On 7/9/10 10:10 AM, Tony Li wrote:
Tony,
Yes, but my point was that we probably cannot assume
that every "device" in the Internet has a unique hardware ID,
because we may have many virtual hosts in the future.
How "smart" individual vendor's solutions are is IHMO out of scope here.
I respectfully disagree. All physical devices have had a MAC address (or
equivalent) for quite some time now. Even the virtual devices are emulating
Ethernet, so for all practical purposes, EUI-48 and EUI-64 are a clear and
distinct identifier space.
The virtualization vendors need to be a little bit more pro-active in MAC
address allocations, independent of everything else, and then IPv6 is good to
go.
Ok. Moreover, I think that using the IPv6 privacy extensions in
combination with ILNP is actually quite useful, i.e., the ID
is fine for servers, but for end-users it's probably a good idea
to have several IDs an change them from time to time.
Some people feel that that's appropriate. That's fine. I'm more of a mobility
junkie myself.
hmm, I am also a 'mobility junkie' but that is precisely why I am
concerned about a globally unique, persistent, hard to change
identifier. Thinking out loud, how about instead of using the privacy
extensions make it possible (for an end user) to perform some kind of
one-way function on the EUI-64 address that can not be related to the
original hardware address. It would then be up to the user to not apply
this function, apply it once or apply it many times, whatever the
trade-off between location privacy and session persistence of the user
duictates.
Klaas
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
