SubNetIO vs. CampusIO (was "Re: ? or two")

Date view Thread view Subject view Author view

Subject: SubNetIO vs. CampusIO (was "Re: ? or two")
From: Dave Plonka (plonka@doit.wisc.edu)
Date: Wed Apr 26 2000 - 14:02:10 CDT

Dave,

I hope you don't mind that I'm sending this response to the "flowscan"
mailing list.  In general, it'd be a great help to me if FlowScan users
would ask such questions in this public forum, as the answers will be
archived and they are obviously pertinent to other users and,
unfortunately, I don't have the time to answer them many times.

On Wed, Apr 26, 2000 at 11:55:06AM -0600, Dave Packham wrote:
> With 4 class b's here at the U would it be really hard to get SubnetIO.pm to
> work

We're in the same situation - you do have records of how your subnets
have been delegated, and who uses each one right?  Using a script, you
can massage that into the ".boulder" format periodically, and restart
flowscan to get it to pick up the changes.  That's what I do.  (If not,
this can often be derived by parsing the addresses and desc$iptions
from your Cisco router configs.)

> and what would I gain?

Well, you'd get per-subnet statistics for flows, pkts, and bytes of
campus I/O.

This data is invaluable if you want to consider usage-based billing fo
Campus departments.  Also, it allows you to answer questions like what
percentage of your precious, costly bandwidth is being used by which
population of users... which is info you'll need to do traffic shaping,
policing, ans QoS policies in the near future as applications will
quickly consume all available bandwidth.

If you think that statistics per-class-B network are sufficient though,
then you could just use CampusIO rather than SubNetIO.  As a point of
comparison, our mgmt. has been most interested in the residence-hall
network (ResNet), which is a whole class B for us.

> Would rrgrapher still work and I would just get
> a rrd file for every subnet?

Yup, exactly.
 
> Not much Doc info in the SubnetIO.pm

> Flow-cache was turned off on one of our core routers  DOH!   Had to shoot
> someone.  Our files are now in the 10-15 meg size per snapshot.

Ya, it takes some time to get the router-heads to understand that
measurment/monitoring will quickly become a "production" service that
needs to keep working just like the packets need to be kept flying.

Later,
Dave

-- 
plonka@doit.wisc.edu  http://net.doit.wisc.edu/~plonka  ARS:N9HZF  Madison, WI

--
Help        mailto:majordomo@net.doit.wisc.edu and say "help" in message body
Unsubscribe mailto:majordomo@net.doit.wisc.edu and say
"unsubscribe flowscan" in message body
Archive     http://net.doit.wisc.edu/~plonka/list/flowscan/archive/

Date view Thread view Subject view Author view

This archive was generated by hypermail 2b25 : Wed Apr 26 2000 - 14:04:42 CDT