On 27-jul-2007, at 16:07, Matt Mathis wrote:
A better approach is to exchange MTU information at the neighbor
discovery / ARP stage. That way, you're not breaking anything that
your correspondent may be depending on.
The problem with an explicit protocol solution is that you have to
get both
parties to deploy it before anybody gets any gain.
That's not completely accurate. If an OS implement such a mechanism,
then it becomes possible for hosts or routers running this OS to
exchange larger packets even on subnets where jumboframes aren't
configured explicitly.
So it could even help vendors sell more gear if they're the first to
implement this. :-)
You could consider adding a padding option to neighbor discovery,
in the style
of RFC4820.
[Can't look up the RFC right now, on my way back home from Chicago.]
Would you send such an ND packet as the first one? That would be
suboptimal, because if the other side doesn't support your packet
size the packet is lost and you need to retry after a timeout. If
you're doing such an ND after a regular one, then there is no real
need for it to be an ND. An ICMP echo request would accomplish the
same. (Except in the case of Windows, of course, which filters them.)
Iljitsch
_______________________________________________
Int-area mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/int-area
