IBM on multiple roots
Richard J. Sexton
From: nickpa...@worldnet.att.net
To: bwg-n-...@fibertron.com
Subject: Re: [bwg-n-friends] Multiple root idea is spreading?
Date: Fri, 17 Sep 1999 14:04:39 +0000
X-Authenticated-Sender: npat...@att.net
Sender: owner-bwg...@spike.fibertron.com
Reply-To: bwg-n-...@fibertron.com
Hi Karl,
Yeah I wanted to see what we would say, as he's
a relatively technical guy (at least I think so).
It wasn't very convincing though, was it?
N
<>
<>
<> > COMPUTERGRAM INTERNATIONAL: SEPTEMBER 17 1999
<> > SECTION: INTERNET
<> > + IBM Set to Donate $100,000 to ICANN; Failure "Not An Option"
<> > By Nick Patience
<>
<> > IBM Corp's VP internet technology John Patrick...
<>
<> > He says that multiple roots would only be feasible if the
<> > internet was being designed now. But with tens of millions of
<> > uses and hosts, it's just not possible he says, because the
<> > internet users around the world will not accept even a tiny
<> > break in the system. "There is no choice," he says, "you must
<> > have one person at the top of the root."
<>
<> Looks like the multiple root idea is spreading.
<>
<> It's kind of funny to read Patricks' statement because it is technobabble
<> that bespeaks a fundamental non-comprehension of what he is talking about.
<>
<> "tiny break in the system"? You mean like sending a relative URL? Or
<> sending e-mail that can't ever be received, like some members of the ICANN
<> board:
<>
<> Mail Queue
<> --Q-ID-- --Size-- -----Q-Time----- ------------Sender/Recipient------------
<> MAA04683* 3119 Sun Sep 12 12:06 <ka...@CaveBear.com>
<> (Deferred: Connection timed out with pegase.bull.fr.)
<> <geraldine....@bull.fr>
<>
<> The system breaks in tiny ways all the time.
<>
<> And if is so concerned about "tiny breaks" then he must be ready to leap
<> out of his chair and request, nay, demand a unified regulatory body
<> sitting on top of the routing of packets.
<>
<> (Packets today are routed end-to-end only because each individual ISP has
<> a notion that it is in its own best interest to make sure that it makes
<> routing choices that promote connectivity rather than damage it.)
<>
<>
<> > We wondered what a governmental-run body might look like if
<> > ICANN should fail and governments step in. Patrick says it's
<> > impossible to tell because such a body has never existed. He
<> > says the internet runs the same way all over the world to the
<> > same standards, which is not true of the telephone, railroad or
<> > any other system that is regularly compared to the internet.
<>
<> You mean satellite frequencies aren't coordinated around the world? And
<> I'm wrong when I think that I hope airports and international airplanes
<> agree on communications, traffic corridors, and landing/takeoff patterns?
<>
<> And I guess I'm wrong that containers (those big boxes on ships, trucks,
<> and trains) are not standardized sizes.
<>
<> I guess he's never looked at the catalog of standards from some group like
<> the International Standards Organization.
<>
<> --karl--
<>
<>
--
Richard Sexton | ric...@tangled.web | http://dns.vrx.net/tech/rootzone
http://killifish.vrx.net http://www.mbz.org http://lists.aquaria.net
Bannockburn, Ontario, Canada, 70 & 72 280SE, 83 300SD +1 (613) 473-1719
Richard J. Sexton
+ IBM Set to Donate $100,000 to ICANN; Failure "Not An Option"
By Nick Patience
IBM Corp's VP internet technology John Patrick has finally got
his way within the company and has apparently secured a
$100,000 unconditional donation to the financially-beleaguered
Internet Corporation for Assigned Names and Numbers (ICANN).
However, at press time yesterday the check had not been written
or sent to ICANN, but it is expected within a day or so, having
been cleared by IBM's senior management, we understand. Last
month ICANN announced one-year loans totaling $850,000 from MCI
Worldcom Inc, Cisco Systems Inc and 3Com Corp, with $500,000 of
that coming from MCI.
Observers were puzzled at the time as to where IBM's donation
was, given that Patrick and MCI's Vint Cerf were the main
protagonists in the desperate fund-raising to bolster ICANN's
bank balance, which was heavily in the red by late June. Around
that time Patrick approached a group of the top venture
capitalists in the US, including John Doerr and Ann Winblad of
Hummer Winblad Venture Partners, only to be turned down.
