For the past two years, Karl Auerbach has made a hobby of criticizing ICANN, the Internet Corporation for Assigned Names and Numbers. He has called the Net's controlling authority over domain names everything from inept to "an organ of the trademark lobby." But on Tuesday the 50-year-old "wild-eyed radical," as he often calls himself, became part of that which he loathes: one of five new members of ICANN's board of directors.
Can the ultimate outsider transform an organization from the inside? Not even Auerbach is sure. "I'm nervous," he says. "I feel like I just signed up to replace Sisyphus, so that he can go to Hawaii while I undertake to roll the boulder up the mountain every day just to have it roll back down every evening."
Auerbach is willing to push anyway. He's confident that the voters who put him in -- everyday Net users who registered through a convoluted online process at ICANN -- made the correct decision. He might be right. The cheerful, bearded 50-year-old has activism in his blood. ("My grandparents worked to promote labor unions, and my father worked to redress imbalances of power in the California TV repair industry," he says.) His own life reflects an eclectic, uniquely Californian mix of technology, law and protest.
This strange brew first appeared during Auerbach's teenage years. Growing up in Van Nuys, Calif., during the 1960s, he developed a love for math, attended the same high school as Internet pioneer Vinton Cerf and began to question what were then the early stages of the Vietnam War. In 1966, he joined an antiwar congressional campaign; in 1970, while studying physics at UCLA, Auerbach found himself running from the Los Angeles Police Department during a riot sparked by the news of the Kent State killings.
Soon after that, computer science struck his gray matter's fancy, and after meeting up with Cerf -- whom he didn't know in high school -- Auerbach began working on the beginnings of TCP/IP. This led him to software, and throughout the '80s and '90s, Auerbach spent most of his time founding or helping to start small infrastructure companies that helped networks work more efficiently. These firms typically ended up being acquired by larger corporations: Epilogue Technology Corp., which Auerbach founded in 1986, is now part of Wind River Networking Products; Precept Software was acquired by Cisco in 1998, which is where Auerbach now works as a researcher.
Through it all, Auerbach maintained his passion for political protest. He says he earned a law degree "because I had been subjected to what I thought was an illegal search by the L.A. sheriff's department and I was curious whether it was, in fact, illegal." (It was.) And when ICANN formed in 1998, under dubious circumstances, Auerbach took up the cause, arguing for a more democratic, less corporate structure. He even formed the Boston Working Group in September of the same year, drawing 1,000 people to the online policy think tank to discuss "the management of Internet names and addresses."
But it's one thing to be a critic and quite another to try to effect change. We spoke to Auerbach about his plan for reform, how he hopes to implement it and what he'd like to see ICANN become.
You've condemned ICANN repeatedly, but in your ideal world, what would it look like? How would you like to see ICANN changed?
We're talking about a California remodeling job, where you knock down the whole house but for one wall and build a new house around it, then tear down the remaining wall. Essentially that's what ICANN needs. It needs a fundamental, ground-up restructuring. I'm talking about a restructuring to the point where the supporting organizations -- such as its law firm -- need to be redefined, if not eliminated; where the board members come exclusively from the at-large membership votes; where everything that ICANN has done so far is subject to a very short sunset provision and has to be reenacted lest it expire. I'm talking about a major overhaul.
How on earth do you go about changing this -- you're one man on a board of 19.
Yeah, well, I'm going to lose a lot. However, there are other things that a member of a board of directors has. The first thing can be seen in the word "direct" -- a director directs. That's an action word. A director must act, a director must direct, must make well-informed decisions. To do that, a director has the right to inquire about and examine all the records, documents and procedures of a corporation. Very little is hidden from a director's eyes.
This gives a director an enormous power to know what's going on in the corporation and to expose improper activity.
Also, by my behavior, I intend to exemplify "open, transparent and accountable." I intend to be a standard by which other board members will be measured.
What else do you bring to the table?
Well, ICANN is moving into a realm where knowledge of technology is actually going to make a difference. And I know the technology. For example, all this business about top-level domains and how many the Net can support -- there's been no technical discussion of how many it can actually support. And most people don't understand the issues of content management, which is very definitely changing domain names from being globally meaningful to locally meaningful.
A lot of people think "domain name" uttered anywhere means the same thing. They think a name is just a name is just a name. Yet it costs a lot of money to move content around the network, both in terms of bandwidth and in terms of time. Users get really tired of waiting for stuff to get dragged across the Net. The idea is -- and Akamai and other companies are doing this -- you move content, you spread it around so it's replicated, so when somebody asks for it, you intercept the domain-name query and you look at it and say: "Where is this user coming from? Where is the closest place he can get the content?" And your DNS [domain name system] answers, then points the user to the place that's closest. Therefore, we've got geographically sensitive domain names.
