From: Barry Shein <bzs@world.std.com> Date: Fri, 12 May 2006 15:45:46 -0400 Subject: Re: MEDIA: ICANN rejects .xxx domain
On May 12, 2006 at 14:51 tv@pobox.com (Todd Vierling) wrote:
The complexity added by TLDs has one extremely critical good side effect: distribution of load by explicitly avoiding a flat entity namespace. The DNS has a hierarchical namespace for a reason, and arguments to the contrary will convince on the order of sqrt(-1) people.
As if you couldn't just hash on whatever the last component is and pick a server on that basis? Query(server[Sum(bytes) mod Nservers])?
That's right, you =couldn't=. In the first case, *WHO* runs that server? What if you are the -only- hit in that hash bucket? What do you do if *nobody* is running a server for that hash bucket when you want to register a name that hashes into it? The current DNS architecture has a 1:1 correspondence with 'levels', 'zones', zone administrators, and administrative authority. Every 'TLD' has its own, *independant*, administrative policies. Some of them have 'structured' second levels, (e.g. .uk., .tw., .jp) others *don't* (e.g. .no, .fr, .ca, .ch). If you just eliminate the top level, then *which* ("in the end, there can only be one") of the various '.com.{CC}" registrars gets to control the 'new' ".com", and what happens to the registrations in all the _other_ '.com.{CC}" 2nd-levels that are now disenfranchised? If you eliminate all the 'structured' name elements, you have a 'mell of a hess' of name collisions to deal have to resolve. *who* gets to use 'McDonalds', for example. the American hamburger chain, or the Scots Clan? Who gets to use "yellowpages"? (anybody remember why Sun had to change the name of their network directory service?) who gets "shaw", 'shaw.ca', or 'shaw.com'? They're *not* the same company. :)
There are probably good answers to people's suggestions for change but working backwards from "that's the way we've always done it" with trailing remarks intended to stifle a response isn't, to my mind, an answer.
It is 'unworkable'. because there's a *whole*lot* more to it than just the technical matter of 'serving' DNS records. And any proposal that views things from _that_ standpoint only, is, by *definition*, defective.
The best answer I can think of off-hand is that dropping .com etc wouldn't add much, if anything. Any savings in typing would be off-set by having to generate non-colliding names which would've been .com and .org, etc. It would just be creating a new TLD, the null TLD moving collision avoidance left by one.
It also eliminates one layer in the 'distribution' of load. resulting in a several (decimal) order-of-magnitude concentration of (a) data, and (b) queries, to the top-level servers. The effect of 'b' can be ameliorated by deploying that same several-orders-of magnitude number of additional root servers. the increase in the size of the 'local' database is a whole different issue -- *and* the fact that is is being more-or-less constantly updated; *unlike* the 'root' zone, where the TLDs are fairly static. An argument could be made for splitting '.com' into, say '.com01' through '.com64', by arbitrarily re-assigning all the existing .com's into random members of the new set. *AND* prohibiting any 'party' from owning/controlling the 'same' second-level name in more than one of those TLDs. Continue to allow 'challenges' to a registration, based on trademark, etc. but "deactivate" the 'infringing' registration, rather than 'turn it over to the other party'. This allows for multiple parties who _have_ rights to the same name to all use the same '2nd level' element w/o conflict. A better argument can be made for eliminating almost all of the 'generic' TLDs. '.int' (or an equivalent) needs to be kept around for organizations that are _not_ under the jurisdication of any national government. The other 'classic' generics (.mil, .gov, .net, .com, and .org) could be folded into '.us' as structured 2nd-level names, as is done in many other national TLDs. Note: this proposal _also_ defuses the issue of having ICANN establish UDRP. Since the all the names conflicts now occur under a single body of law, already-established court procedure in *each* jurisdiction can be used to resolve issues in _that_ jurisdiction. "Bye-bye" to the cross-border conflicts/issues on *that* matter. *IF* one is going to have 'generic' TLDs, I'd suggest they should be reserved for major multi-national operations. e.g. the '.net' TLD reserved for network operators with 'physical presense' in at least (say) 25 countries, a .net.{CC} domain in each of those countries, _and_ 'direct' allocation of an IP address block for each of those 'national' operations.
