Re: Disabling temporary addresses by default?

[Michael Richardson <[e-mail]>]

Fred Baker <[e-mail]> wrote:
    > +1 on this. We can discuss whether a temporary address should have a
    > limited lifetime or whether a given interface might have at most some
    > number of temporary addresses at any given time, but to say that in the
    > default (e.g., usual) case each system has zero seems like a bad idea.

I agree with what you said, and the way that you said it.

In particular, I think that it would be better to have a temporary address
per user and/or threadish-pool (e.g. Chrome) and/or container, and keep this
for a longer time.
I think that if a host/network is going to support multiple temporary
addresses, that they should be used in parallel and used for a long time.

I'm aware that Android (Lorenzo, Jen, etc.) been resistant to supporting
DHCPv6.  This is causing a bit of schism, because we of how we have made the
stateful/stateless split.  Maybe we need a stateful SLAAC that isn't DHCPv6
to make this work.  Or something like that.  Or find a way to address the
issues that Android has with supporting DHCPv6 client.

(Please feel free to point me again to some summary discussion from 2003,
when the objection was first detailed)

-- 
Michael Richardson <[e-mail]>, Sandelman Software Works
 -= IPv6 IoT consulting =-