DDNS and DHCP

Kevin Darcy kcd at daimlerchrysler.com
Thu Apr 12 18:28:37 UTC 2001


Adam Lang wrote:

> ----- Original Message -----
> From: "Kevin Darcy" <kcd at daimlerchrysler.com>
> To: <bind-users at isc.org>
> Sent: Wednesday, April 11, 2001 5:44 PM
> Subject: Re: DDNS and DHCP
>
> > > Well if I have a network printer that is going to be used 95% of the
> time by
> > > a subnet and 5% of the time outside of it, I want that printer in that
> > > subnet to control traffic.  I also need the printer to be static.  How
> would
> > > I have the printer in that subnet, but on a different DNS zone than the
> > > dynamic clients in the subnet?
> >
> > If the printer is static, how does Dynamic DNS come into play at all? You
> just
> > define the printer and never change it. Whether you define it by editing
> the
> > zone file, or through "nsupdate" or some other Dynamic Update mechanism,
> is
> > kind of irrelevant if it's not changing.
>
> Because the clients would be dynamic.  And that was what I was asking.
> About having static ip's in a zone that is handled by dynamic updates.

If you allow your zone to be updated by a DHCP server dynamically, then this
just becomes a matter of DHCP server configuration. The DNS server has no way
of knowing whether an incoming Dynamic Update is for a statically-assigned or
dynamically-assigned address.

>
>
> >
> > And what do you mean by "control[ling] traffic"? Do you have some
> proprietary
> > system in place which controls print-output traffic using DNS names?
> Please
> > elaborate; it's not clear at all what it is you're trying to do here.
>
> If I have 100 employees in my company and 4 printers and the network is
> divided into 4 subnets with a printer in each.  If 90% of printer traffic
> comes from inside the subnet it resides in, it helps control traffic (as
> opposed to sticking all the printers on the same subnet).
>
> This is basically what I wanted to do:
>
> When I configure a PC, I'll give it a name (no big deal since I have to
> configure Client Access for AS/400 emulation and have to give it a "device
> ID".  Figured I'd just use the same name for the hostname the client tells
> DHCP it wants).
>
> DHCP would hand out an IP address, take the requested hostname, and stick
> them both in DNS.  This way it would be possible to track user traffic no
> matter what their IP address is.  Well, all printers are network attached
> and I would want them to have static IPs and not to change and I was
> inquiring on how to do that since it says that you can't have static content
> in a dynamic zone.

Okay, I think you're confusing statically *assigned* addresses (which is a
DHCP concept) with statically *maintained* zones (which is a DNS concept). You
can have statically-assigned addresses in a dynamically-maintained zone, but
your DHCP has to support that feature.

> As for multiple zone's on the same subnet, I am not grasping how that would
> be done.

As I said, the reverse namespace of DNS is based on octet boundaries. So if you
have a subnet that spans multiple /24's (e.g. 129.9.26/23), then you could have
multiple zones for that subnet (i.e. 26.9.129.in-addr.arpa and
27.9.129.in-addr.arpa). If your subnets are /24 or smaller, then this does not
apply to you. In that case, it's one zone per subnet, unless you get into
"classless delegation" as per RFC 2317...


- Kevin





More information about the bind-users mailing list