Help with DHCPv6 client-identifiers

[Ted Lemon]

On Nov 20, 2011, at 12:12 AM, "Peter Grandi" <pg_dhcp at dhcp.for.sabi.co.UK> wrote:
> This I sort of agree with but not to the extent of "completely
> inexusable". I think that tools should be able to be configured
> to so "odd" things, for people who do know what they are doing,
> as the IP (not Internet!) tradition is to provide low level
> features even if it means giving people enough rope to hang
> themselves with it.

We went with a DUID instead of a hardware address because hardware addresses aren't unique, and this has been a problem in the past even within a single subnet.   It would be nice if the MAC address could be assumed to be unique, but it can't.

There is a draft working its way through the DHC working group that proposes a mechanism for using the hardware address not as an identifier, but as a hint, and I expect it will be adopted and standardized.   It will probably cause the occasional operational problem in dual-stack environments when it incorrectly identifies a DHCPv4 client as being the same device as a DHCPv6 client when actually they aren't, but it shouldn't cause any harm in DHCPv6-only environments.   It should also allow you to do the bar code scanner DHCP server setup, as long as you follow the spec and don't get carried away.

But the idea that it's completely inexcusable that we didn't use the MAC address as a host identifier is pretty absurd.   There was a time when I used to get flamed on this very mailing list because using the MAC address as a host identifier caused severe brokenness.   The bottom line is that you just can't win: someone will always have a legitimate complaint no matter how you address any particular problem.