[comp.dcom.sys.cisco] cisco's FDDI MTU vs. RFC1188

robelr@ucs.indiana.edu (Allen Robel) (05/22/91)

A show interface on a cisco FDDI interface shows an MTU of 4470. 
RFC1188 specifies an MTU of 4352 octets for FDDI:

      Therefore, the MTU of FDDI networks shall be 4352 octets.  This
      provides for 4096 octets of data and 256 octets of headers at the
      network layer and above.  Implementations must not send packets
      larger than the MTU.

I don't think that there is any practical harm in this discrepancy
in our environment as we're just using the FDDI to connect
ethernets and I don't think that the cisco would ever generate
frames this large on its own.  But I was wondering if some
enlightened soul might shed some light on why the descrepancy
exists.  Too, I'm bored and couldn't think of a better way
to spend my time than to post this :-)

thanks,

Allen Robel                         robelr@mythos.ucs.indiana.edu 
University Computing Services       ROBELR@IUJADE.BITNET 
Network Research & Planning         voice: (812)855-7171
Indiana University                  FAX:   (812)855-8299

William "Chops" Westfield <BILLW@mathom.cisco.com> (05/22/91)

    A show interface on a cisco FDDI interface shows an MTU of 4470. 
    RFC1188 specifies an MTU of 4352 octets for FDDI:

      Therefore, the MTU of FDDI networks shall be 4352 octets.  This
      provides for 4096 octets of data and 256 octets of headers at the
      network layer and above.  Implementations must not send packets
      larger than the MTU.

The RFC reserves some space for "expanding the MAC header" (what a bad
idea (expanding it, not reseving the space)).  The cisco implementation
pre-dates RFC1188, and is based on RFC1103, which specifies 4470.

Of course, you can change the MTU on the cisco to whatever you would like,
assuming it is smaller than that actually supported by the media...

Bill Westfield
cisco Systems.
-------

fin@unet.unet.umn.edu (Craig A. Finseth) (05/23/91)

	...
   The RFC reserves some space for "expanding the MAC header" (what a bad
   idea (expanding it, not reseving the space)).  The cisco implementation
   pre-dates RFC1188, and is based on RFC1103, which specifies 4470.
	...

Actually, this is not a bad idea.  For example, the 802.10 (secure
LANs) standard specifies additional header information by necessity.
This extra buffer prevents fragmented Ethernet packets, which are a
much worse problem than losing a tiny percentage of an FDDI packet,
the vast majority of which will be either 576 or 1536 bytes anyways.

Craig A. Finseth			fin@unet.umn.edu [CAF13]
University Networking Services		+1 612 624 3375 desk
University of Minnesota			+1 612 625 0006 problems
130 Lind Hall, 207 Church St SE		+1 612 626 1002 FAX
Minneapolis MN 55455-0134, U.S.A.