From: fwells12 (fwells12@xxxxxxxxxxx)
Date: Wed Feb 14 2001 - 04:27:01 GMT-3
I get what your trying to say here Fred, but I don't get your example too
well. The routers address space - 172.16.32.0-172.16.63.0 would be a /19.
If you had a requirement of a /22 network (say 172.16.32.0 -172.16.35.0, the
172.16.33.0/172.16.34.0 and 172.16.35.0 would overlap into the existing
space. What I don't get is why you would continue using spaced subnets. It
seems like it would make summarization not quite so granular because you
would now have to summarize that router for example by using something like
172.16.32.0 255.224.0.0 which is fine but it covers a huge amount of unused
address space. By using the 172.16.32.0-172.16.35.0 space for your /22
requirement and then going back to the original mask of /24 for the other
connected networks say 172.16.36.0, 172.16.37.0, 172.16.38.0 and 172.16.39.0
etc you could then summarize with a more granular area-range like
172.16.32.0 255.240.0.0 right?
Can you see any reason that would not work?
I guess I am expecting a lab requirement of something like: ' Summarize all
OSPF areas using summary routes which keep unused address space to a
minimum' etc.
Cheers
----- Original Message -----
From: Fred Ingham <fningham@worldnet.att.net>
To: fwells12 <fwells12@hotmail.com>
Sent: Tuesday, February 13, 2001 5:18 PM
Subject: Re: IP addressing thoughts...
> Frank: Let's say there is a requirement in the main network address
> space for a /22 network. The standard mask is specified as /24. Now
> lets take any router's address space
> 172.16.32.0 to 172.16.63.0. You could choose any networks in this range
> to satisfy the requirements.
>
> One choice would be 172.16.33.0, 172.16.34.0, 172.16.35.0, etc.
> as needed. What I recommend is that the subnets be spaced out
> such as 172.16.33.0, 172.16.37.0, 172.16.41.0, etc. In this way you
> will never have an overlap with the /22 network.
>
> Hope this clears this point up. If not, e-mail and I'll provide other
> examples.
>
> Fred.
>
>
> fwells12 wrote:
> >
> > I appreciate your feedback Fred. Solid advice. One thing I did not
quite
> > grasp was your statement 'I would space my subnets acccordingly for all
> > routers' would you be kind enough to elaborate on what you meant by it
> > please.
> >
> > ----- Original Message -----
> > From: Fred Ingham <fningham@worldnet.att.net>
> > To: frank wells <fwells12@hotmail.com>; <ccielab@groupstudy.com>
> > Sent: Tuesday, February 13, 2001 11:13 AM
> > Subject: Re: IP addressing thoughts...
> >
> > > Frank: The basic idea is sound but before you assign a given range to
a
> > > router I suggest you read the instructions for any fixed addresses.
If
> > > there is a required loopback of 172.16.225.0/30 on r2, for instance,
> > > your scheme falls apart.
> > >
> > > I suggest ranges of 32 rather than 16 (will work for 8 routers) and
also
> > > space out the networks to the lowest mask requested. For instance if
> > > there is a required network of /22 somewhere in the network I would
> > > space my subnets acccordingly for all routers. This guarantees no
> > > overlaps.
> > >
> > > There are no points in the lab for saving address space. You should
> > > complete your addressing before configuring the routers and also draw
a
> > > network diagram. Going into the lab with a fixed addressing scheme
can
> > > cost you time. Rather, you should be able to construct addressing
> > > quickly after reading the requirements. On rebooting - in the lab I
> > > recommend you do this often - there are no points for not rebooting
and
> > > it can
> > > clear up nagging problems.
> > >
> > > My opinion, Fred.
> > >
> > > frank wells wrote:
> > > >
> > > > I am giving a little thought to how I am going to handle the IP
> > addressing
> > > > in the lab. Regardless of the IP address/mask I am given I want to
come
> > up
> > > > with a scheme which allows me easy summarization of all routers and
> > their
> > > > connected OSPF network segments.
> > > >
> > > > I am almost certain Cisco will throw the VLSM-FLSm issues at me and
I
> > want
> > > > to be prepared.
> > > >
> > > > Lets say we get the following address/mask given to use: 172.16.0.0
and
> > we
> > > > are able to cut it up any way we want.
> > > >
> > > > Give me your thoughts on the following idea:
> > > >
> > > > 172.16.0.0 use mask 255.255.240.0 to get 14 useable subnets (not
> > including
> > > > subnet zero) I chose the 24 bit subnet mask to anticipate being
given
> > > > possible RIP/IGRP FLSM addresses to
> > > > deal with. I assign a contiguous range of subnets to each router
like
> > so:
> > > >
> > > > R1=172.16.16.0 - 172.16.31.0
> > > > R2=172.16.32.0 - 172.16.47.0
> > > > R3=172.16.48.0 - 172.16.64.0
> > > > R4=172.16.64.0 - 172.16.71.0
> > > > R5=172.16.72.0 - 172.16.87.0
> > > > R6=172.16.88.0 - 172.16.95.0
> > > > R7=172.16.96.0 - 172.16.111.0
> > > > R8=172.16.112.0 - 172.16.127.0
> > > >
> > > > Now when asked to do something like intra-area summarization to the
max,
> > I
> > > > can easily summarize each routers connected networks with a simple
> > command
> > > > like area 1 range 172.16.16.0 255.255.240.0 right?
> > > >
> > > > Also, by planning this way I would like to be able to address my
network
> > > > before I begin typing it into the routers. This would give me an
> > advantage
> > > > as I would be able to assign the highest IP addresses in the
respective
> > > > subnets to a/the loopback addresses etc. Ultimately I want to not
have
> > to
> > > > reboot the routers because I created an OSPF adjacency prior to
adding
> > the
> > > > loopbacks etc.
> > > >
> > > > I am also thinking about drawing a line on top of my picture and
> > deviding it
> > > > into as many segemnts as I have routers like so:
> > > >
> > > > R1 R2 R3 R4 R5 R6 R7 R8
> > > > |-------|-------|-------|-------|-------|-------|-------|-------|
> > > > 16 32 48 64 72 88 96 112
127
> > > >
> > > > By doing this I can easily mark off along my line which subnets I
have
> > > > already used too which may prove to be useful. Plus, it takes up
little
> > to
> > > > no room on your paper.
> > > >
> > > > Hope the ASCII pic isn't mangled too bad...
> > > >
> > > > Good idea/bad idea?, thoughts please.
> > > >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:28:48 GMT-3