🚀 go-pugleaf

In article <2ldkpk$sq5@dumbcat.sf.ca.us>, marc@dumbcat.sf.ca.us (Marco S Hyman) writes:
> In article <CM8Kny.8Ho@eecs.nwu.edu>, H. Peter Anvin N9ITP <hpa@nwu.edu> wrote:
>  > > PS:  Some implementations do permit one-bit subnet ids, but this is non-
>  > > conformant.
>  >
>  > No, they are not.  Any application that does *not* permit a one-bit
>  > subnet id is nonconformant.
>
> RFC 1009 states "No subnet should be assigned the value 0 or -1 (all one
> bits)".  See second paragraph of page 6. Has this been superseded by a later
> RFC?

No.  The current "Requirements for Internet Hosts -- Communication Layers"
(RFC 1122), which supplements, clarifies, and/or supersedes the primary
protocol documents, states in section 3.2.1.3 on addressing on page 31:

            IP addresses are not permitted to have the value 0 or -1 for
            any of the <Host-number>, <Network-number>, or <Subnet-
            number> fields (except in the special cases listed above).
            This implies that each of these fields will be at least two
            bits long.

The purpose of RFC 1122 (actually, of any of the "standard" RFCs) is to
help vendors and implementors avoid problems of lack of interoperability,
bad performance and/or robustness.  So there's nothing wrong with an
implementation allowing the use of subnets 0 and/or all-ones *as an option*,
especially if there is a big demand for it, as there is in this case in
these days of increasingly scarce IP addresses.  However, to allow it by
*default* would be a bad idea, since

o   commonly-used "classful" routing protocols (e.g., RIP) can't distin-
    guish a network route from a subnet route for subnet zero, and

o   some hosts may confuse IP addresses on the all-ones subnet with
    broadcast addresses.  True, the code would have to be pretty sleazoid
    to do this, but any confusion about broadcast addresses is potentially
    disasterous.

Another poster mentioned varying-length subnet masks.  This is a separate
issue, but they also require classless routing protocols like OSPF.  Often
one can workaround these kinds of problems by using static routes or assigning
multiple subnets on the same "wire."  Cisco routers (and probably other brands)
have had the latter capability for quite some time -- see the description of
"secondary address" in the Cisco docs.

I hope this helps.  I wanted to direct followups to comp.protocols.tcp-ip
but my news software wouldn't let me, too many x-posted groups already...
--
James Harvey   harvey@iupui.edu   IUPUI OIT Technical Support/VMS/Unix/Networks
Disclaimer:  These are my own opinions.  I do not speak for Indiana University.
"Contrariwise," continued Tweedledee, "if it was so, it might be, and if it
were so, it would be; but as it isn't, it ain't.  That's logic!" -Lewis Carroll