The emergency fundraising effort was not publicly announced
until the money was more or less already in the bank and the
approach to VCs was not announced at all. Patrick acknowledges
that the fund-raising drive should have been more part of the
public debate on administration of the internet: "perhaps it
could have been handled differently," he says. Many in the
internet community feel that by not announcing it and then
accepting large loans from corporations, ICANN is in danger of
being seen as being in the pocket of corporate America, but it
regularly denies such charges.
Patrick believes the failure of ICANN "is not an option."
However, he also insists that its scope is "very narrow." He
says "it's not trying to boil the ocean" and solve all the
problems associated with running the internet's name and
numbering systems, but it must be seen to be representing all
the various interests that are involved in the internet. Asked
whether he feels the administrative procedural and technical
remit given to ICANN could in fact be very broad and
far-reaching, Patrick disagrees. He says it is a simple set of
initiatives, including deciding how many registrars and
registries there should be and some sort of fee structure to
support ICANN. He also believes that there can only be one root
from which all name servers take their domain name-IP number
mappings.
He says that multiple roots would only be feasible if the
internet was being designed now. But with tens of millions of
uses and hosts, it's just not possible he says, because the
internet users around the world will not accept even a tiny
break in the system. "There is no choice," he says, "you must
have one person at the top of the root."
We wondered what a governmental-run body might look like if
ICANN should fail and governments step in. Patrick says it's
impossible to tell because such a body has never existed. He
says the internet runs the same way all over the world to the
same standards, which is not true of the telephone, railroad or
any other system that is regularly compared to the internet.
At its meeting last month in Santiago, ICANN revealed that
after using the $850,000 in loans, it expected to have just
$37,000 cash at the end of August and already had at that time
unpaid invoices and bills totaling around $250,000 generated in
July and August alone. It is hard to see how ICANN will pay
those loans back within one year, if at all.
Greg Skinner
>+ IBM Set to Donate $100,000 to ICANN; Failure "Not An Option"
>By Nick Patience
>[Patrick] says that multiple roots would only be feasible if the
>internet was being designed now. But with tens of millions of
>uses and hosts, it's just not possible he says, because the
>internet users around the world will not accept even a tiny
>break in the system. "There is no choice," he says, "you must
>have one person at the top of the root."
Hmmm ...
I wonder if the multiple root world might look something like what we
had in the early to mid 1980s, where there were lots of "networks",
each with their own addressing syntaxes and rules. To send email from
one "network" to another, you had to put part of the address in
quotes, or use a % sign, and forward it through a relay site. (For
example, someone wanting to send mail from an ARPAnet site to a CSnet
site might use the address
my-csnet-contact%their-csnet...@csnet-relay.arpa.) I can
imagine that for many, if not most protocols, modifications can be
made that allow for the proxying or gatewaying of certain functions;
such functionality already exists in several protocols already. This
doesn't necessarily mean we would see the end of e-commerce, although
we might see minor flaws in some types of e-commerce as they modify
their protocols to compensate for proxying and gatewaying functions.
Likewise, other non e-commerce services could adapt.
>We wondered what a governmental-run body might look like if
>ICANN should fail and governments step in. Patrick says it's
>impossible to tell because such a body has never existed. He
>says the internet runs the same way all over the world to the
>same standards, which is not true of the telephone, railroad or
>any other system that is regularly compared to the internet.
Perhaps it might look like the FCC. Historically, the FCC was created
to resolve the type of contentious issues facing Internet name and
address policy today.
--gregbo
gds at best.com
Richard J. Sexton
It's happening already, Greg. Anybody can send mail to
orange%dns....@vrx.net.
The irony that Dave Crocker invented this use of % is not
lost on me.
Maybe we should refer to it as dcrocking the addres.
Jim Kingdon
> had in the early to mid 1980s, where there were lots of "networks",
> each with their own addressing syntaxes and rules.
Yeah, that strikes me as an interesting analogy.
There are a couple of things to point out about it:
(1) The tendency is to evolve away from fragmentation. Connectivity
is its own reward. This is why I don't think we need a
heavy-handed bureaucracy to enforce a single DNS root - as much
consistency as the world needs will happen more or less naturally.
