Verisign and ICANN

February 20, 2006 7:18 PM

Bob Parsons of Go Daddy warns about the proposed contract between VeriSign and ICANN, allowing VeriSign a permanent monopoly on .COM and price increases without regulation.
posted by Protocols of the Elders of Awesome (29 comments total)

I'm torn... on the one side, a monopoly over the primary source of information distribution for the foreseeable future seems bad, but on the other hand, anything Bob Parsons is against I'm usually in favor of. That fucker's insane.
posted by jonson at 7:32 PM on February 20, 2006 [1 favorite]

""
posted by Fezboy! at 8:06 PM on February 20, 2006

What jonson said.
posted by purephase at 8:14 PM on February 20, 2006

Maybe Bob can rough up ICANN a bit?
posted by MikeKD at 8:28 PM on February 20, 2006

The irony of a man crying wolf over a potential monopoly in the same blog where he exhorts the business tactics of John D Rockefeller & Standard Oil is pretty amusing.
posted by jonson at 8:31 PM on February 20, 2006

Good points, jonson, but the ICANN/Verisign chicanery is of longstanding notoriety and really isn't refutable by dint of a self-interested rebuttal.

It's apparent that ICANN is broken in several major ways, probably permanently. There's little citizen input, and the US is not going to allow ICANN to internationalize.

I say this bullshit gets routed around, as damage. Let .COM atrophy to fresher, more sanely managed domains.
posted by dhartung at 8:35 PM on February 20, 2006

What would be a saner domain than .COM? .MUSEUM? .US?
posted by Nelson at 8:39 PM on February 20, 2006

.HAWT

It rhymes!
posted by WolfDaddy at 8:46 PM on February 20, 2006

Isn't this the same guy who advocates torture?
posted by Malor at 8:55 PM on February 20, 2006

Malor: yup.
posted by puke & cry at 9:12 PM on February 20, 2006 [1 favorite]

Proof that the world isn't black or white and that sometimes you're on the same side on an issue with your enemy on a different issue. What ever shall we do?
posted by smackfu at 9:36 PM on February 20, 2006

A monopoly over domain names? Deja vu.
posted by TwelveTwo at 9:58 PM on February 20, 2006

smackfu: "Proof that the world isn't black or white and that sometimes you're on the same side on an issue with your enemy on a different issue. What ever shall we do?"

Hitler was a vegetarian, ya know :P Figured I'd try to Godwin it early.
posted by symbioid at 9:59 PM on February 20, 2006

Continuing the derail: I think Hitler was a vegetarian for Gastrointestinal reasons, he would eat meat when there were no physical ramifications.
posted by cloeburner at 10:09 PM on February 20, 2006

Take anything Bob Parsons says with a grain of salt. I don't understand why ICCAN dosn't give .com to someone else though. bleh.

ICCAN is pretty pathetic.
posted by delmoi at 10:16 PM on February 20, 2006

What did ICCAN ever do wrong?
posted by Mijo Bijo at 11:03 PM on February 20, 2006

The only solution is to broaden the top level domain space. Make literally every combination of 1,2,3,4, and 5 letters a valid TLD then assign them at random and then depreciate .com .org .net. Let the current holders of the .com .org .net domains have 6 years to transition then shut them down.

That would fix anticompetitive deals like this and cut down on the squatting. Even at $5 a piece trying to register all 26^5 11881376 combinations of any string is going to be a problem.
posted by Mitheral at 11:17 PM on February 20, 2006

Mitheral... Since when is "solution" a euphemism for "destroying the internet"?

Are you kidding?

metafilter.qfxpm ? Yeah, the marketability of that will be fantastic. It's hard enough getting people to remember .net or .info instead of .com.

Cute, though.
posted by disillusioned at 11:22 PM on February 20, 2006

I'm not advocating metafilter.qfxpm (though I'm sure someone would want line noise like foo.hewpw) but how about metafilter.pine or metafilter.tree or metafilter.rock or metafilter.movie or metafilter.abcde or metafilter.zip or metafilter.link? There must be thousands of five character or less words and useful strings just in the english space. Add in other languages using the same character space and it must be tens of thousands.

One of the reasons the huddling masses have so much trouble with anything but .com is they aren't exposed to it. Like when you give someone the address http://ask.metafilter.com and they type in http://www.ask.metafilter.com or even www.http://ask.metafilter.com (shudder). People only have problems with .org because they are only exposed to .com; the abnormal, semi random TLDs would be no more confusing then area codes once people got use to it. And if they had to go to msn.ms and yahoo.party they'd get used to it really quick.

Oh ya one further improvement: you can't sue for trademark infringement at the TLD level. So if Bob's computer service registers computer.apple Apple can't litigate to get it turned over to them.

On the other hand this would put a serious load on the primary name servers so there is some work to be done there.
posted by Mitheral at 11:40 PM on February 20, 2006

or even www.http://ask.metafilter.com

Not having access to any kinda web stat logs here, I can only hope you're joking/exaggerating. Do people ever really do that? Type www BEFORE the http?
posted by jonson at 11:45 PM on February 20, 2006

Yes, they do.
posted by TwelveTwo at 11:51 PM on February 20, 2006

As long as "the network of networks" is built around something as dunderheaded as DNS, the domain racket will be sweet for racketeers like Verisign. The whole idea of DNS was kind of a hack, in the days when Jon Postel and others were still envisioning DARPAnet as some kind of egalitarian non-heirarchal routing structure. But that's all been out the window in any real sense, since the mid-90's.

