Re: [sidr] BGPsec without Extended Messages (draft-ietf-sidr-bgpsec-protocol)

[Matthew Lepinski <[e-mail]>]

Alvaro,

Thanks a lot. That makes perfect sense.

I support this change.

- Matt Lepinski

On Apr 4, 2017 1:18 PM, "Alvaro Retana (aretana)" <[e-mail]> wrote:

> On 4/4/17, 12:44 PM, "Matthew Lepinski" <[e-mail]> wrote:
>
> Matt:
>
> Hi!
>
> > The proposed changes seem reasonable, but I want to make sure that I
> > understand the path forward clearly.
> >
> > My understanding is that if we were to reach a future where
> > draft-ietf-idr-bgp-extended-messages is widely deployed, then the BGP
> > speaker's maximum message size will just be larger (than it is today)
> > and as a result we avoid reaching the point where Section 9.2 (of
> > 4271) guidance is needed.
> >
> > Is my understanding correct?
>
> Not exactly.
>
> The text in 9.2/rfc4271 is generic, it doesn’t apply to a specific message
> size; the maximum size is defined elsewhere.  The current text of
> draft-ietf-idr-bgp-extended-messages changes the size of the messages in
> Section 4 (of rfc4271), which is the same place where 9.2 points to.  IOW,
> the text in 9.2 would not change and still be applicable, the limit would
> just be reached later.
>
> > (I want to make sure that future
> > implementers will find our text clear and we won't need to revise this
> > spec to add clarity if extended messages ends up in widespread use.)
>
> To me, the main purpose of changing the BGPsec spec is to depend on
> whatever BGP does, and not on a future extension that may or may not be in
> the form it is today.  However, if we keep the reference to the known
> standard (rfc4271), then we should not have to update this document because
> we would just inherit whatever BGP does.
>
> Thanks!
>
> Alvaro.
>
>