Re: [Idr] WGLC for draft-ietf-idr-mrai-dep-03

[<[e-mail]>]

Paul,

Thanks for your reply. More inline

> From: Paul Jakma [mailto:[e-mail]]
> On Tue, 29 Mar 2011, [e-mail] wrote:
> 
> > Sorry if I'm missing something obvious or already discussed but is
> > there a need to recommend that an AS (SP/ISP) use (hence configure)
> > the same MRAI value for all its IBGP sessions?

Any recommendation on the above point?

> > Also, RFC 4271 made
> > a distinction between iBGP and eBGP sessions (basically MRAI value
> > on iBGP should be lower than on eBGP). Does this need to be added
> > or is this also deprecated?
> 
> That's out of scope. Implementations may or may not wish to do that.
> An earlier form of this draft (draft-jakma-mrai-02) gave more of an
> outline, but it doesn't seem there is a consensus on anything more
> specific than that the current RFC4271 defaults are not appropriate.

>From your above statement, I understand that the draft only deprecate
the 2 default values defined in RFC 4271. Hence the following RFC 4271
text is still applicable: " Since fast convergence is needed within an
autonomous system, either (a) the MinRouteAdvertisementIntervalTimer
used for internal peers  SHOULD be shorter than the
MinRouteAdvertisementIntervalTimer used for external peers, or (b) the
procedure describe in this section  SHOULD NOT apply to routes sent to
internal peers."

In that case, IMHO, the draft sentence " The appropriate choice of
default values is left to the discretion of implementors." Should be
augmented with "but as per RFC 4271 default value should be chosen lower
for iBGP peer than for eBGP peer".


Thanks,
Regards,
Bruno

> 
> regards,
> --
> Paul Jakma  [e-mail]  twitter: @pjakma  PGP: 64A2FF6A
> Fortune:
> *** Rince is [e-mail] (We have Joey, we have Fun, we
have Linux on a Sun)
>  	-- Seen on #Debian