host-identifier with IPv6

Simon Hobson dhcp1 at thehobsons.co.uk
Mon Mar 2 21:12:59 UTC 2009


Ted Lemon wrote:

>On Mar 2, 2009, at 1:17 PM, sthaug at nethelp.no wrote:
>>Why is this difficult to understand?
>
>Actually, no feature is being removed - I don't know where you got 
>that.   But in any case, this is a classic example of the situation 
>where no matter what you do, someone doesn't get what they want. 
>You seem to have concluded, based on no evidence, that what most 
>people want most is a less-stable unique identifier that allows them 
>to track devices by the sticker on the side.

This may be a fundamental difference of opinion, but the MAC address 
of the devices interfaces is probably as stable as anything you are 
likely to get on current devices. It only changes if you physically 
change the interface or reprogram the flash chip (both generally rare 
events). Furthermore, barring mistakes, it *should* be globally 
unique.
Unless I'm misunderstanding, each OS will generate a unique ID when 
it first boots on a device - each OS will generate a different one, 
and each time an OS is re-installed it is likely to change. And just 
for good measure, there is no way to know it in advance of 
configuring the device - precluding pre-configuring of provisioning 
systems ready for the device to "plug in and go".

>In my personal experience, this is an edge case.   Yes, there are 
>people who want to do this, but they are by no means the majority of 
>users.

Actually, I don't think it's such a small minority as you suggest by 
calling it an edge case. A minority probably, but probably the 
minority running large installations where DHCP should be of the most 
benefit.
But it doesn't even have to be a large installation for this to be of 
benefit. At my last job, I ran the DHCP and just looking at printers 
it's easy to see how it's "quite useful" to be able to configure the 
systems so that they know "device X is to be called Y" - even if you 
use dynamic addressing and refer to them by DNS name. BEing able to 
read the identifier required to program the system(s) off the label 
on the side of the box saves a fair bit of work.

>And we have a solution that addresses their needs.

Do we ? The impression I have from this thread is that we "probably" 
have a solution for "most" clients, and it will "most likely" work ...

Please correct me if I've got this wrong, but I had the impression 
that there is no unique and immutable identifier sent in DHCP6 
requests - and certainly not a 'plain text' MAC address for an 
interface. At least one of the ID's sent by the client "most likely" 
contains the MAC address, but at the same time the RFC says that 
servers should not try and interpret the contents of this ID (because 
the content is not defined).


Am I missing something ?

PS - I'm not trying to "have a go" at you or anyone else, but it does 
seem that "something is missing" here.

-- 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.



More information about the dhcp-users mailing list