For good or ill, the Internet now essentially runs on heirarchal routing. As such, it makes more sense to flatten name resolution into routing services than it ever did, and it has for years. Dump DNS, and its archaic "roots," let the routing infrastructure directly support name resolution to end nodes on arbitrary but unique strings, and disband ICAAN asap.

If I'm the first guy to identify myself as paulsc.swell why shouldn't I be able to be that indefinitely, for as long as I'm able to squat it, without asking anyone else about it? Why shouldn't paulsc.swell be simultaneously resolvable to as many independent addresses as I like, and am using at the same time? Why should anyone else really care what I've told the machines handling my various kinds of traffic to do with it, if my "zone of authority" is just paulsc.swell? It's not like the machines really care, after all.

This would be funny, if, like most of the rest of this madness that we call the Internet, wasn't so damn silly-sad...
posted by paulsc at 12:07 AM on February 21, 2006

jonson writes "Type www BEFORE the http?"

Oh ya. Confusion with SSL sites are also popular, our Outlook Web access server is encrypted so is at https://owa.foo.org. I seen http://https://owa.foo.org https://http://owa.foo.org and http://owa.foo.org.https. Keep in mind [e-mail] is their email address so the last is especially perplexing.

It points to a fundemental lack of how URLs work. I believe it is at least partially because the average low end user never need type more than somecompany.com into their browser (thank you very much Netscape and Microsoft).
posted by Mitheral at 12:08 AM on February 21, 2006

Mitheral writes "It points to a fundemental lack of how URLs work"

It points to a fundemental lack of understanding of how URLs work.
posted by Mitheral at 12:10 AM on February 21, 2006

Happily an ORG and will stay there until my dying day...
posted by Samizdata at 2:09 AM on February 21, 2006

let the routing infrastructure directly support name resolution to end nodes on arbitrary but unique strings

Well, it has done that since the beginning, if you think of your IP address as a 4-byte arbitrary but unique string. Of course, the unmentioned features you are likely looking for in addition to that are long-term persistence even when switching ISPs and some form of choice over your identifier string.

Technically, we could upgrade to a DNS replacement like the one you want with no trouble at all; in fact several P2P networks already use their own adressing schemes that could be easily extended to this end. The problem comes in convincing a significant part of end points to "upgrade" to this new scheme. But thats a social problem, not a technical one. (Although that doesnt make it any easier)

Still, the routing layer would be a horrible place for something like that, since all routing hardware would have to be replaced (which is/was also a major problem with IPv6) and you would have a lock-in when something even better comes along. Just put it on top of the IP/Routing-Layer, the way DNS works today.
posted by Cironian at 3:03 AM on February 21, 2006

"Well, it has done that since the beginning, if you think of your IP address as a 4-byte arbitrary but unique string."....
posted by Cironian at 6:03 AM EST on February 21 [!]

Not really, thanks to "rules" for IP networking and routing, which resist division of nets into discrete subnets, without a horrible amount of address waste. So, there's nothing "arbitrary" about an IP address, and that, too, is part of the problem. It's a major reason why routing has become hierarchal.

What I'm talking about is more like a re-think & implementation of IPX NLSP, for a global scale, hierarchal network. Mainly, a discovery/exploration mechanism for distant subnet resources, that is not unduly expensive (maybe something along the lines of gopher), needs to be built. Why would routing hardware need to be replaced? Routing hardware isn't likely to be replaced for IPv6 either, and even for a more radical protocol revision such as I'm talking about, mostly software just needs to be upgraded. Most Cisco routers can still route IPX, although the "big iron" used for backbone routing has no reason to still have such software loaded. I concede that for storage and operation of foreign subnet resources in an expanded NLSP type scheme, it may be desirable to retain and use the current DNS hardware a bit differently, but as you say, there is little point in having this remain a centrally "managed" (i.e. controlled) resource. The whole point is to do something far more efficient, flexible and robust, down to the client level. So it had better not be more resource intensive.

Once we have this all thrashed out here on MeFi, we will need to schedule an Internet Maintenance Day for the changeover, unless we can get this done in time for this year's event...
posted by paulsc at 3:53 AM on February 21, 2006

disillusioned: metafilter.qfxpm ? Yeah, the marketability of that will be fantastic.

...while the useability will remain undiminished, which is all that matters. people thinking like you, in terms of marketability of the internet, are what's fucking it up.
posted by quonsar at 4:26 AM on February 21, 2006

Yes, there is no more need for your domain name for a server to be marketable than there is for your telephone number. As long as you can communicate it with little trouble the domain name has served it's purpose.

Really what difference in the grand marketing scheme would there be if IBM's web address was ibm.tree or ibm.comp or ibm.serv or ibm.ibm or ibm.blue?
posted by Mitheral at 8:13 AM on February 21, 2006