From owner-freebsd-questions Tue Apr 15 10:40:00 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id KAA24303 for questions-outgoing; Tue, 15 Apr 1997 10:40:00 -0700 (PDT) Received: from d2si.com (macbeth.d2si.com [206.8.31.2]) by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id KAA24260 for ; Tue, 15 Apr 1997 10:39:42 -0700 (PDT) Received: (from alec@localhost) by d2si.com (8.8.5/8.8.5) id MAA25341; Tue, 15 Apr 1997 12:39:16 -0500 (CDT) From: Alec Kloss Message-Id: <199704151739.MAA25341@d2si.com> Subject: Re: Subnets of all 0's/all 1's In-Reply-To: from Guy Helmer at "Apr 15, 97 10:50:30 am" To: ghelmer@cs.iastate.edu (Guy Helmer) Date: Tue, 15 Apr 1997 12:39:16 -0500 (CDT) Cc: freebsd-questions@freebsd.org X-Mailer: ELM [version 2.4ME+ PL31 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Guy Helmer is responsible for: > I'm helping a FreeBSD system administrator whose class C network is > subnetted at 255.255.255.192. He would dearly like to use the subnet with > systems numbered x.x.x.1-63; I have held up RFC 950 to say "this isn't > allowed", but the RFC doesn't say specifically why this wouldn't work. > > Should FreeBSD be able to support a network with a subnet of all zeros or > all ones? If not, could someone give a short technical explanation as to > why? > > Thanks, > Guy Helmer > > Guy Helmer, Computer Science Grad Student, Iowa State - ghelmer@cs.iastate.edu > http://www.cs.iastate.edu/~ghelmer > I did not know that this isn't allowed. I checked things out and RFC certainly does imply that this is a no-no. In RFC 950, RFC 943 this referenced as justification for this. RFC does not seem to use imperitive language (you shall or you must) but something much less harsh, as in "When called for, the address zero is to be interpreted as meaning "this", as in this network." This policy seems a bit strange to me. For example, take a subnet mask 255.255.255.128 on a class C network (ie two sub-networks). By the above rule, addresses xxx.xxx.xxx.0-127 would all be illegal. So, you've just cut you available hosts in half for no reason. Basically, this rule makes it illegal to only chop a network in half, which seems strange to me. RFC 950 doesn't cover (as near as I could tell) subnetting a subnet. For example, say I've subnetted my class C (in fact, I have) with netmask 255.255.255.224, 8 networks of 30 hosts each. Now, I realize that my servers (had I more than one) are eating up a bunch of bandwith talking with each other, so I want to isolate them on their own network. Say I take all the hosts in the range xxx.xxx.xxx.64 to xxx.xxx.xxx.95 and set their netmasks to be 255.255.255.240, 16 networks of 16 hosts each. With proper configuration, I expect this would work with existing software. Well, I already use hosts on the "reserved" subnet in the range 1-31 without any problems. Anyone out there have a definitive answer?