[dns-operations] stealth slaving the root zone
warren at kumari.net
Wed May 16 21:56:43 UTC 2012
On May 16, 2012, at 3:46 PM, Suzanne Woolf wrote:
> On Wed, May 16, 2012 at 08:11:45PM +0100, Jim Reid wrote:
>> On 16 May 2012, at 18:07, Joseph S D Yao wrote:
>>> On Wed, May 16, 2012 at 09:31:29AM +0100, Jim Reid wrote:
>>>> easily serve a zone containing a few million names. [And FWIW I very
>>>> much doubt the vanity TLD madness will continue long enough for the
>>>> root zone grow to anywhere like that size: maybe a few thousand new
>>>> TLDs at most.] There are of course non-trivial problems making every
>>> Jim, if vanity has started to have bounds, I've not been aware of it.
>> True. However the $186k entrance fee and other unavoidable costs will
>> set some limits.
> The thing to watch for is pressure to lower that.
> The current commitment on ICANN's part, in deference to the findings
> from the discussions of root scaling in SSAC, RSSAC, and elsewhere,
> limits growth in the root zone to no more than 1000 new delegations in
> a year. DNS experts told ICANN quite clearly that slow change could be
> quite large before it should cause concern, and a growth factor of 10
> against the current zone is not large. They also warned that abrupt
> change could be more of a concern, depending on the magnitude.
> Of the 2100 or so applications currently received
> evaluation will take some time and not all requested names will be
> delegated, with no clear way to predict how many actual delegations
> will result or when (the evaluation phase determines that, according
> to the guidelines in the New gTLD Applicants' Guidebook as published).
> In the short-to-medium term, I'm not worried about this; a factor of
> 10 growth in the number of delegations, with no corresponding change
> in update rates, seems unlikely to cause operational stress on any
> timescale that wouldn't also allow appropriate steps to be taken to
> mitigate it-- and a number of parties, including ICANN itself and an
> assortment of governments via the GAC, are keeping a close eye on the
> specifics of how this plays out in practice.
> In the long term, I expect pressure to lower both the entry fee and
> the technical requirements, just because that's the way of the
And I (personally) don't foresee the actual *number* of delegations causing scaling issues (RAM is cheap, records are small, etc), but rather the increase in traffic / query rate to the root caused by recursives not having TLDs cached (currently a busy, well resourced recursive has basically all (interesting) TLD delegations cached basically all the time).
Yes, it is an effect of more delegations, but a: second order and b: only if the shiny new delegations actually get *used*.
I *really really really* don't foresee this becoming a scaling concern, but even if it somehow did, it still wouldn't call for big-iron boxes, but rather an increase in anycast capacity.
Much more concerning (and annoying) is simply the amount of bogus / garbage queries that root gets -- if folk want a windmill to tilt at...
> Much depends on how well new gTLDs do in the next few years.
: and would be handled by the current capacity planning.
> (NB: I serve as the RSSAC liaison to the ICANN Board of Directors, but
> I don't speak for ICANN and have no non-public knowledge on the
> current program status.)
And if we're doing full disclosure: I serve on the ICANN SSAC and we had *many* *long* discussions on this topic… I seriously considered taking up cutting myself again….
> dns-operations mailing list
> dns-operations at lists.dns-oarc.net
> dns-jobs mailing list
More information about the dns-operations