(2) If anyone tries to own one of the networks, it is doomed (and the
more control freakery going on, the more so). Perhaps the best
example is fidonet, in which the political struggles were
legendary and fidonet survived by not letting them get away with
it (too much). Or other examples might be proprietary protocols
like DECnet and the IBM send/receive protocols. Or the OSI
protocols - this is a great example of how you can screw up by
trying to say "there is one true way, it is very important that
everyone do things this way".
> Perhaps it might look like the FCC. Historically, the FCC was created
> to resolve the type of contentious issues facing Internet name and
> address policy today.
Gawd help us, I hope not. Way too many lawyers, lobbyists, and
agendas, and not enough making the customers happy and moving bits and
introducing great new products. The current fight over whether
satellites {can,must} carry local TV stations is a case in point.
Greg Skinner
>orange%dns....@vrx.net.
I know this. However, this "backdoor" knowledge of how to use email
is not widely documented.
The question becomes what the common Internet user will do when faced
with this type of problem.
--gregbo
gds at best.com
Greg Skinner
Jim Kingdon <kin...@panix.com> wrote:
>Greg Skinner wrote:
>> Perhaps it might look like the FCC. Historically, the FCC was created
>> to resolve the type of contentious issues facing Internet name and
>> address policy today.
>Gawd help us, I hope not. Way too many lawyers, lobbyists, and
>agendas, and not enough making the customers happy and moving bits and
>introducing great new products. The current fight over whether
>satellites {can,must} carry local TV stations is a case in point.
Agreed. However, the DNS mess has brought about FCC-type
bureaucracy.
--gregbo
gds at best.com
Greg Skinner
>agendas, and not enough making the customers happy and moving bits and
>introducing great new products. The current fight over whether
>satellites {can,must} carry local TV stations is a case in point.
You might be interested in
http://www.wired.com/news/news/email/tip/politics/story/21985.html
Note comment on peering arrangements.
Jim Kingdon
> http://www.wired.com/news/news/email/tip/politics/story/21985.html
> Note comment on peering arrangements.
Yeah, that is a prime example of why we want the government as far
away from the net as possible (which I assume was your point).
If I understood the comment correctly, they want to make the internet
some kind of metered thing. This might sound sort of reasonable until
you realize that it makes no sense in the internet context.
Telephones are based on a model where you know which end of the call
pays (usually the caller, or recipient for 800 numbers and such). At
the level of an internet router, there isn't even such thing as a
"call" or anything equivalent.
In a sense, there is a real problem (companies like GTE either don't
have the incentives to make their internet networks reliable, or think
they don't - I'm suspecting the latter). But letting individual
peering/transit points/contracts experiment with different pricing
models is going to work better than having some idiot economist in
Washington, DC concoct some crazy scheme which may or may not have
anything to do with actually giving the customers what they want.
Oops, I'm ranting. You guys probably are already with me on this....
Jim Kingdon
Yes, ICANN is too bureaucratic. And I hope we can do something about
that (by reforming ICANN, by reforming CORE, by alternate DNS roots,
probably some of each and perhaps a few others I forget to mention).
But being the pessimist I am, I am quite willing to say "it could be
worse" :-).
Speaking of alternate DNS, anyone written a rebuttal to the IAB's
paper on how there can be only one root? I don't mean a
point-by-point rebuttal, that would be tedious, I just mean a white
paper giving the case for alternate DNS. I can think of technical
details (the IAB paper seems to be worried that data from one DNS
hierarchy will leak into servers from another - which was more of a
problem with older DNS implementations which need to change for other
reasons anyway, as I understand it). Plus policy details
(e.g. different roots are likely to agree on many delegations), &c.
I mean, parts of the IAB's paper are correct - about how a shared zone
needs some kind of "registry" function - but they kind of wave their
arms when they jump from that to "there can only be one root".
Richard J. Sexton
Jim Kingdon <kin...@panix.com> wrote:
>> Agreed. However, the DNS mess has brought about FCC-type bureaucracy.
>
>Yes, ICANN is too bureaucratic. And I hope we can do something about
>that (by reforming ICANN, by reforming CORE, by alternate DNS roots,
>probably some of each and perhaps a few others I forget to mention).
>
>But being the pessimist I am, I am quite willing to say "it could be
>worse" :-).
Not really. What do you know about OSI ?
