Re: BFD WG adoption for draft-haas-bfd-large-packets

[Jeffrey Haas <[e-mail]>]

Les,

A brief note, and one worn as a chair:
Despite having strong personal preferences about mail quoting, I realize
that not everyone gets a chance to choose a client of their own choice.
But that said, your responses in this thread have been impossible to parse
out from the surrounding reply in text-plain readers.  This includes the
IETF mailarchive.

Please make adjustments to your mail reply styling so as to be more
readable.  Thanks. :-) 

And now on to non-chair commentary:


On Tue, Oct 23, 2018 at 02:57:17PM +0000, Les Ginsberg (ginsberg) wrote:
> Reshad –
> 
> I think there are two cases:
> 
> 1)A client who uses encapsulation (MPLS is an easy example) and really needs to know if an MTU sized packet plus the encapsulation can be delivered vs a client who does NOT use encap and only cares about Layer 3 sized MTU packets
> 
> 2)A client which has deliberately limited the max-sized packet they will ever send and really only cares about (MTU –X)

While I agree with your sentiment from the perspective of BFD clients, I
think the use case goes slightly bigger than this.

In the cases driving Albert's original scenario, we're not protecting BGP,
IGPs or other control plane protocols for large packets.  What's really
being protected is user payload.

So, to some extent, the use case piggy-backs on existing protection of a
client to guarantee the path has a sufficient MTU.

-- Jeff