Thomas,

> One of the original motivations for disabling an interface if DAD
> fails for the LL address is that if you are running DAD on a LL
> address generated using the EUI-64 format, in the absence of a DOS
> attack, if DAD fails, that really means duplicate ethernet addresses
> are in use. In this case, the device is hosed and configuring another
> address won't result in a usable network. In fact, it probably makes
> things worse by creating problems that harder to diagnose than just
> shutting down the interface. The user will see a partially working
> network where some things work, and other things won't.
> 
> BTW, some of the above text might be worth putting into 2462bis as
> background information.

Understood about the original intention, but this may not be the case.

2462 currently says:

   Note that interface identifiers will typically be 64-bits
   long and based on EUI-64 identifiers as described in [ADDR-ARCH].

This may not always be the case. See 3GPP, for example.  Also,
I remember that the MIP folks wanted to be able to use a random
number in place of the EUI-64 in case the DAD failed.  

My question, is the original quote above still a valid assumption?
If so, what does 'typically' mean?

John

--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to