Re: CIDR on IP hosts

Lars Poulsen <lars@rns.com> Wed, 16 October 1996 18:10 UTC

Received: from cnri by ietf.org id aa25429; 16 Oct 96 14:10 EDT
Received: from nico.aarnet.edu.au by CNRI.Reston.VA.US id aa18299; 16 Oct 96 14:10 EDT
Received: from RNS.COM (SPECTRUM.RNS.COM [131.143.16.1]) by nico.aarnet.edu.au (8.6.10/8.6.10) with SMTP id DAA02750 for <cidrd@iepg.org>; Thu, 17 Oct 1996 03:13:24 +1000
Received: from cab.RNS.COM by RNS.COM (SMI-8.6/SMI-4.1(Spectrum)) id JAA02165; Wed, 16 Oct 1996 09:55:15 -0700
Message-Id: <199610161655.JAA02165@RNS.COM>
X-Sender: lars@azure.rns.com
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 16 Oct 1996 10:03:50 -0700
To: Benny Rodrig ESD-Tel <brodrig@madge.com>
From: Lars Poulsen <lars@rns.com>
Subject: Re: CIDR on IP hosts
Cc: cidrd@iepg.org
X-Mailer: <Windows Eudora Version 2.0.2>

>Does anybody know how users typically configure their hosts, when 
>allocated CIDR blocks of contiguous class C networks? Are the hosts 
>configured with a classless netmask (e.g. 193.1.1.1, 22) to form one 
>supernet, or is the net configured as multiple class C nets connected to 
>each other by routers? 

Typically, users don't choose how to configure their hosts, network
managers configure the hosts for the users. The hosts are configured
to reflect the layout of the network.

An organization with enough hosts to reasonably use a /22 address range,
(300-800 hosts ?) will typically need a segmented network architecture.
The 10 bits of local address might typically be 4 bits of subnet-ID
and 6 bits of local address.

This type of question might more appropriately be asked on the
comp.protocols.tcp-ip USENET group.

/ Lars Poulsen            lars@RNS.COM    +1-805-562-3158
  RNS / Meret Communications
  7402 Hollister Avenue
  Goleta, CA 93117