Newsgroups: comp.dcom.lans
Path: utzoo!utgpu!dennis
From: dennis@gpu.utcs.utoronto.ca (Dennis Ferguson)
Subject: Re: Packet Drivers or NDIS Drivers for Netware
Message-ID: <1991Mar8.033923.13241@gpu.utcs.utoronto.ca>
Keywords: NDIS,Netware,Drivers
Organization: none
References: <1991Mar6.201431.18839@novell.com>
Date: Fri, 8 Mar 1991 03:39:23 GMT

In article <1991Mar6.201431.18839@novell.com> donp@na.excelan.com (don provan) writes:
>The existence of ARP shows that IP itself isn't media-independent, so
>of course complete media-independence for an IP implementation is not
>possible.  Other protocols, however, such as CLNP, XNS, and IPX,
                                              ^^^^
>require no media specific address mapping because they pass media
>addresses around at the network layer.  Here, ODI has an overwhelming
>advantage because these protocols can, in fact, be completely
>media-independent.
>
>In other words, IP's use of logical addresses is not typical.

I don't think it is true that CLNP passes media addresses around at
the network layer.  Or more specifically, if your CLNP implementation
assumes that a useful MAC address is encoded in the NSAP I think you
are going to be unhappy with it.

GOSIP-format NSAPs do include a 6 byte System ID, but it is not a
a requirement that this contain a MAC address.  Hosts which autoconfigure
their NSAPs may put a MAC address in there to guarantee the generation
of a unique number (this is why the field is 6 bytes), but you can't
assume that all hosts have done this.  And even if you could, multihomed
hosts (with more than one interface but perhaps just one NSAP) would still
break the assumption that you can use the address for sending.

ES-IS (which is sort of an ARP-replacement rolled in with a gateway
discovery protocol) is used to do media-dependent address mapping for CLNP,
in a way which is quite analogous to ARP and IP.

I'm not really sure that the list of protocols which use logical addresses
is limited to IP and CLNP, either (SNA seems to, though in a way which
certainly has no analogy to IP).  What is "typical" may depend a lot on
the particular set of protocols you consider.

Dennis Ferguson