And there's this strong financial pressure by data providers and ISPs to manage the content so it can get to the users more efficiently, in terms of not generating so much bandwidth consumption. This is big bucks -- we're talking billions of dollars here. ICANN doesn't understand this. It's assuming that domain names are this globally unique name space, when in fact they are evolving to become more of a personal domain space.
So unless one understands the technology and the implications of it, when one enacts rules regulating it, you essentially pour concrete around the technology. You inhibit, if not prevent, innovation. By creating all of these laws about rights to names and structures and DNS, we don't know what innovation we're impeding.
Do you remember the Hushaphone decision back in 1956?
No. I'm much too young to remember anything that happened in 1956.
Ah, it's a very interesting case. In the 1950s, AT&T was The Phone Company, this huge monolith, and it and the Federal Communications Commission -- which was pretty much subject to AT&T's whim -- had an iron fist of control over the telephone system in the United States. But there was a little company called Hushaphone, which built a plastic and aluminum widget that clamped onto the mouthpiece of the telephone. As with when you put your hand around the mouthpiece, it cut out exterior sound and helped focus your voice into the microphone. It was a totally passive piece of equipment. AT&T went ballistic. It said, "You can't sell that product; it will damage and destroy the telephone network."
And the FCC said AT&T was right. "We can't have linemen blown off the telephone poles from a high-voltage shock because someone's using a Hushaphone" -- that sort of thing. It actually took something like 20 or 30 years to get to court, but the U.S. District Court took a look at this thing and said, "AT&T, you're full of it. This little widget cannot possibly harm the telephone network. Plain common sense tells us that."
So the District Court enacted a standard that said any action that is privately beneficial is permissible on the telephone network, as long as it's not publicly detrimental. That's an important standard; that was the crack that broke AT&T apart. Hushaphone was the start.
We need to revisit that rule. We need to realize that large technical entities sometimes don't tell us the truth. We have to realize when we're being subjected to a snow job. It takes some technical savvy to look at something and say, "This emperor has no clothes."
Where does your relationship with Cisco fit into this?
I don't speak for Cisco and I never will. Cisco is not like the old companies where the president gives an order and everyone marches off wherever he says. We are more of the cats. And most of us are financially independent, so we don't have to follow orders anyway.
Are you going to have enough time to do for ICANN what you would like to do?
Yeah, I think so. It will of course interfere with my work, but it's been interfering with my work for a couple of years. I'll probably continue about the same kind of schedule, which is work, then say "Hi" to my wife in the evening, then go do ICANN stuff. She doesn't like that.
Aren't you also doing DARPA [Defense Advanced Research Projects Agency] research?
I'm not sure how much of what I'm doing is yet open for public release, so I'll be a bit vague. But I'm doing research with folks at UC-Berkeley on means to build what I like to call "autopilots" for the Net. The idea is to add intelligent control systems and make use of process control principles to exercise that control and keep the Net operating within limits established by the network administration. This kind of thing will become increasingly important as the Internet moves toward providing "lifeline" services and time-sensitive applications, such as I.P. telephony.
OK, let's get into some of your more specific gripes about ICANN. What do you think is ICANN's biggest flaw?
First of all, ICANN is not a legislature, yet it is enacting what amounts to a law that is above all other nations on the planet. The Uniform Dispute Resolution Policy [for domain names] supersedes U.S. trademark law or any other trademark regime in any other country. ICANN will claim that it is merely private and contractual, but the 55-mile-an-hour speed limit was not imposed by federal law directly. It was imposed by the federal government saying to the states, "We will withhold money unless you -- the states -- enact a 55-mph speed limit." So just because it's done by contract doesn't mean ICANN is not enacting what amounts to a worldwide law.
And ICANN doesn't have the structures associated with classical government: a full-participation review, a tension between various opposing forces -- a separation of powers. It doesn't even have a notion of due process. It essentially is an oligarchy that operates on the same principles that Louis XIV did. So it's an inappropriate body to enact what amounts to worldwide legislation.
Look at most of the UDRP cases. We have the Brazilian soccer team taking away Corinthians from someone in the U.S., for example. I'm just waiting for the second shoe to drop, which is when the city of Corinth in Greece goes after the soccer team in Brazil.
But if your vision of pure democracy is realized, won't ICANN be crippled -- even slower than ever? After all, not even the Founding Fathers decided that pure democracy would work.
