IP Address Management Tool (IPAM) for DNS and DHCP

[Persiko, Mark]

Hello,

I would submit that IPAM solutions help manage IP address and name spaces, as well as producing outputs (named.conf and zone files, in the case of BIND) to support DNS resolution.

IP address and name spaces could be seen, perhaps, as "inventory" of two types of network resource.  While BIND 10 could be augmented to support inventory functions, one would want to be careful in how tight a coupling one makes between the two, and whether it would weigh BIND down in the process.

I would certainly like to see XML-based configurations.  Many API's support XML output now, and conversely, can read XML input.

True, BIND provides validation as a "protocol engine", as it's harder to write an RFC "rules engine!"  :)

Mark

-----Original Message-----
From: bind-users-bounce at isc.org [mailto:bind-users-bounce at isc.org] On Behalf Of Paul Vixie
Sent: Tuesday, March 04, 2008 12:09 PM
To: bind-users at isc.org
Subject: Re: IP Address Management Tool (IPAM) for DNS and DHCP

"Persiko, Mark" <Mark.Persiko at Level3.com> writes:

> BIND-DLZ looks highly desirable as an augmentation to a DNS management
> tool (by that, I mean a database with DNS information could be seamlessly
> tied to BIND servers.)  However, is that part of the standard
> distribution now, or would it need integration (and optimization) to work
> its way into BIND 10?

to the best of my knowledge, DLZ is a standard feature in late model BIND9.
equivilent functionality will almost certainly find its way into BIND10 (but
i hope we have hot-spot caches with SQL-triggered invalidation, and i hope we
can accept RFC2136 updates and back-propagate them into SQL, both of which
prevent me from running DLZ on my own zones.)  we (ISC) love that BIND9 is
seen as a general DNS protocol engine for other folks' DNS storage systems.

what i'm looking for in this thread, though, is management features like
clustering, XML-based config, better support for GUI, or other reasons why
people aren't running raw BIND9 and instead pulling in something like
InfoBlox, M&M, etc.  how can BIND10 better support this functionality, and/or
better support these vendors, than BIND9 does?
--
Paul Vixie