ICANN/GNSO GNSO Email List Archives

[ga]


<<< Chronological Index >>>    <<< Thread Index >>>

Re: [ga] A Windfall for VeriSign?

  • To: ross@xxxxxxxxxx
  • Subject: Re: [ga] A Windfall for VeriSign?
  • From: Karl Auerbach <karl@xxxxxxxxxxxx>
  • Date: Wed, 17 Oct 2007 00:00:07 -0700


Ross Rader wrote:

I'll leave the rest for others to comment on, but I will flatly state that the DNS and the registration is a public resource and it needs to be managed as such. This isn't an analogy, its a statement of fact.

I take a rather different point of view.

I perceive DNS as an application that is layered onto the internet.

It is a useful and important application, but it is just an application that users may chose to use, chose to ignore, deploy their own, or invent something else.

Most people do not like surprises and would not invite inconsistency into their lives, so most people chose to ride on the main deployment of DNS.

And it is appropriate for that main deployment to manage itself.

But it is not appropriate for that main deployment to eject the choices and innovations of others who, to usurp a phrase, think different.

That right to think and do differently is the heart and soul of the concept of innovation at the edge, the end-to-end principle, and the first law of the internet (http://www.cavebear.com/cbblog-archives/000059.html)

I don't see DNS deployments as "resources" in the sense that they are some sort of exhaustible pool - As I mentioned elsewhere the number of possible TLDs is a number rather larger than this:

11,144,421,984,854,529,111,291,814,965,840,121,701,917,784,688,171,700,627,654,810,062,931,821,453,496,825,690,394,892,284,041,625

In other words, we could assign a trillion trillion TLDs to every electron in the universe and still have mountains of TLDs to spare.

And inside *each* TLD we can have that same number of second level names. And inside each second level name we can have that many third level names, and so forth until we hit the maximum size of a DNS name.

And if that isn't enough we can have as many different DNS systems on the net as we chose to construct; in fact each user of the net could have his/her own private DNS system - I've tried it, it does work.

The issue is not "resource" or "public resource".  Rather it is consistency.

There is little incentive for people to create clashing DNS systems; there is much incentive to create DNS systems that mesh without inconsistencies.

In other words, the idea that we need a system to coerce consistency is an idea that is duplicative with, redundant to, and repetitive of a natural self interest towards consistency.

Do we have a regulatory body to ensure that objects fall in response to gravity? No, that would be silly even though quantum physics says that there is a very small chance that a thing could fall upwards. Yet we have an ICANN that tries to mandate that there will be no DNS but ICANN's DNS as an equally strange means to prevent DNS inconsistencies that will, through the natural course of events and the drive of human self interest, rarely arise.

Had ICANN never occurred it is my belief that we would have had, by now, a viable internet with a multiplicity of DNS systems, nearly all being consistent with one another. (Some groups would chose to be intentionally different, but that would be their choice and their burden to bear.) On such an internet we would not have the singular point of failure and attack that today's DNS represents, and it would have eliminated the vast internet tax - over $500,000,000 every year as measured in excessive domain name fees - that ICANN imposes on the net.

What is "consistency"? Is it perfect alignment of name spaces? No, not even ICANN's singular DNS can provide that - DNS is always shifting as servers update information asynchronously from one another, as caches fill and time out, and as resolvers make random choices of the order in which they package resource records into responses. (And DNS names are terribly unreliable as stable names over even short periods of time.)

Consistency is a mode in which, if a TLD exists at all, its contents are the same, modulo updates and caching, no matter where a name in that TLD is uttered and no matter by whom that name is uttered. And we have a time honored means to enforce this - the trade and service mark law of individual nations and the associated inter-national mechanisms.

But consistency does not require that every TLD exist everywhere any more than every supermarket is obligated to carry exactly the same brands of products - as long as those who consume names are satisfied or have the ability (directly or via ISPs acting as their agents) to chose a DNS provider more in keeping with their expectations.

Had we had such a mechanism, a mechanism that is precluded by ICANN, we would not have the tremendous TLD selection issue.

Rather, new aspiring TLDs could try to make their way into the inventory of DNS providers, much as new TV stations try to find their way into the offerings of cable and satellite TV providers, much as new brands of laundry soap fight for space on the shelves of supermarket chains.

ICANN represents a deep lack of faith in the workings of the marketplace and, instead, harkens back to the ideas of centrally planned economies such as espoused in the Five Year Plans of the now defunct Soviet Union.

So, at the bottom of things, I find the words "public resource", and "trust", and even "property" to be conclusory terms that keep us from getting down to the level where we can find answers that give us the kind of internet that we want.

                --karl--




<<< Chronological Index >>>    <<< Thread Index >>>