steve, all. On Tue, Jul 05, 2005 at 10:01:22AM -0700, Steve Gibbard wrote:

problem. Right now, if you're an end user doing your DNS lookups via the ICANN root, you can get to just about everything. If you're something that end users want to connect to, using an ICANN-recognized domain will mean almost everybody can get to you, while an "alternative" TLD would mean only a tiny fraction of the Internet would be able to get to you. So, if you're a content provider, why would you use anything other than a real ICANN-recognized domain? And, if the content providers aren't using real domain names, why would an end user care about whether they can get to the TLDs that nobody is using?

s/ICANN root/real Internet/ s/"alternative" TLD/IPv6/

The exceptions to this that I see would be either when somebody comes out with something that is so much better that it's useful in spite of a lack of an installed userbase (Skype may be doing this to phone calls), or when something is rolled out to a large enough self-contained user community that the lack of ability to communicate outside that region won't be a significant barrier. [...] But still, anybody wanting to communicate outside that region or userbase would probably find they were much happier using addresses that met global standards.

all of this applies directly to lack of IPv6 adoption, again.

So anyhow, that's a long way of saying that, just as this hasn't gone anywhere any of the many other times it's been raised over the last several years, it's unlikely to go anywhere, or cause problems, this time.

so does this. IPv6: unlikely to go anywhere or cause problems. good to know. funny. all threads eventually merge. (and then someone mentions the nazis and they end. i think meta-mentions like this explicitly don't count so we may have to suffer through this thread for a while longer). t. -- _____________________________________________________________________ todd underwood director of operations & security renesys - interdomain intelligence todd@renesys.com www.renesys.com

On Tue, Jul 05, 2005 at 10:01:22AM -0700, Steve Gibbard wrote:

On Tue, 5 Jul 2005, Jay R. Ashworth wrote:

...

But Steve appeared to be suggesting that there was no reasonable way to *avoid* problems -- and that's clearly not the case. If I misinterpreted Steve, no doubt he'll correct me. But there are two fairly prominent,

I don't think that was what I said. What I was attempting to say is that the issue of alternate roots probably isn't something that's worth worrying about. I see no reason why they'll catch on, other than perhaps in limited cases where they'll work ok.

Catch on in the consumer sense? No, probably not -- though the question is "will IAP's switch their resolver servers to an alt-root".... which leads directly to:

In the general case, with alternate roots, there's a chicken and egg problem. Right now, if you're an end user doing your DNS lookups via the ICANN root, you can get to just about everything. If you're something that end users want to connect to, using an ICANN-recognized domain will mean almost everybody can get to you, while an "alternative" TLD would mean only a tiny fraction of the Internet would be able to get to you. So, if you're a content provider, why would you use anything other than a real ICANN-recognized domain? And, if the content providers aren't using real domain names, why would an end user care about whether they can get to the TLDs that nobody is using?

Two points: 1) this speaks to the same issue as my comments the other day on the IPv6 killer app, though it's admittedly even harder to posit a site which would do this. 2) Based on the events earlier in the week, I believe that's a "US Department of Commerce" approved TLD... which changes the game a little bit.

This is the same phonomenon we saw ten years ago, as the various "online services," GENIE, Prodigy, MCIMail, Compuserve, AOL, etc. either interconnected their e-mail systems with the Internet or faded away and died. As the Internet got more and more critical mass, there was less and less incentive to be using something else. It's been a long time since I've seen a business card with several different, incompatible, e-mail addresses printed on it, and that's because something simpler worked, not because people screamed loudly about the falling sky.

Certainly. But there weren't geopolitical implications there, merely commercial ones. I think the stakes may be a bit higher here, particularly in the case we were using as an example: China.

The exceptions to this that I see would be either when somebody comes out with something that is so much better that it's useful in spite of a lack of an installed userbase (Skype may be doing this to phone calls),

Yup. Killer apps are great. Hard to predict; *really* hard to invent.

or when something is rolled out to a large enough self-contained user community that the lack of ability to communicate outside that region won't be a significant barrier. If a few large countries were to roll out alternate root zones nation-wide, in such a way that they worked well for domestic communication, but couldn't be used for international stuff, *maybe* that would be good enough to catch on. But still, anybody wanting to communicate outside that region or userbase would probably find they were much happier using addresses that met global standards.

But again, you're positing that someone would create a root zone that *purposefully* conflicted with the current one, which doesn't seem supported by history, much less common sense. Am I wrong that you mean that?

So anyhow, that's a long way of saying that, just as this hasn't gone anywhere any of the many other times it's been raised over the last several years, it's unlikely to go anywhere, or cause problems, this time.

Maybe. China's *really* big. America's *really* unpopular, in some places. Cheers, -- jra -- Jay R. Ashworth jra@baylink.com Designer Baylink RFC 2100 Ashworth & Associates The Things I Think '87 e24 St Petersburg FL USA http://baylink.pitas.com +1 727 647 1274 If you can read this... thank a system administrator. Or two. --me