John Riordan gave a good probable reason behind the choice, hearsay or not, and it matches both MAP's and several other people's public and private speculations. (Though I did like the comments about military-industrial complex conspiracies, the military being privatized, and the Great Conspiracy by Yet Another Branch of the US Government to Put Money Into SAIC's Pockets $50 at a Time). I'd like to take a quick poll of feelings about this sort of thing. Many organizations (mine included) have in the past gone for numbers of domain-names, each reflecting some different operating unit or other. A number of these happen because of technical concerns on the part of the holder of a more generic domain, or because of concerns about bureaucratic such-and-so on the part of the people looking for a subdomain. Looking around a bit, completely unscientifically and without more than eyeballing things, it appears that this practice is continuing, despite the back-pressure of a registration fee levied by the InterNIC. I was wondering if, firstly, anybody else thinks that having many second-level domains per company is a real issue on any front, and whether it really needs fixing by perhaps us suggesting that subsequent domains be charged on an exponential scale, with proceeds going into the costs of maintaining the worldwide DNS, particularly with respect to user-and-administrator education. Note that I shall happily ignore the question of who we should suggest this to, or who "we" should be (NANOG or I*-something-or-other); I'm just wondering if I'm completely out-to-lunch on this one. Also I shall happily ignore the issue of how one decides which organizations are considered part of the same <splat> of operations/control/management/ownership, as there be ratholes lurking there. However, I'd like to see something that handles not only <eighty-product-names>.COM being registered by one party, but also <eighty-quasi-independent-orgs>.COM, all clearly able to fit (at least for the moment) into <eighty-subdomains>.BigCompany.COM. The "(for now)" is key; some of the interesting things seen in the past has included people rushing out and grabbing lots of similar-sounding domains prior to a naming decision of a new spin-off-company being made by its board of directors, and also lots of rushing out and acquiring domains by products which think that they _might_, some day, be spun off into completely independent entities. If people think this sort of thing is OK, I'll shut up now. If not, I'm interested in hearing about it. Sean.
On Mon, 22 Apr 1996, Sean Doran wrote:
Looking around a bit, completely unscientifically and without more than eyeballing things, it appears that this practice is continuing, despite the back-pressure of a registration fee levied by the InterNIC.
Sean, you misunderstand. The fee is not a back-pressure, on the contrary, it is an IMPETUS. It's like this... Company: I want one of thos domain thingies. Internic: That'll be 50 bucks. Company: 50 bucks? I'll take 5 then; put it on my VISA.
issue on any front, and whether it really needs fixing by perhaps us suggesting that subsequent domains be charged on an exponential scale, with proceeds going into the costs of maintaining the worldwide DNS, particularly with respect to user-and-administrator education.
Someone posted some stats on com-priv a week ago that would indicate that Internic is raking in $10 million per annum with the current fees. Are you sure anything needs fixing here?
(NANOG or I*-something-or-other); I'm just wondering if I'm completely out-to-lunch on this one.
You have a good idea there in establishing a funding link between the registrants of domain names and the people who operate the registries and the root nameservers. Of course, most registries are already funded by domain fees like the Internic's $50. The missing link is funding for the root nameservers. Some people also feel that DNS registrations should fund ISOC/IAB/IETF/IANA or some sub or superset thereof. ftp://ietf.cnri.reston.va.us/internet-drafts/draft-ymbk-itld-admin-00.txt covers some of this as does the POISED95 WG
all clearly able to fit (at least for the moment) into <eighty-subdomains>.BigCompany.COM.
I think human beings like flat hierearchies. In management. In GUI menu systems. And now in DNS. Certainly they seem to be more efficient in management and in GUI menu systems and even B+ Trees vs. binary trees. Maybe they really are more efficient in the DNS as well.
If people think this sort of thing is OK, I'll shut up now.
I don't think there really is anything that we can do about it other than to accomodate it. I fully intend to have a couple of dozen domains just for myself and run virtual WWW servers on my home LAN in the next century. But by then that will be the normal state of affairs. The plans to open up the top level domain space by adding 15 or more new international top level domains per year will make the landscape rather more interesting. http://www.kirk.tlhIngan.alt :-) Michael Dillon Voice: +1-604-546-8022 Memra Software Inc. Fax: +1-604-546-3049 http://www.memra.com E-mail: michael@memra.com
On Mon, 22 Apr 1996, Sean Doran wrote:
Looking around a bit, completely unscientifically and without more than eyeballing things, it appears that this practice is continuing, despite the back-pressure of a registration fee levied by the InterNIC.
