Re: Comparing an old flow snapshot with some packet size data

"Karl Denninger, MCSNet" <karl@mcs.com> Thu, 08 August 1996 21:03 UTC

Received: from ietf.org by ietf.org id aa21888; 8 Aug 96 17:03 EDT
Received: from cnri by ietf.org id aa21884; 8 Aug 96 17:03 EDT
Received: from murtoa.cs.mu.OZ.AU by CNRI.Reston.VA.US id aa14608; 8 Aug 96 17:03 EDT
Received: from mailing-list by murtoa.cs.mu.OZ.AU (8.6.9/1.0) id GAA13476; Fri, 9 Aug 1996 06:56:09 +1000
Received: from munnari.OZ.AU by murtoa.cs.mu.OZ.AU (8.6.9/1.0) with SMTP id GAA13449; Fri, 9 Aug 1996 06:44:45 +1000
Received: from Kitten.mcs.com by munnari.OZ.AU with SMTP (5.83--+1.3.1+0.56) id UA25820; Fri, 9 Aug 1996 06:44:39 +1000 (from karl@mcs.com)
Received: from venus.mcs.com (root@Venus.mcs.com [192.160.127.92]) by Kitten.mcs.com (8.7.5/8.7.5) with SMTP id PAA11379; Thu, 8 Aug 1996 15:41:22 -0500 (CDT)
Received: by venus.mcs.com (/\==/\ Smail3.1.28.1 #28.5) id <m0uobtR-000IDOC@venus.mcs.com>; Thu, 8 Aug 96 15:41 CDT
Message-Id: <m0uobtR-000IDOC@venus.mcs.com>
Subject: Re: Comparing an old flow snapshot with some packet size data
To: Andrew Partan <asp@partan.com>
Date: Thu, 8 Aug 1996 15:41:21 -0500 (CDT)
Sender: ietf-archive-request@ietf.org
From: "Karl Denninger, MCSNet" <karl@mcs.com>
Cc: pferguso@cisco.com, jhawk@bbnplanet.com, kwe@6sigmanets.com, big-internet@munnari.oz.au
In-Reply-To: <199608081958.PAA01711@home.partan.com> from "Andrew Partan" at Aug 8, 96 03:58:00 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Precedence: bulk

> 
> This is all interesting stuff.
> 
> One question that I have been trying to figure out is 
> 	What size MTU should an ISP support on its backbone?
> 
> If we view the future where lots of hosts are connected via ethernet
> and fast ethernet & the like, then a MTU of 1500 would be 'correct'.
> 
> If we think that the future will have lots of hosts connected via
> Fddi or similar, then a MTU of 4470 would be 'better'.
> 
> Any ideas?
> 	--asp@partan.com (Andrew Partan)

Well, 4470 is the common FDDI and HSSI MTU, and so that's a pretty common 
one for high-speed links...

The trade-off IMHO has to do with the technology changes.  Cram 4470 into 
53-byte cells for ATM, and you end up needing 85 of them for each segment!
I suspect that some of the buffering problems we've seen with things like
Netedge boxes can be traced to that kind of encapsulation change problem...

Cut that MTU to 1500 and now you only need 29 cells for a segment.  Much
less likely to run into trouble.

The bigger problem is that some hardware out there can't hope to keep up at
DS-3 rates and above with very small segment sizes.  I don't know how much
of a problem this is in the real high-end hardware, but I do know that it
shows up instantly for those people trying to do the "cheap router in a PC
box" solutions.  Reports from the field are that a "100Mbps" network 
on which a PCI Pentium is sourcing or sinking traffic can not really expect 
to see more than about 50Mbps due to the packet processing overhead in 
this environment with a 1500 byte MTU -- but that number rises to nearly
85Mbps with a 4470 MTU.  Its not moving the data that is killing the
throughput -- its handling the encapsulation overhead.

I don't know how many people are stressing their backbone hardware far
enough in a restricted-MTU environment (intentionally) to know if this 
is really a problem out there or not in the real world on the core.  

Can a CISCO 7513, for example, handle near-theoretical throughput on a FDDI
ring if you change the MTU size to 1500 from 4470?  Or does its routing
performance suffer?

--
--
Karl Denninger (karl@MCS.Net)| MCSNet - The Finest Internet Connectivity
http://www.mcs.net/~karl     | T1 from $600 monthly; speeds to DS-3 available
			     | 23 Chicagoland Prefixes, 13 ISDN, much more
Voice: [+1 312 803-MCS1]     | Email to "info@mcs.net" WWW: http://www.mcs.net/
Fax:   [+1 312 248-9865]     | Home of Chicago's only FULL Clarinet feed!