I perhaps; should have been *clearer*!
Default-MTUs for different media-types are *different*!

That does not by any stretch of the imagination *imply* that a router/switch's 
BVI/SVI/L3-interface will *auto-magically* coalesce(INCREASE) transit-datagrams 
to fit the MTU of the outgoing-interface.

"Nagle" in this case does not apply since we are talking transit-traffic as 
opposed  to locally-generated from said switch/router.

Going back to my original-response:

Fragmentation/Re-assembly happen at L3 NOT at L2.(as long DF bit is not set
by initiator.(seperate-tutorial?) Also-Applicable is the 
"lowest-common-denominator" mentioned in my original response.

What you are trying to accomplish:

1) auto-magically-coalesce datagrams to fit the larger-mtu-of the outgoing-int 
doesn't work.

As mentioned before, the lowest-common-denominator is what works.

./Randy

-- On Mon, 8/22/11, Sridhar Ayengar <[email protected]> wrote:

> From: Sridhar Ayengar <[email protected]>
> Subject: Re: [c-nsp] BVI MTU Question
> To: "Randy" <[email protected]>, "Cisco NSPs" <[email protected]>
> Date: Monday, August 22, 2011, 3:35 PM
> Randy wrote:
> > I was thinking all ethernet for some reason...but
> translation between different media-types should work.
> > My bad....
> 
> Is gigabit ethernet (1000BASE-SX) considered the same media
> type as fast ethernet (100BASE-TX)?  Because that's my
> configuration.
> 
> Peace...  Sridhar
>


_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

Reply via email to