>Speaking of alternate DNS, anyone written a rebuttal to the IAB's
>paper on how there can be only one root? I don't mean a
>point-by-point rebuttal, that would be tedious, I just mean a white
>paper giving the case for alternate DNS. I can think of technical
>details (the IAB paper seems to be worried that data from one DNS
>hierarchy will leak into servers from another - which was more of a
>problem with older DNS implementations which need to change for other
>reasons anyway, as I understand it). Plus policy details
>(e.g. different roots are likely to agree on many delegations), &c.
>
>I mean, parts of the IAB's paper are correct - about how a shared zone
>needs some kind of "registry" function - but they kind of wave their
>arms when they jump from that to "there can only be one root".
karl@cavebear (Karl Aurbach) is leading that effort. People
should contact him.
Richard J. Sexton
Jim Kingdon <kin...@panix.com> wrote:
>Oops, I'm ranting. You guys probably are already with me on this....
There's lots to rant about. Here's my last one.
>I thought it was to move away from a monopoly registry model to a shared
>one so that no registry-registrar could obtain any lock-in advantage over
>competing registrars. Isn't this what the testbeds are complaining about
>(that no registrar is able to compete effectively against another
>registrar who also owns the registry?) Otherwise there's no incentive to
>generate technology to enable portability.
So all this crap going on now is for my benefit ?
Was I asked ?
Can I pay twice as much to have it the way it was before
ICANN was born ?
There's one database. One zone file. Ona database manager.
Get over it. I don't really care who sells me the domain,
either the registry works or it doesn't for some definition
of works. If I don't like it I can go somewhere else. There
is no hard evidence anything is broken or in danger of
being broken. Just a lot of FUD.
I don't feel locked in. yeah, the price might go up, but the
price might go down, too. If we presume change to be inevitable
which forms the basis of the "lock in" argument, there is nothing
to prevent change to whatever we claim is the solution to this
problem! And in fact, this is what'a happened.
In a post ICANN world, in the name of protecting the
onsumer, domains don't cost any less, all the software
on the net that used to email a domreg template to
hostm...@internic.net is now broken, I have to use
a credit card at time of registration to lease a domain
I used to be able to own, that has in increasing amount
of oneraous and blatently wrong trademark regulation
picking away at it.
This is protected ?
I have to hand it to them. They havn't actually stopped the
net from working yet, but it's pretty fucked up.
C'mon Mr. Peabody, set ther wayback machine to 1996.
Richard J. Sexton
>Jim Kingdon <kin...@panix.com> wrote:
>>Oops, I'm ranting. You guys probably are already with me on this....
>
>There's lots to rant about. Here's my last one.
Here's another
>Has competition beein increased in the registrar front end? Not really.
>NSI was allowing hundreds of companies (call them "registrars" if you like)
>to register names in com net and org prior to the SRS system. It also allowed
>individual end users to do it for themselves. The difference between now and
>then? NTIA has fixed the wholesale price, and has interposed between the end
>user and the registry a single, proprietary Shared Registration System (SRS)
>software. In other words, we have regulated NSI's price. We could have done
>that with or without separating the registry and registrar functions.
>
>Ultimately, ICANN's goal is to make it impossible for the end user to access
>the registry directly and make their own registration. They want to force you
>to use an intermediary.
Or in other words, ICANN is saying "you can't buy things from that company
any more. You have to go throuhg one of our exlusive distributors that
we've sold a franachise to. Can you say MLM boys and girls ?
So instead of complaining about a "government contract granted monopoly
for NSI to sell domains" we can now talk about "a government awarded monopoly
on selling of domain franchises to ICANN".
In other words we've traded one monopoly for a bigger version of that
monopoly. Of course when we said 3 or 4 years ago "the only way to
solve this monopoly is to make lots and lots of them so they're so
dilute no large problem with one of them affects the Interent in toto
very much" we were told that manopolies were simply not acceptible
thank you and good bye.
>The other point of serious confusion that you reveal concerns portability.
>There is no "portability" of the domain name itself. What "portability" means
>in the SRS system is simply that you can change the person who bills you for
>the domain name. IT DOES NOT CHANGE THE BUSINESS THAT PROVIDES THE DOMAIN
>NAME.
I said about 2 years ago that the easiest way to get from where there
were now (1997) to where they want to be (1999) would be to add a checkbox
to the domreg form so you could select the CORE registrar you wanted
NSI to send $20 to. It could have been implemented overnight and nothing
would be different.