new.net: yet another dns namespace overlay play

Mailman List Mirror (Read Only)
1

-----BEGIN PGP SIGNED MESSAGE-----

So the short answer minus the marketing-speak is that new.net is offering
domains in new TLDs without having them in *.root-servers.net. The idea
isn't new; people have been doing this for years with limited success.

As domains in the new TLDs are obviously only useful if people can reach
them, new.net is offering a variety of ways to be able to resolve them.

Anyone in control of a recursive nameserver can:

- - Slave a root zone with the new TLDs appended.

- - Add stubs for the new TLDs pointing at udns[12].newdotnet.net. (I'm
  pleased with the UltraDNS infrastructure, which is serving similarly here
  to the gtld-servers.net but for new.net's TLDs. For those using the
  stub-zone hack, these are the only servers they ever see.)

- - Replace the root cache with one pointing to ns[123].newdotnet.net, which
  will delegate as appropriate to the various TLD servers. (There are
  obviously too few of these available with not enough diversity, but more
  will be added soon.)

The "launch partners" are listed as Earthlink, Excite@Home, NetZero. These
shortly will be using one of the above to allow their users to reach the new
TLDs. Given that the whole goal is for the domains to be resolvable, expect
to see more.

But these options aren't available to the average end user. Rather than
setting up a few centralized recursive nameservers and trying to get every
desktop in the world to send their traffic to them, new.net has elected to
also make the domains reachable with ".new.net" appended. This allows for
proper local caching and search-path related tricks. For all desktop users,
the option is available to just add new.net to the search path.

One of the functions of the Windows client offered on the site is a resolver
plugin that appends ".new.net" on to some queries. It tries to cut down on
the number of failed DNS requests that just would be generated by just
adding new.net to the search path. It will also act as a sub-proxy for
proxy users, appending ".new.net" on to some HTTP proxy requests. I'm not
sure whats going on with a patent application on this.

This is obviously less clean than just having ICANN list new TLDs. But
given the ongoing frustration with ICANN, there is significant interest in a
different approach, and new.net is taking one.

I'm pretty sure that I didn't want it to come to this, and I'm not entirely
convinced that anyone should be doing it. But personal reservations aside,
its happening. And I intend to see that its done as well as possible.

                             Aaron Hopkins
                             Systems Engineer, idealab!
                             Acting VP of Engineering, new.net

2

So the short answer minus the marketing-speak is that new.net is
offering domains in new TLDs without having them in
*.root-servers.net. The idea isn't new; people have been doing this
for years with limited success.
[...]
The "launch partners" are listed as Earthlink, Excite@Home, NetZero.

NetZero is an idealab! Capital Partner, so I'm not terribly surprised
there. EarthLink and Excite@Home are big and well-respected names,
and surely do carry _some_ weight as to setting precedents. But these
organizations do not the internet make, and the success of new.net's
new TLD's is dependent upon widespread adaptation. What incentive --
operational, financial, and otherwise -- is there for other providers
to follow suit?

What differentiates new.net from other people who might want to
utilize similar tactics for TLD's of their own? And what happens when
there's TLD overlap between you and other alternate registrars?
Clearly some sort of authority structure, be it ICANN or something
idealab!-created to "compete" with ICANN, is needed.

I'm sure Dr. Joe Baptista and Roeland Meyer would like it if the
global Internet could resolve .god domains, and certainly Jim
Phlemming has a TLD up his sleeves to be used in conjunction with IPv8
rollout... :wink:

This is obviously less clean than just having ICANN list new TLDs.
But given the ongoing frustration with ICANN, there is significant
interest in a different approach, and new.net is taking one.

I agree, ICANN (or is ICan't more appropriate?) is a very broken
organization, and I wholeheartedly appreciate the efforts made by you
and others to circumvent them through innovation such as this.

Still, they're the folks most(?) Internet operators recognize as being
in charge, and until that changes, I question the usefulness of
alternate root servers.

-adam

3

I'm pretty sure that I didn't want it to come to this, and I'm
not entirely
convinced that anyone should be doing it. But personal
reservations aside,
its happening. And I intend to see that its done as well as possible.

Then why did you and David ignore my plea to cooperate with the extant TLD
managers, with whom the new.net TLDs now collide? You could have launched
the new.net TLDs with a bunch of in-place registrants already hosting sites
under the TLDs you have collided with. You could have built a shared
registration system that could have encompassed all the non-ICANN TLDs, and
helped create something that would have really given serious challenge to
ICANN. But instead, you chose to ignore me, and the others. Now we have a
mess on our hands, for example, who is really the registrant of
warren.family, the one who has held the Pacificroot warren.family for 4
years, or the one who just got warren.family from new.net on Monday?

What will new.net say to their customer when one of Pacficroot's registrants
sends a C&D to a new.net registrant who has collided with their SLD? Is
new.net's indemnity clause sufficient to protect you from liability to your
customers, after you sold them something which subsequently got the new.net
customer into a lawsuit? Maybe you want to run that one by David Hernand,
see what he thinks.

What a can of worms you have opened. I wish you had listened to me. And
no, I'm not interested in a job at new.net anymore, in case you were
wondering.

4

But if they had done the net-friendly thing (created a
partnership/coalition/whatever with existing alternate roots), that
would have.. been the net-friendly thing.

Sure they could have done this, instantly strengthened their
position, and maybe created a serious enough force that ICANN may
have felt it and reacted intelligently. As it is they're just
another alternative root, albeit with more $$ than most.

Will they survive? maybe. Given the fate of similar idealab!
creations it's certainly not a statistical probability.

...

Of course, who are we to challenge new.net, with their
patent-pending technology for appending ".new.net." to hostnames,
leading partnerships with exciting companies like earthlink!
and they exist to a whopping 16 million users, are easily accessed
by everyone else willing to fiddle with their resolver (well, not
mail, sorry we need to invent a sendmail plugin for that one).

Besides, nothing that ever came out of palo alto and was spun
up by a "think tank" with an exclamation mark appended to their name
ever went wrong, right?

Last I checked guesstimates of internet users globally was something
like 400 million people, ~30% were in north america,
~25% in europe, ~20% in asia/oceania, ~10% in south america.
Maybe these figures seem high, but remember that only a considerable
subset will be regular [ab]users.

16 million, probably an aggressive figure to start with, any way you
look at it, nothing but a drop in the bucket.

aaron: when you guys end up on fuckedcompany, can I get a deal on
some of that nice hardware you [probably] have over there? Maybe a
terabyte disk farm, or a highend server (I'll pay shipping).

cheers,
- wolfie.