[36976] in North American Network Operators' Group
Re: jumbo frames
daemon@ATHENA.MIT.EDU (Richard A. Steenbergen)
Fri Apr 27 15:46:19 2001
Date: Fri, 27 Apr 2001 15:43:43 -0400 (EDT)
From: "Richard A. Steenbergen" <ras@e-gerbil.net>
To: Stephen Sprunk <stephen@sprunk.org>
Cc: nanog@merit.edu
Message-ID: <Pine.BSF.4.21.0104271516310.98098-100000@overlord.e-gerbil.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Errors-To: owner-nanog-outgoing@merit.edu
On Fri, Apr 27, 2001 at 02:10:48PM -0500, Stephen Sprunk wrote:
>
> Or unless you actually have >1500 MTU to the hosts, which is quite
> possible. A traffic study from MCI's backbone (obviously years ago)
> showed nearly 40% (byte-wise) of their traffic was in packets >1500
> bytes. With the death of FDDI, this has probably come down, but
> GE-attached servers in colos should push it back up.
Probably not since GigE defaults to 1500 bytes and there is no agreed upon
standard for jumbo frames. Right now it doesn't make much sense to attempt
jumbo frames for packets headed outside of your administrative control,
because of PMTU-D problems.
> Router performance is, however, directly related to packet size, since
> forwarding overhead is per-packet and not per-byte. It is much easier
> to fill big pipes with 9000 byte packets than 1500 byte packets.
More importantly it's easier to generate the packets in the first place.
The amount of overhead that goes into generating a packet on a host is
really quite a bit, and operations on 1500 bytes are exactly the opposite
of how most hosts are optimized. At gigabit speeds (125MB/sec) it only
takes 12 microseconds to transfer 1500 bytes.
--
Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/ras
PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)
