Re: [v6ops] Same interface ID under several prefixes

[Fernando Gont <[e-mail]>]

Hi, Brian,

On 17/6/22 01:45, Brian E Carpenter wrote:
> Hi,
> 
> As a side effect of a WG Last Call over on 6man, I've come up with a 
> question about SLAAC implementations.
> 
> On Windows 10 (patched up to date), if you find the Interface ID in a 
> stable GUA or ULA, it is also the Interface ID in the device's 
> link-local address. I see the same identifier used for three different 
> prefixes right now (GUA, ULA, LL). All you have to do is type "ipconfig" 
> at the Windows command prompt.

FWIW, Windows has implemented this behavior since a long time (pre RFC7217).

IIRC, they essentially generate a token at systems installation time, 
and then they use that token in replacement of the "Modified EUI-64" IIDs.

IIRC, they did this at the time to mitigate address scanning attacks, 
before we produced RFC7217/RFC8068.


> However, it seems to me that this practice violates RFC 7217, which says:
> "   o  The resulting Interface Identifiers must change when addresses are
>         configured for different prefixes. "
> 
> Am I right to consider this a bug?

In all fairness, I'd call it a non-standard implementation, which does 
not comply with the requirements in RFC8064, and certainly is not 
compliant with RFC7217.

I assume they don't claim to implement RFC7217. -- If they did, then 
yes, it would be fair to call that a bug. :-)

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492