The Routing Over Large Clouds Mailing List Archive by date

[Paul Koning 1695 <pkoning@chipcom.com>]

Dave Katz wrote:
>The mistake of putting hardware type instead of address type into ARP
>resulted in no end of ugliness when multimedia bridges came along (which
>is why FDDI uses the "Ethernet" hardware type).  In retrospect, what should
>have been defined was an "IEEE 802" address type used across all media;
>this is in effect what the Ethernet hardware type has become.
>
>It seems as though what's trying to be accomplished is a binding
>between a subnetwork address and a network layer address; by
>introducing the media type into the process it only makes multimedia
>interworking below the network layer difficult (regardless of whether
>or not I might think this is a good idea).

Dave and I share some of the same scars from this one.  I'll have to
endorse what he said, enthousiastically.

The bottom line is that the presence of a field called "hardware type"
in the classic ARP protocol always was a mistake.  It didn't hurt us until
someone introduced a new code for "IEEE 802", oblivious to the fact
that this uses the same addresses as Ethernet and therefore should not
be treated differently.   Fortunately, this was caught during FDDI 
development
and at least in that case did not persist past the early deployment
stage.  Even so, it caused plenty of pain then, and most inplementations
probably still have extraneous code lying around to treat hardware type
6 in an ARP as equivalent to type 1.

Based on this experience, I feel that there should NOT be a "hardware
type" field in any address resolution protocol.  It would be useful to have
an "address type" to handle those cases where multiple address formats
are in use on the same medium.  However, in that case it is absolutely
essential that the SAME "address type" value be used on different
media to represent the same addresses, or addresses that differ only
in a trivial one-to-one mapping.  (For example, Ethernet and ALL 802
family 48-bit addresses are ONE address family, not two in spite of the
dual bit order.  However, 60-bit 802.6 addresses (if they still exist) are
presumably a different address type.)

     paul koning
     (pkoning@chipcom.com)