[dns-operations] 答复: L-Root IPv6 Address renumbering
stlagerh at microsoft.com
Mon Mar 14 02:52:06 UTC 2016
I find the renumbering schedule rather aggressive and I disagree that DNS resolver operators are treated as children should the timeframe be extended.
The operation of the root zone is critical, there is a reason why there are 13 operators. And no recursive software would try all 11+13 existing addresses before giving up and returning SERVFAIL, at best I would expect 3 or perhaps 4 tries before that happens.
Has the timeline been discussed in the Root Server System Advisory Committee where you are a member? Are all the members in agreement that best operational practice is to shut down an old address 6 months after renumbering?
It does not seem to be to be a large operational burden to keep the old address longer, what is the driving force behind this aggressive timeline?
> -----Original Message-----
> From: dns-operations [mailto:dns-operations-bounces at dns-oarc.net] On
> Behalf Of Terry Manderson
> Sent: Sunday, March 13, 2016 6:59 PM
> To: Shane Kerr <shane at time-travellers.org>; Davey Song(宋林健)
> <ljsong at biigroup.cn>
> Cc: dns-operations at dns-oarc.net
> Subject: Re: [dns-operations] 答复: L-Root IPv6 Address renumbering
> On 11/03/2016, 7:41 PM, "dns-operations on behalf of Shane Kerr"
> <dns-operations-bounces at dns-oarc.net on behalf of shane at time-
> travellers.org> wrote:
> >The root server system is highly redundant, and we are only talking
> >about 1 of 11 IPv6 addresses for these servers, or 1 of 24 addresses
> >overall. I doubt a single resolver would fail to resolve if any of the
> >root servers turned off its IPv6 today. (Probably true for IPv4 too,
> >although A is probably still "special" in some setups.)
> >old address doesn't provide any real benefit, apart from academic
> >interest. (I admit to looking forward to a presentation about this at a
> >future DNS OARC meeting!) :)
> Could be persuaded to do something there :)
More information about the dns-operations