Distributing DNS servers

[Barrett Richardson]

On 28 Aug 1999 peter at nospam.se wrote:

> Barry Margolin <barmar at bbnplanet.com> wrote:
>

<cut>
 
> : In article <Pine.BSF.4.01.9908270950530.19335-100000 at phoenix.aye.net>,
> : Barrett Richardson  <barrett at phoenix.aye.net> wrote:
> : >
> : >I want to distribute my primary across a network topology for
> : >various reasons. I intend to have an ip address for the primary
> : >attached to a loopback interface on multiple machines at
> : >multiple points in my network (and use OSPF or BGP to establish
> : >reachability to various nameservers in various locations thru out
> : >the network).
> 
> : We're doing a similar thing.  If you traceroute to 4.2.2.1 from different
> : parts of the country you'll get a different machine.  We're not doing it
> : with a looback interface, but with a virtual address on the ethernet
> : interface.
>

<paste>

> Mr, Margolin:
>
> I must confess, i do not understand the motivation for these schenarios.
> Please enlighten me (and maye others) why this is done, and
> what benefits are gained.
>
> Thankx
> Peter h


Notice how Mr. Margolin's replies appear beneath quoted text
to establish context, and items that do not pertain to a reply
have been trimmed? This is likely intentional.

Think "high availability" and "essential service".

Draft draft-ietf-catalone-hadns-00.txt may be worth a gander.
It (and others) are available for perusal at

http://info.internet.isi.edu/1/in-drafts


-

Barrett