Re: [v6ops] A common problem with SLAAC in "renumbering" scenarios

[Fernando Gont <[e-mail]>]

Hello, Lorenzo,

On 26/2/19 11:19, Lorenzo Colitti wrote:
> On Tue, Feb 26, 2019 at 7:59 PM Fernando Gont <[e-mail]
> <mailto:[e-mail]>> wrote:
> 
>     >>                             Default: 2592000 seconds (30 days), fixed
>     >>                             (i.e., stays the same in consecutive
>     >>                             advertisements).
>     >
>     > So given that document is 12 years old, with that default copied
>     from one that is 21 years old, is an update required? 
> 
>     That's my take.
> 
>     > And if so, to what?
> 
>     At the very least, never larger than the "Router Lifetime". Please see:
>     https://github.com/fgont/draft-slaac-renum/blob/master/draft-gont-6man-slaac-renum-02.txt
>      (particularly Section 5.1.1)
> 
> 
> Disagree that prefix lifetimes must never be larger than the router
> lifetime, particularly because 0 is a perfectly valid default lifetime.

How does a Lifetime of 0 fit in RFC8028?



> Also looking at section 5.1.1, the default MaxRtrAdvInterval of 5
> minutes violates the best practices in RFC 7772.

Thanks for the pointer!

Looking at RFC7772, it says:

      Thus, the minimum lifetimes for
      RA configuration parameters such as default router lifetime might
      reasonably be 5-10 times the RA period, or roughly 45-90 minutes.


Question: I'm trying to parse this paragraph. It would seem that, in
order to ahve an appropriate RA frequency, you need to have a Default
Router Lifetime of 45-90 minutes, but the default is actually 30 minutes.

Am I missing something?

And, how do you get the 45'-90' numbers?

FWIW, with the default values in RFC4861 you get, for the *average*
case, ~ 6.6 RAs /hour, and a DEfault Lifetime of 30 minutes...

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