Sean, you misunderstand. The fee is not a back-pressure, on the contrary, it is an IMPETUS. It's like this...
Company: I want one of thos domain thingies.
Internic: That'll be 50 bucks.
Company: 50 bucks? I'll take 5 then; put it on my VISA.
This also, has the effect, before the fee, the Internic could say one to a company, with now with the fee the are on shaky ground, when combined with their trademark policy. (if you own the trademark, and pay the fee, the probably can't prevent you from registering as many a you want.) It does ohav ethe positive effet of putting the cost of domain speculation, into the non-trivial range.
issue on any front, and whether it really needs fixing by perhaps us suggesting that subsequent domains be charged on an exponential scale, with proceeds going into the costs of maintaining the worldwide DNS, particularly with respect to user-and-administrator education.
Someone posted some stats on com-priv a week ago that would indicate that Internic is raking in $10 million per annum with the current fees. Are you sure anything needs fixing here?
Hm... Internic is raking in "millions" with registration fees, but the root name server operators are getting ZERO. Yet, the real service that you are registering for, is root name service, not the WHOISS database, (root name service is even debateable, since most of the domains are in .com, but the same machines serve .com as ., so its moot.) If I were a root name server opeartor, that could cause major havok, if I decided too, I would be asking Network Solutions, aka internic.NET., where my cut is. If all the root name servers, decided for just 1% of that fee, or the'd drop "internic.net" out of the root zone files, they could probably get away with it. It would also solve the whole issue of resrouces to support root DNS.
(NANOG or I*-something-or-other); I'm just wondering if I'm completely out-to-lunch on this one.
You have a good idea there in establishing a funding link between the registrants of domain names and the people who operate the registries and the root nameservers. Of course, most registries are already funded by domain fees like the Internic's $50. The missing link is funding for the root nameservers. Some people also feel that DNS registrations should fund ISOC/IAB/IETF/IANA or some sub or superset thereof.
ftp://ietf.cnri.reston.va.us/internet-drafts/draft-ymbk-itld-admin-00.txt covers some of this as does the POISED95 WG
There is a very dangers path, because once anyone can buy a top level domain form the root name server opeartors, people will, how long before "@aol.", and "@home." show up? About 2 seconds. You would also need some sort of feedback into the referal rate, but with good postive and negative caching that wouldn't be a huge hurdel.
all clearly able to fit (at least for the moment) into <eighty-subdomains>.BigCompany.COM.
I think human beings like flat hierearchies. In management. In GUI menu systems. And now in DNS. Certainly they seem to be more efficient in management and in GUI menu systems and even B+ Trees vs. binary trees. Maybe they really are more efficient in the DNS as well.
Flat space will fail when 50% of the humans on this planet have their own domain. (You think Internet growth is bad now, just wait until a Internet terminal can be had for $100, the $500 ones are only a year or so around the corner, and based on historical trends, the rgowth rate increases every time the the price point drops. (Which could happen, 2.5 billion people on the Net in 6 years...) (Assuming population iof the Net is doubling every 9 months, like bandwidth, and addressable hosts, and domains).
If people think this sort of thing is OK, I'll shut up now.
I don't think there really is anything that we can do about it other than to accomodate it. I fully intend to have a couple of dozen domains just for myself and run virtual WWW servers on my home LAN in the next century. But by then that will be the normal state of affairs.
Luckyly, all the good names will be taken well before this happens, and there just isn;t that much motiviation, for people to get 6 different domains and pay for them when they look like fsdhwue47.sdfj, etc. We are quite limited by the fact that the DNS protocols seem to limit us to [A-Z][a-z][1-0]-_. We would be really bad off if the DNS implementers had used UNICODE.
The plans to open up the top level domain space by adding 15 or more new international top level domains per year will make the landscape rather more interesting. http://www.kirk.tlhIngan.alt :-)
Who has proposed this? I don't see it happening. A. it won't scale B. it looks ugly Thus violating the two key reasons the DNS system is so popular today. I think more TLD will be opened, but it will be for a couple of reasons, first off some re-writes DNS so that the clunky concept of a "root" name server is dead, then the commerical giants buy Microtsoft, or your favorite software vendor to implement the new protocol, and with their millions of users for everyone elses network to work they would need to upgrade to the new software to support "aol.". The social reasons with regard to Paul Vixie's "dot envy", are quite clear. I expect there to be a good bit of money made in the "domain length shortening business". I can see that ad now: Domains R Us, get your domain shortened here. Be the envy of major corporations and national governments, meet exiciting women and make money in your new home based business, with Domains R Us. sigh. -- Jeremy Porter, Freeside Communications, Inc. jerry@fc.net PO BOX 80315 Austin, Tx 78708 | 1-800-968-8750 | 512-339-6094 http://www.fc.net
Hm... Internic is raking in "millions" with registration fees, but the root name server operators are getting ZERO.