As to .XXX:
To my mind the real camel's nose in the tent is that to create it would seem to urge or at least validate its enforcement and coercive means would necessarily arise (civil lawsuits, criminal charges, regulatory apparatus.)
*ABSENT* a prohibition on such 'content' under other TLDs, there is nothing to enforce. It was proposed as a place where porn operators _can_ register, not where they _must_ register. *WITHOUT* even getting into the differences in 'definiton' for what would have to be so registered, depending on geography. What do you do with stuff that is _legally_ displayed in public, and openly sold, in Denmark, that mere possession of would get you arrested in the U.S.A.?
Otherwise of what use would it be, in terms of the conceptions of its champions as opposed to unintended consequences?
The champions "think": 1) that all the purveyors of such would _voluntarily_ choose to register there, because, by having a 'place' where such stuff is 'officially sanctioned', they won't get hassled for having/selling such stuff. 2) that those who want to block access to such stuff would have a very simple and effective means of doing so.
The deeper problem is the conception by many (unwashed) that someone must be in charge,
Yeah, it's a misconception, but can you point to _anything_else_ of a man-made nature where there is nobody that is in charge, nor ultimately responsible for it? What we have here is an 'act of man' that is on a par with an 'act of nature', or alternatively, an 'act of God'. It's not really suprising that the 'unwashed' don't 'get it' -- it just may qualify for the term 'unique' in that respect. Whether this uniqueness is a 'good thing', and how long it can survive, remain to be seen.
we used to get calls asking for contact info for the Internet complaint dept, and they didn't mean us. People were often shocked to hear that we had no answer.
I'm glad _I_ didn't get any of those calls -- it would have been *VERY* *DIFFICULT* to resist giving out Al "I was instrumental in the founding of the Internet" Gore's name and number. If he wants the credit, he gets the blame, too! :)
And widespread conceptions like that have a way of materializing, sans some force of resistance.
I suppose some may say it's 10 years too late for that comment.
-- -Barry Shein
The World | bzs@TheWorld.com | http://www.TheWorld.com Purveyors to the Trade | Voice: 800-THE-WRLD | Login: Nationwide Software Tool & Die | Public Access Internet | SINCE 1989 *oo*
From owner-nanog@merit.edu Fri May 12 14:51:36 2006 From: Barry Shein <bzs@world.std.com> Date: Fri, 12 May 2006 15:45:46 -0400 To: "Todd Vierling" <tv@pobox.com> Cc: "Jim Popovitch" <jimpop@yahoo.com>, nanog@nanog.org Subject: Re: MEDIA: ICANN rejects .xxx domain
On May 12, 2006 at 14:51 tv@pobox.com (Todd Vierling) wrote:
The complexity added by TLDs has one extremely critical good side effect: distribution of load by explicitly avoiding a flat entity namespace. The DNS has a hierarchical namespace for a reason, and arguments to the contrary will convince on the order of sqrt(-1) people.
As if you couldn't just hash on whatever the last component is and pick a server on that basis? Query(server[Sum(bytes) mod Nservers])?
There are probably good answers to people's suggestions for change but working backwards from "that's the way we've always done it" with trailing remarks intended to stifle a response isn't, to my mind, an answer.
The best answer I can think of off-hand is that dropping .com etc wouldn't add much, if anything. Any savings in typing would be off-set by having to generate non-colliding names which would've been .com and .org, etc. It would just be creating a new TLD, the null TLD moving collision avoidance left by one.
As to .XXX:
To my mind the real camel's nose in the tent is that to create it would seem to urge or at least validate its enforcement and coercive means would necessarily arise (civil lawsuits, criminal charges, regulatory apparatus.)
Otherwise of what use would it be, in terms of the conceptions of its champions as opposed to unintended consequences?
The deeper problem is the conception by many (unwashed) that someone must be in charge, we used to get calls asking for contact info for the Internet complaint dept, and they didn't mean us. People were often shocked to hear that we had no answer.
And widespread conceptions like that have a way of materializing, sans some force of resistance.
I suppose some may say it's 10 years too late for that comment.
-- -Barry Shein
The World | bzs@TheWorld.com | http://www.TheWorld.com Purveyors to the Trade | Voice: 800-THE-WRLD | Login: Nationwide Software Tool & Die | Public Access Internet | SINCE 1989 *oo*