Re: defining service ceilings

In article <d41bid$s3e$1@sf1.isc.org>,
Brad Knowles <brad@stop.mail-abuse.org> wrote:
>At 12:13 PM +0200 2005-04-18, Ted Lindgreen wrote:
>
> Right. It's designed for root servers and TLD servers. It's not
>designed for general-purpose authoritative-only purposes.

I don't see a real difference between an root server, a TLD server
and a general-purpose authoritative-only server, but that's probably
my defect.

>
> Even for root and TLD servers, there are internal issues with the
>code that prevent it being used for some TLD providers, primarily
>having to do with the day the program wants to load all data in
>memory (and the particular type of uncompressed jump table that is
>built), etc....

All servers I know serve from data loaded in memory, at least nsd
loads much faster then bind.

>
> That is unless you folks have changed a lot more since I last
>looked at the program than I had thought.

The last time you looked(*), you looked at NSD 1.02b1 (beta 1), the
current version is 2.2.1. The chance that things have changed is high.

But since this is a bind newsgroup, this discussion is better carried
on on the nsd-users mailing list
(http://open.nlnetlabs.nl/mailman/listinfo/nsd-users).

(*) At least, according to the ripe talk in january 2003 you keep
referring to.
(http://ripe.net/ripe/meetings/ripe-4...ns-dnscomp.pdf)

jaap

--
Member of the human race since 1951