Fellow, Berkman Center for Internet & Society at Harvard Law School
Project Manager, Harvard-Jamaica Technology Education Project
We gather today for the Jamaica Internet Forum, among us are many who are experts in various aspects of Internet business, development, and technology. But how many among us know how it all got started? And how many know who governs it? How can we put so much trust in the Internet—investing millions in its development, in technology and Internet education, and in businesses that depend on it—when we don’t know who’s making sure it will still be there tomorrow?
I’m a teacher of Internet law at Harvard, so it’s my business to know the history of the Internet and to keep up with who is pulling all the strings behind the scenes. You might have guessed from the description of this panel that much of the answer lies in a mysterious organization called ICANN (the Internet Corporation for Assigned Names and Numbers). I’ve had the benefit of working closely with ICANN, first by webcasting their meetings, then as a member of their policy staff. For this panel I’ve been asked to speak to the significance of ICANN for Jamaica and Jamaican ISPs. Rather than tell you what I think Jamaica should do, I’ll use the next few minutes to tell you the story of the Domain Name System and to give you an idea of what ICANN does. Then you can decide if and how it is in your interests to get involved. For those already in the know, you’ll find that my history elides many of the details of the development of the Internet. It is my intention to get some of the important developments across in the short time allotted.
In the beginning, (sometime in the 1960s), there were a few big computers running at some famous academic institutions, such as Harvard, MIT and Stanford. Funded by the United States department of defense, researchers at those schools developed a way to connect those computers together and allow the transfer of information from one to another. At that point if you were working on one computer and you received information from another computer, there was no confusion about where it came from: there was only one other computer in your universe. The reverse is also true, if you wanted to send information from your computer there was no confusion: there was only one place to send it.
And why was the U.S. Department of Defense interested in this seemingly academic endeavor? They didn’t want to connect just two computers in a room, they wanted to connect computers across great distances in such a way that if a nuclear attack took out an important computer, the information would still be safe elsewhere and the rest of the network would still be functioning. To make this work, the computer scientists used the existing telephone system to allow information to flow everywhere the phone lines could go. Now there was a problem if you received information: who did it come from and was it for you? And if you wanted to send information, how did you address it? Each computer was assigned a number. You just had to keep a list of who was at each number and then address information to that number. This worked well at first, but soon it became difficult to remember the numbers. To solve this problem, computers were assigned names to go with the numbers. While the number of computers on the network was small, this was a pretty good system, but as more an more computers joined, problems arose: it began to take a long time to look up something on the list, and the poor guy who was maintaining the list was spending all his time adding new entries.
The Domain Name System, as we know it today, was the solution “they” came up with for this problem. That leaves at least 2 big questions: How does the DNS work? And who are “they”? The Domain Name System is actually quite simple. It uses the same list idea, but breaks it up into a pyramid-like structure. This is best illustrated by a diagram:
When your computer asks for the website http://law.harvard.edu, your computer first asks the list at the root where to find the list for .edu. The response of the root is the IP address of a computer that maintains a list of all the domains inside .edu. Your computer then asks the computer maintaining the .edu list where it can find harvard. Finally it asks the computer maintaining the harvard list where it can find the law computer. The law computer returns the web page you wanted. The benefit of this system is that it speeds up lookup (because the list to look through is relatively short at each level) and that it breaks up the work of maintaining the list, meaning that the list can be updated much more easily. That is, if Harvard decides to add another school, it does not have to contact “them” to do it. It maintains the computer that has the Harvard list, (currently including entries like law, business, medicine, education, etc.,) and can simply add the new school to the list itself.
So who are “they”? This is the most incredible part of the story. “They” (at this time, which is throughout the 70s, 80s, and much of the 90s) are an informal association of computer geeks who call themselves the IETF (Internet Engineering Task Force). They are almost all male, almost all American, almost all white, and almost all have big bushy beards and long hair. They are not paid for the work they do, but do it because they want the Internet to work. When a problem arises, such as the list of names and numbers getting too long, they discuss the best solution and make a decision by consensus. They determine consensus by getting together in a room and humming for proposals of which they approve. The idea is that no one person can hum any louder than any other, so in this way they are able to determine if there is rough consensus around a particular proposal. This system worked until the mid-90s, as we’ll see.
Among this group there was a man named Jon Postel. He was a student and then a professor at the University of Southern California. Because he was respected and trusted in the group, and because he volunteered, he was given the job of maintaining the list. When the DNS was created, he maintained the list for the root and for most of the top-level domains. If you wanted a new domain added, you wrote him to ask for it and he made a decision about it. He was not paid for this service, and the Internet was essentially run out of his office. Amazing.
At some point a friend of his in Britain called up and asked him to create a domain for the U.K. He reasoned that the U.S. had the .gov and .mil domains for themselves, leaving the rest of the world to use the inappropriate .com, .org, and .net for their government business. This made sense to Jon and he created a domain called .uk. Almost immediately after he realized he had a big problem on his hands. Soon lots of countries would be asking for domains and he would suddenly be in the middle of a serious geopolitical situation. What if two countries wanted the same extension? What if a group that was not internationally recognized as a country wanted a domain? He solved the problem by finding an International mailing codes list and creating domains for each entry on the list, giving Jamaica .jm. This was the beginning of a landslide of issues that turned on social and political concerns—as well as technical—that could not and should not be decided by the IETF alone.
After a while the .com domain was getting very large and Jon was no longer enjoying running it. He asked the U.S. government to put out a contract for a company to run the domain, and Network Solutions, formed for the purpose, was given the contract. A few years later Network Solutions proposed to the government that they stop paying on the contract and instead authorize Network Solutions to charge customers for registering domain names. The government agreed. Network Solutions became very rich.
To make a long story shorter, many people were unhappy. Consumers didn’t want to pay Network Solutions prices for domains, which were quite high because of their monopoly market. Other companies wanted to get in on the business to compete with Network Solutions. As the rest of the world was coming online, people were unhappy about the U.S. domination of the Internet. Postel and the IETF were unhappy with the way in which Network Solutions was usurping their authority.
Out of this mess Postel managed to precipitate an even greater crisis that caused the governments and big businesses of the world to take notice. Postel, wanting to show NSI who was boss, demonstrated that he could take control of the DNS from Network Solutions. Remember the root lists I discussed before? The ones that told your computer where to find the list for .com. We’ll there were 13 of them spread around the world. All of them except for the two at NSI were run by friends of Jon. Currently they were pointing all .com requests to NSI. Postel asked half of them to switch so they pointed to his .com list instead. Careful not to create a real disaster, Postel made sure his list was the same as the NSI list, but it demonstrated to NSI and to the world that the entire Internet could actually break in a matter of hours.
ICANN was formed to solve all these problems. It was a non-profit organization, structured to take input from people all over the world and tasked with managing the domain name system of the Internet. Since its formation it has operated on shaky foundations with people all over the world doubting its legitimacy and its decisions. No one has sought to overthrow or replace it though, perhaps fearing what would happen to the Internet if ICANN were no longer there. It successfully broke up the Network Solutions monopoly, introduced a procedure for solving disputes over domain names, and introduced new top-level domains.
The one thing it has failed to do is draw widespread participation from Internet users or even ISPs. From my perspective, it appears that the work of ICANN is exceedingly important. However, it also appears that if we feel we can trust ICANN to manage it in a reasonable way, there is little reason for the average Internet user or even ISP to get involved. It is in ICANN’s interest, our interest, big business’ interest, and the world’s interest to see the Internet operating well, spreading widely, and costing as little as possible. As long as these interests are aligned we have little to fear from ICANN.