The Internet is Missing
The simple architecture that gave us an environment so hospitable to innovation and growth is missing. Here are some steps to bring it back. [Bob Frankston, 2002-03-13]
The Internet is Missing: We Must Find It!
The Internet is Missing
Very simply, the end-to-end architecture of the Internet has given all of us the ability to innovate and create new services. Everyone created new services at the edges of the network as peer participants. The network itself provided unfettered transport.
But We Have Lost The Simplicity:
The lack of addresses in IPV4 means that most systems can only be secondary participants without a public presence.
The .COM mania has hijacked the DNS and we no longer have the ability to maintain a stable presence and the net is guaranteed to unravel as registrations expire. And, worse, the gets scrambled as they names get reassigned.
The lack of encryption has encouraged meddling by those who second-guess the content, whether in a misguided attempt to help or a misguided attempt to have an Internet free of disruptive innovation.
Finding it Again:
Connectivity. The current telecommunications industry is a legacy of television and telephony which are now just minor applications. We must create a new industry focused on connectivity and without the conflict of interest inherent in also being a provider of content and services. See "Connectivity: What it is and why it is so important".
The DNS mechanism is viable but we must separate the social policies inherent in ".com". The DNS must be a simple technology without imposing its own semantics. The ".com" names belong in the marketplace along with trademarks. See "The Tragedy of the .Coms".
We need encrypted IPV6. By definition, if we cannot give everyone as many IP addresses as they can use, then we are out of addresses. We are ten years past the point of crisis and must deploy IPV6 at the edges now and not wait for any changes to the Internet infrastructure itself. The current Internet can be used as a transport. But, just as important, we need encryption since the Internet is part of the real world. Just as we lock our doors, maintaining the integrity of our information is the responsibility of each of us. Encryption also assures that the Internet provides unbiased connectivity rather than favoring familiar services--the ones are already understood rather than the new ones that we can't anticipate. See "The Importance of Encrypted IPV6".
The result will be a revitalized Internet that can again become a fount of innovation and, very important these days, economic value. We should also appreciate the importance of a resilient infrastructure for our security.
To Understand This, We need To Step back...
The Internet doesn't make Sense!
There is a tendency to confuse artifacts with concepts. More important than the current Internet is the concept behind it – extreme simplicity. Traditionally telecommunications is defined in terms of services such as telephony and television. The Internet itself just carries packets of bits and doesn't even guarantee that they will be delivered. The Web and, for that matter, telephony and television, are applications that are created outside of the transport network itself.
To be technical, there is something called the "Internet Protocol" that defines the format of these packets and some housekeeping rules. There is no "Internet"; that's just a term we use for the community that cooperates using this protocol. When all the members of the community are able to exchange packets across a common set of connected interconnected networks something very powerful emerges.
Unlike television, each connected system is a contributor. Unlike the telephone system, each system is potentially connected to all other systems simultaneously and in any combination, not just in pairs or small groups.
Since the Internet does not try to define services (other than transport), each participant is able to take advantage of this connectivity without being limited to ways anticipated by a service provider. Because the Internet doesn't treat each use as special, the connectivity itself becomes a commodity whose price declines very rapidly and thus makes imagination, not cost, the limiting factor. This is counter-intuitive to those whose business revolves around maximizing the value of their capital current assets.
To put it another way, traditional telephony is about maximizing the value of a scarce resource while the Internet is about creating abundance by trying to stay a little ahead of demand. Even though the cost per unit decreases rapidly, the aggregate value still increases!
Too much theory, what about the Web and Email?
After all, who care about the boring technical stuff and economic models when what is really important is the Web. Isn't that what's really important?
Well, yeah. The Web is important but it's not what is important about the Internet.
What is important is that a researcher in an office in Geneva wrote a simple program to play with the idea of sharing documents with other physicists. A while later another group of programmers at a university in the middle of the United States built on that and created a program, which they called a browser, that could be run on any personal computer. More important, these personal computers also acted as servers since the idea was to share not just to publish.
What is important is that the Internet provided the connectivity that made such sharing easy and the community effect meant that the participants could not only view web pages but create their own. Not only could they create their own web pages but they were able to rapidly experiment and redefine the Web. Yet the basic protocols for the web are very simple. To get bold, you should mark it with <B>important</B> which simply means start bold and end bold.
And no one in charge! In fact the Web would not be possible if there were anyone in charge because being "in charge" generally means trying to keep things under control.
Instead of trying to maintain control, we put the onus on the participants to protect themselves. For example no matter what garbage a web server sends, your browser needs to be able to take it in stride. But this is no different than any social interaction that requires judgment. The Internet just provides connectivity and doesn't try to prejudge the uses of the connectivity.
Whatever … I Got My Broadband And My Web and Also My TiVo So Why Are You Complaining?
That's like saying that you have the golden egg, why do you need the goose? Or you had fish for dinner so why do you need to learn how to fish? (Apologies for these cultural allusions, I presume there are similar allegories in other cultures.)
You can think of the bursting of the so-called Internet bubble as a pileup as we hit a wall. While much of the value of such companies was indeed fantastic (in the sense of a fantasy) there was reality in the idea that anyone can innovate and create something new. After all, if the Web was so easy why stop there? But most of the innovation was about new ways to use the Web for commerce. What was less obvious but more important was the difficulty of creating new innovations like the web and email. We have been simply mining just one small vein when there are so many others.