Sure, it's chaotic, but we only had 3,000 people from North America vote in this election. And we're talking about a group that has communications possibilities well in excess of anything that was imaginable to people in the 1780s.
And what's wrong with moving slowly? We're in a whole new environment.
How do you think ICANN should handle the addition of new top-level domains, which may be the first piece of policy you have a hand in?
ICANN should be in the business of giving away top-level domain slots -- not names, but slots. A slot is a chance to put the name of your choosing into the root zone. ICANN should not look at the semantics of that name whatsoever. ICANN should be absolutely blind to the meaning of that word in any given language. All ICANN should do is check to make sure that the name is not being used already.
Secondly, ICANN should keep its hands absolutely out of the way the top-level domain is operated. It should allow them to go out of business. It should allow them to impose their own charters at their own choosing. If someone gets a slot and wants to call it "foo" and only include Web sites about "foo" birds, that's the business of foo's top-level domain. ICANN should not try to be a policeman of charters or anything else.
How many names should there be?
We can easily handle 10,000 top-level domains per year. By actual experiment, I've determined that the domain-name system can technically hold at least 1 million top-level domains with no problems whatsoever, probably many times that.
Should companies get any protections from ICANN?
No. There are existing trademark laws that they can use. What the trademark people want is fast recourse. They don't want to have to go to Brazil, which happens to have a domain they want; and they've gotten this and more. But why should we give trademark people this speedy access to supernational jurisdiction? I'd rather let the law evolve along the lines that it evolved over the past 2,000 years, which is by trial and error, finding out what works a case at a time.
The legal system has these techniques. And of course, there will be injustices along the way, but we already have plenty of injustices with the UDRP. And what's more frightening is that the whole ICANN legal system doesn't have the same checks and balances that the true legal system does. The true legal system is very much aware that it makes mistakes: It has notions of review, of overturning prior judgments. It's all built into the system. But ICANN doesn't have any of that.
Sounds like a libertarian stance -- are you in fact a libertarian?
Not really. I'm very much in favor of strong government regulatory regimes when there is a need for them. I've read a fair amount of history -- particularly the history of the U.S. since the Civil War -- and it is clear to me that the development of the regulatory state was a good thing and well justified by the abuses. I don't see that those abuses have waned, and thus there are still good reasons for strong regulatory bodies.
That said, I do recognize that there is value in recognizing that there are areas in which nongovernmental coercive forces -- such as Adam Smith's "invisible hand" -- are still a valid, and even preferable, alternative to a governmental regulatory body.
In the DNS space, it's my feeling that ICANN should limit itself to handing out lots and lots of slots in the root zone and let the operators of those slots run them as they please. It's my feeling that this is an area in which we can allow nongovernmental and nonregulatory forces to work (in conjunction with the preexisting legal framework that penalizes things like defamation and trademark infringement).
In the I.P. address allocation area, I take the reverse point of view -- that a strong regulatory body, and ICANN, are indeed needed. I don't think in the I.P. address area that economic forces will produce a solution that is good in terms of being open to new entrants or future flexibility of the Net.
And as for ICANN's third role -- protocol parameters -- that's simply a useless appendage to ICANN. It reminds me of something that Dickens came up with -- the Circumlocution Office -- a governmental entity with no positive function but that took it upon itself to make sure that no other governmental entity could do its job. I tend to feel that ICANN ought to lop off its useless "protocol parameter" appendage. I wouldn't call that "libertarian"; I'd call that rational.
Ultimately, do you really think you fight the system while working within it?
Of course. If I didn't believe that, I'd be fighting ICANN's systems tooth and nail. But I'm afraid of the vacuum that would result if ICANN were to disappear completely.
Where do you see the organization going?
I believe ICANN was given two distinct and contradictory roles, one of which was to establish itself, the other of which was to try to come up with policies regarding domain names. ICANN is a new experiment in international government. There's no source of authority for that. The only place it's going to come from is historical acceptance, and the only way it's going to get that is by doing things that are right, so that over time, people say, "You know, ICANN's doing the right thing." They'll just come to accept it, and ICANN will grow to power by that means. To do that, you have to get people to accept the decisions and to accept that they've had a meaningful participation. So to me, ICANN's race to put policy into place without creating the structural integrity and trust was a complete error. And it's suffering from it now. No one trusts ICANN anymore, and it may not be recoverable. I'm going to try to recover it. It also means maybe we have to throw out a lot of what's been done and start all over again, with everybody involved this time. It's a brave new world.