Re: draft-van-beijnum-multi-mtu-05.txt

[David Lamparter <[e-mail]>]

On Mon, Apr 11, 2016 at 05:18:07AM -0700, Sowmini Varadhan wrote:
> On Wed, Apr 6, 2016 at 12:55 PM, Erik Nordmark <[e-mail]> wrote:
> > On 4/6/16 12:18 PM, David Lamparter wrote:
> >> On Mon, Apr 04, 2016 at 11:37:17PM +0200, Mikael Abrahamsson wrote:
> >>
> >>> I'll be presenting
> >>> https://tools.ietf.org/html/draft-van-beijnum-multi-mtu-05 in 6man on
> >>> Wednesday. There are operational considerations here, so I encourage
> >>> people to drop in and/or comment on this draft.
> >>>
> >> Ok... easy comments/opinions first:
> >> - should definitely be ICMP, not UDP.  Erik Nordmark's suggestion to
> >>    piggyback on NUD sounds great, though admittedly I haven't thought it
> >>    through looking for pitfalls.
> >>
> > One issue compared to a dedicate probe is that a NUD (a unicast NS) of
> > size X wouldn't necessarily result in a unicast NA that is also padded;
> > perhaps one can add such behavior so that we can get the bidirectional
> > probe in both directions.
> 
> 
> there's also the question of what mtu to advertise in the face of
> pvlans, when there is a router on a promisc port proxying the NA.

A system doing proxy NA is ultimately a router and should be handled as
such; i.e. if done "right" there is only one 1:1 relationship for which
first-hop MTU discovery is done.

> In practice, since a node typically uses a single interface to talk
> to multiple onlink neighbors, and jumbo mtu etc is configured per
> interface, you would end up getting metered down to the lowest
> common denominator for the subnet, right?

No, the per-interface configuration is exactly what this draft is trying
to work around.

Regards,

-David