In dollars, yes. I run a root name server because even with all of its warts, I rather _like_ the Internet and I want it to continue to exist. What I get from running a root name server, other than a chance to hack BIND to work better on root name servers, is the satisfaction and safety of knowing that at least one root name server is run _correctly_. Which is not to say that it's the only correctly run server, but you get the point.
Yet, the real service that you are registering for, is root name service, not the WHOISS database, (root name service is even debateable, since most of the domains are in .com, but the same machines serve .com as ., so its moot.)
If I were a root name server opeartor, that could cause major havok, if I decided too, I would be asking Network Solutions, aka internic.NET., where my cut is.
If all the root name servers, decided for just 1% of that fee, or the'd drop "internic.net" out of the root zone files, they could probably get away with it.
It would also solve the whole issue of resrouces to support root DNS.
InterNIC is buying a pair of machines to replace my root name server; ISC's budget could not sustain another 128MB hit in the same fiscal year. InterNIC's offer of equipment support went out to all the root servers, not just mine. I don't think the root server operators have any gripes. We generally do what we do out of, um, well, love for the net in some form or another. It's nice that InterNIC is buying us hardware but we'd find a way to do it even without.
On Tue, 23 Apr 1996, Jeremy Porter wrote:
the price point drops. (Which could happen, 2.5 billion people on the Net in 6 years...)
I think the number of people with a phone at home is less than that. Don't forget that most of the world's population lives in China and underdevelopped countries.
We are quite limited by the fact that the DNS protocols seem to limit us to [A-Z][a-z][1-0]-_. We would be really bad off if the DNS implementers had used UNICODE.
Assuming 9 letters in a domain name this would be 37 ^ 9 which is 1.29e14. I think that leaves a fair amount of room for choice considering that some domains will be more or less than 9 characters and there are currently over 100 top level domains with more planned.
The plans to open up the top level domain space by adding 15 or more new international top level domains per year will make the landscape rather more interesting. http://www.kirk.tlhIngan.alt :-)
Who has proposed this? I don't see it happening.
ftp://ietf.cnri.reston.va.us/internet-drafts/draft-ymbk-itld-admin-00.txt
A. it won't scale
Then push for the "million domains" experiment to go forward sooner rather than later. This pushes the scaling one level deeper in the tree so if scale problems do develop we can simply open another subtree. Only at the top level do we need to worry about scale, but there should be no problem with 2 or 3 times the current number of TLD's.
B. it looks ugly
Maybe you misunderstand. tlhIngan is the Klingon language word for "klingon" thus the English equivalent would be http://www.kirk.klingon.alt This is no uglier than http://www.apple.austin.tx.us
I think more TLD will be opened, but it will be for a couple of reasons, first off some re-writes DNS so that the clunky concept of a "root" name server is dead,
I see. Well then, you will probably be interested in the new Internet technology for transporting live humans across the net that was being discussed on the inet-access list this past month. There is an archive of inet-access available at ftp.earth.com if you would like see how Internet Service Providers are currently handling this. Obsmiley ;-) any directory service needs a place to start the search. That's what the roots are for and I don't think anything could possibly replace them except a different set of roots or a different kind of root. The concept is inherent in the system. Michael Dillon Voice: +1-604-546-8022 Memra Software Inc. Fax: +1-604-546-3049 http://www.memra.com E-mail: michael@memra.com
On Mon, 22 Apr 1996, Michael Dillon wrote:
Someone posted some stats on com-priv a week ago that would indicate that Internic is raking in $10 million per annum with the current fees. Are you sure anything needs fixing here?
I would gladly take a $1 million chunk of that to run a root server. $1 million would pay my salary, as well as a few others, in addition to paying for a DS3 and a machine to handle all the requests. And in future years, I could give myself a huge raise and pay for loads of upgrades to my DNS root server. /cah
participants (5)
-
Craig A. Huegen -
Jeremy Porter -
Michael Dillon -
Paul A Vixie -
Sean Doran