Re: Address privacy

[Fernando Gont <[e-mail]>]

On 26/1/20 18:59, Brian E Carpenter wrote:
> On 27-Jan-20 10:39, Manfredi (US), Albert E wrote:
> ....
>> This seems like a good default, no? The business about "privacy" concerns not so much the type of privacy that data encryption would provide, but rather, being able to track an individual in his/her travels. Changing the IID only when the prefix changes should prevent that well enough?
> 
> The IID is normally set *before* the host generates its Link Local address, which is normally before it starts to listen for RAs from which it will learn the current prefix(es).
> 
> So you'd have to make the IID for SLAAC independent of the IID for LL (which is of course exactly what RFC4941 does).

Technically speaking, LL are also generated with SLAAC -- albeit ont in 
response to RAs.

So the IID of LLs should also be generated with RFC7217. If the 
implementation employs the SSID as one of the arguments to F(), the LL 
address will also change as a node roams across wifi networks.


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