Re: Address privacy

[David Farmer <[e-mail]>]

On Thu, Jan 30, 2020 at 1:03 PM Michael Richardson <[e-mail]>
wrote:

>
> Tom Herbert <[e-mail]> wrote:
>     > I don't understand why 24 hours is considered the preferred lifetime.
>
> I think that this is a good and definite RFC4941bis comment that we should
> go
> into.
>
> I think that using temporary addresses in series, for short periods of
> time,
> is useless: WHEN IT COMES TO PREVENTING TRACKING WHEN NOT MOBILE.
>

I agree RFC4941 does little to prevent sophisticated tracking
especially when a device is stationary, what it does do is prevent trivial
tracking especially over time frames of months or years. Without RFC4941, a
stationary device could be using the same address for months or even years
at a time, making it trivial to track these devices.

Devices that are linked to or primarily used by humans frequently have
diurnal cycles just as the humans that are using them usually do. Changing
addresses once a day is unlikely to cause too much disruption to
these diurnal cycles, the more frequent the changes the more likely there
will be unanticipated consequences. Even today most applications are
written using an always-on model of networking, where addresses never
change.  And, since most applications are used by humans in a diurnal
cycle, a daily address change is likely to go completely unnoticed by an
application, or at least not have any impact that is noticed by the actual
human user.

In my opinion, anything between 8 and 24 hours in most cases will be more
or less equivalent in most use cases, from privacy and disruption
perspective. Going shorter than 8 hours could provide minor privacy
benefits but is likely to create much more disruption.

Finally, I doubt changing addresses will have any
significant impact on sophisticated tracking even at change frequencies in
the range of minutes and at those change frequencies, network stability is
likely to be impacted, especially with many hosts making address changes at
those frequencies.

Basically, temporary addresses changing on a daily basis seems like a
prudent thing to do, this shouldn't cause too many problems and should
prevent the most trivial forms tracking especially over long time periods
like months and years.

By the way, I think calling them privacy addresses sets the bar way too
high, maybe the title should be changed as well, "Privacy Extensions for
Stateless Address Autoconfiguration in IPv6" is probably setting the bar
too high.   Maybe "Temporary Address Extensions for ..."

Thanks

-- 
===============================================
David Farmer               Email:[e-mail]
Networking & Telecommunication Services
Office of Information Technology
University of Minnesota
2218 University Ave SE        Phone: [phone]
Minneapolis, MN 55414-3029   Cell: 612-812-9952
===============================================