It was simpler in the early 1990’s when the Internet was small and most participants had their own presence on the net and a high speed connection through institutions like universities. "Presence" means that others can easily connect to owns systems. It means they have a valid address on the network. To be very specific they had a 32 bit IPV4 address and anyone who knew that address could connect to them. To make this work better the DNS (Domain Name System) was created to provide stable names. Thus you could connect to tlb.lcs.xyz.edu instead of 123.321.123.321. In a small community, like a few schools, choosing names wasn't a big deal.
The Web gave companies a reason to connect their internal networks to the Internet and it gave home users a good reason to buy a PC. But Computers within these networks systems were second class citizens. They were either hidden behind firewalls or the whole network shared a single public IP address.
The community effect was so powerful that the relatively small community of participants who had a public presence provided enough value to drive the process. And, for each special case, such as the web, we could create a way of working around limitations. You could host a web site on an external server farm which seemed like a good idea anyway since they also came with 24x7 support. Home users rely on address translation (typically done by router boxes) because their access providers would provide only one or a small handful of addresses. At least they could browse the thing as big as the Web.
Remember that the Web itself was a simple mechanism that built upon earlier ideas like Gopher and SGML. The most exciting applications are often the most mundane. Just like the power of the Internet comes from its simplicity, the mundane ideas are typically not exciting because they are just plane useful and, once you see them, obvious.
That's Fine. But Don't Take Away My TiVo Or My Telephone!
OK, you're getting the sense that there is a conflict between the Internet and telephones and television. And you also notice that while Internet radio is starting to work you still don't make phone calls over the Internet and now that you have TiVo (the few of you who do) how can things be any better?
But you should ask why the cost of your telephone service has not had the same price/performance improvements you've seen in computers and in the rest of the Internet where a factor of a thousand or a million is not uncommon. Even more to the point, a telephone connection is a small audio stream compared with even a cable modem that has 10 to 100 times the capacity yet is priced about the same. And the cable modem is available 24 hours a day and you only use the phone part of the time. What's so special about a phone call that you need to pay a separate fee for that one application?
Television is even simpler though more bits. Simpler because it is a one way stream so we don't need the same kind of round-trip performance we need for a phone conversation. And a cable that carries a billion bits per second to one hundred homes can provide each of them with their own personal video stream that can be connected to any provider over the Internet. There is no need to have someone select what you can watch when you can make the choice yourself. In fact, DSL was originally created for delivering television not for connecting to the Internet. And that was at least ten years ago.
But it is really unfair to ask the companies whose profits come entirely from charging you for telephone calls or from charging you for subscribing to their choice of television stations to provide enough capacity in the first mile (or kilometer) from your home so you wouldn't need to pay an extra monthly fee for those applications. But it's far more unfair to tolerate the inherent conflict of interest between controlling transport and selling services that use the transport.
Where once there was a telecommunications industry there still is a telecommunications industry resisting change. But the Internet has proven that what we need are separate industries. One whose incentive is to provide more (and more and more) connectivity. And others to provide services and content. Many others. In fact there were many ideas that attempted to take advantage of the promise of the Internet. Some were very good and some were very bad. But regardless of their merits they ran into imposed by the current telecommunications companies.
The FCC is attempting to rescue these good ideas by providing the appropriate regulations. Unfortunately, doing so requires prejudging the winners and the loser. One lesson of the last ten years is that this kind of central planning doesn't work. What we can do is provide a level playing field for the ideas to vie for success. And that level playing field was the Internet. Was.
Bringing Back the Internet
The goal is to make the Internet simple again. Once again I should be able to just connect to the Internet and be a full participant and not just a visitor. What will it take?
Connectivity! The current telecommunications industry is the result of years of reinforced artificiality defined and redefined regulation rather than the marketplace. The Internet is built on connectivity which is a very simple commodity. With connectivity, services such as telephony and television delivery are very simple applications. Yet we have a telecommunication industry whose profits come nearly entirely from billing for these services and thus it must, in order to stay alive, assure that we do not have sufficient connectivity to do it ourselves. Were it not for the legacy of regulation the industry would have reinvented. At this point we must separate connectivity from content/services in order for both to survive and even thrive.
Every system must have a "real address". The four billion addresses of the current Version 4 of the "Internet Protocol" is far from adequate. In reality only a fraction of those are available for technical reasons. Version 6 not only gives us 128 bits for addresses (that's big enough to count all the subatomic particles in the universe), it allows us to wrap the new addresses around the existing addresses. This means that a local network that must share a single address can now build on that address so that every system has a publicly accessible version 6 address and thus it can be a full participant. This means that we can build V6 connectivity as an application on the current version 4 Internet – a dramatic demonstration of the power of the idea of moving the power to the edges.
We must encrypt all traffic. Encryption sounds like something only bad people do but in reality it is just like locking your front door and verifying your bank statement. It is simply responsible behavior. Protecting your use of the Internet from tampering is no different. It is also important because it protects the Internet from reverting to the old model of the phone network which was tuned for a single application. There is a tendency to forget that we don't know what the next innovations will be and thus favor the existing way we use the Internet. Encryption discourages such meddling and thus forces a focus on connectivity rather than counterproductive "favors".
Stop the .COM tragedy from unraveling the Internet. I've written about this in great length elsewhere. It is vital that we can rely on the DNS as a source of stability. As long is it must also serve complex and ambiguous social needs it can't be stable. And it is naively foolish to try to make the DNS serve as the equivalent of the trademark system. It seemed easy when all we had were the names of a few universities and not millions of John Smiths or Lees. We need to make the DNS a boring source of unique handles and leave defining names to the marketplace where it belongs.
While each of these steps is vital in its own right, together they give us the real Internet. The one we pretend we have but lost years ago.
© Copyright 2002, 2003 by Daniel Bricklin, Bob Frankston, and David P. Reed
All Rights Reserved.