down after 5pm each day. Gee, is that PST or GMT? I told him the problem is almost certainly the capacity of his local ISP. And if they didn't fix it, he should switch. That's the beauty of moving the Internet to the private marketplace. Each ISP competes by providing better service for the dollar, and this means adding resources to the Internet. If nothing else, the AOL debacle last August - where 6.2 million subscribers lost access to the Internet for 19 hours - points up that performance problems for most people are caused less by the Net global infrastructure than by their local provider.

Re-enforcing the investment by industrial users and the ISPs is the growth of the telecom's investment in new high-volume lines and the development of new NAPs, better router switching technology, and faster protocols (for example, HTTP 1.1). The Internet is not going down; it's stretching its backbones, growing up.


The Internet may certainly look different as an adult than it does as this gangly teenager. For example, independent ISPs may be a temporary phenomenon now that telecoms are changing their business model from selling data pipelines to offering Internet access. Broadcasting over "push" channels may catch on. (It's certainly appealing to those who want to cash in on a new TV-like industry.) Intranets and extranets and Internet 2 are leading the way to private nets. ATM is becoming increasingly used.

Flat rates are under attack by the US telecoms at the same time the world markets have opened up for competition. The domain namespace is exploding. The biggest new Internet markets may be in Europe and Japan rather than the US. Some form of e-cash is around the corner, and Metcalfe is pushing e-mail postage.

These are the kind of changes that make the Internet newly fascinating. Who knows how it will turn out?

We're facing an emerging Internet wrapped into our toasters and morphed into a jungle of connections and economic policies that will make your current long-distance phone choices look simple. It's not your dad's Internet anymore. The amazing news is that the basic technology designed to survive WWIII seems to be surviving the WWW.

The Internet is dead. Long live the Internet.

Back to IC Online

Copyright (c) 1997 Institute of Electrical and Electronics Engineers, Inc., All rights reserved.