[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: compatible address support [Re: WG Last Call: draft-ietf-v6ops-6to4-security-02.txt

Pekka Savola wrote:
On Thu, 3 Jun 2004, Fred Templin wrote:

The example in 6.1 cites compatible addresses.   While these are in use,
they are being deprecated in draft-ietf-v6ops-mech-v2-02, so I suggest
the example is reworked a little and that the reference is to the new
draft and not RFC2893 (reference [4]).


I agree with Tim about updating the [MECH] reference, but I was also
wondering about when it would be OK for implementations to begin
dropping support for IPv4-compatible addresses? I guess there could
be several alternative approaches, including:

 1) remove the IPv4-compatible address support code completely?
 2) leave the code, but surround it with compile-time directives?
     (default-disabled vs. default-enabled is another decision point)
 3) other?

I seem to recall seeing some earlier discussion on this, but perhaps
there are more current viewpoints based on the extensive  scenarios
and analysis work done by 'v6ops'?



IMHO, this is probably an implementation choice, and not something we
can or should specify here.  Personally, I see very little use for
them..


We had a similar discussion around site local and in the end decided that
all the IETF should say is:

   Existing implementations and deployments MAY continue to use this
   prefix.

In other words, it's not our business how implementors choose to
roll back a deprecated feature.

Brian