Re: Re :[DNA] Best current practice for DNA

[Pekka Nikander <pekka.nikander@nomadiclab.com>]

Daniel,

> I am simply pointing out what DNA BCP is...
> DNA BCP is may composed of
>
>  1) L2 Triggering

Yes.

>  2) DAD Potimization (for LLA)

No, not really.  For the BCP, we can't consider changing
current mechanism, just how to use them more effectively.
Hence, we may consider leaving out DAD if the host is
perfectly sure that it is on a link where it was before,
but not otherwise.  However, in any case where we leave
out DAD, we have to do a very good analysis of the potential
problems, e.g., some other host may have taken the address
in the use while the now-entering hosts has been off-line.

>  3) RS (reduce random delay on the host)

No, not really.  Again, we do not want to change the
current specs for the BCP.

> 4) Fast RA (remove random delay on the router)

Again, same as above.

Using optimized DAD (if that proceeds in IPv6), shorter
random delays etc MAY be considered for the protocol
enhancements, when we start working on that.  Hence, it
is good to keep thinking about them, and to do simulations
and experiments.

--Pekka