Limitations of the Internet
Excerpted from my keynote speeches
I’d like to discuss with you the emergence of a universally-connected world, and talk with you about the opportunities that this new world presents, and explore with you the possible limitations.
What are the limits? And will they slow us down from this vision of a universally connected world? And for the empowerment that e-business will bring to all of us. Well, I like to think of the glass being half-full, not half-empty. But there certainly are some possible limitations. I think there are five that are most meaningful.
First, security. Will people learn to trust security? I believe they will. I believe that very soon people will say “Gee, security is not actually a problem. In fact, it’s an opportunity.” Because of the great power of encryption, no longer will we have to call an 800- number and give our credit card number to a complete stranger. No longer will we have to take business critical secrets, put them in a brown envelope, and stamp it “For Your Eyes Only” and then hand it to somebody in the mailroom. No longer will we have to worry about authentication. No longer will we have to worry about the integrity of messages. No longer will we have to worry about repudiation: “I didn’t order that. I didn’t send that e-mail.”
That’s what encryption gives us: Authentication, scrambling of information for confidentiality, integrity and non-repudiation. Incredibly powerful things. Today, nobody would buy a word processor without a spell checker. We know what it is, we know what it does, we trust it. Do we know how it works? No. And that’s the way encryption will be. Very soon, you won’t buy any software that doesn’t have Public Key Cryptography built right in. All data will become encrypted. And this power of encryption will enable e-business, it will enable us to reach out and do business in a community of billions with confidence on a global basis. So security is not a limitation. It’s an enabler.
What about bandwidth? Well, bandwidth is a long discussion and I’m going to try to give you a view of it that’s very net — no pun intended. You break it into two parts: the backbone and the last mile. The backbone was built by IBM and MCI in 1988. In 1991, it was upgraded to 45 million bits per second. On April 1st, 1995, it was more or less disbanded and replaced with a structure of network access points. It served us quite well but it will not move us to the future. About 100 universities have now agreed on a new concept called [Internet II]. [Internet II] will employ a technology of gigaPOPs, a SONET-based synchronous optical networking technology that will facilitate up to ten billion bits per second. It’s under construction right now, and over the next couple of years these gigaPOPs — probably 50 or so of them in the U.S., maybe a dozen or so in Europe and Asia — will all work on consistent standards and will provide of incredible amounts of bandwidth. And ultimately subsume the exiting the DBMS and the existing network as we know it and become the next-generation Internet. So the backbone is going to have the oomph that we need. And the standards will enable it to work reliably and predictably.
The last mile: Many say, “Well, that’s the problem. The last mile.” But when you look at what’s going on here and really get down into the details, you can get quite optimistic about the last mile for two simple reasons: technology and competition. A Digital Subscriber Line is an awesome technology. It’s just a matter of building the modems. Do you think someone will figure out how to make those modems work? Of course. And Synchronous Digital Subscriber Line will be possible, six or eight million bits per second will be possible, into 90% of the homes of America. But meanwhile satellite technology is bringing increased bandwidth. Cable modems are beginning to be built with 25- and 30- and 35-million bit per second capabilities. Wireless is coming on very strong. So these various technologies will leapfrog each other and provide tremendous amounts of bandwidth. I believe the bandwidth will be there.
Scalability is an issue that we should all think about. Scalability has to do with stepping up to the tremendous demands that will be placed on servers. The Patent Database, a free public service put up two months ago by IBM, is an example of what’s possible. Here’s a database of millions of patents — all the U.S. patents in fact — complete bibliographies of those patents, including the images. It’s a really fun place to visit, by the way, to see what patents might have been filed by one of your college buddies or to see who patented a bagel slicer or anything else you can think of. What’s the weird patent of the week? Twenty-eight hundred CD-ROM’s full of data. Terabits of information, being delivered by multiple nodes of a supercomputer, as a service on the Web. Showing businesses of the world that you don’t have to have just a little Web server over in the corner, you can have major databases of customer inventory files, engineering drawings, enormous databases of all kinds. [They] can now be dealt with in a scaleable environment on the Web.
Now, two more possible limitations. Regulation is something we should all really think about. Many key government leaders around the world are looking at this huge growth of the Internet and saying “This looks like something we ought to tax. This looks like something we should regulate.” And this is not a good idea. So the Global Internet Project was formed and announced in December. Jim Clark from Netscape, John Gerdelman from MCI, and myself and a number of other executives in the industry have collaborated to write papers and put together a story for key leaders of the world about the good of the Internet, the power of the Internet to deal with disease management, to deal with disasters, to deal with improving our lives, to deal with education, to deal with many powerful capabilities that will improve productivity and improve the quality of life around the world.
And lastly, proprietary thrusts. Some companies may possibly get an idea about cornering the Internet. Bring it to the desktop, or perhaps corner it over in some portion of the network. This is a really bad idea. The Internet was built on a model of cooperation. It has thrived on a model of cooperation. And its future depends on that model of cooperation. Any computer being able to reach any other computer. Can you imagine some new great resource on the Web — like the Patent Database — and up comes the first page and it says “Caution: This will only work if you have Release 9.3.4 of the XYZ Operating System.” That’s a really terrible idea.
The world is one Internet, but no one company should nor can dominate this. It must be open industry standards, cross-platform and utilizing a language that anybody can write and anybody can read. And it can run on any platform. This particular chart here, you may see our name on it but this is not an IBM chart. This is a chart about the Java World Tour. And on it you see IBM, Netscape, Sun and Novell. Arch-competitors in the marketplace collaborating, arm in arm, to bring the world of components, Java components, to the world. So that components written by any of us for our companies can be treated as though they were written by any of the other three. This is a powerful idea. To allow the world to re-engineer itself using modular, granular components built out of 100% Pure Java. It’s a really good idea. There are no losers with this concept. Now initially it’s these four companies. I hope it will become all companies in our industry creating componentized software that can run on any platform.