Home page logo

nmap-dev logo Nmap Development mailing list archives

Re: Nmap-based mapping/monitoring tool
From: Mike Slifcak <slif () bellsouth net>
Date: Tue, 09 Mar 2004 22:47:17 -0500

Times between ICMP echo  and ICMP echo reply are point-to-point
measurements, they may be routed differently outbound than return,
and they are critically dependent on where you measure _from_.

There may be intra-segment delays that you won't see with this

With one or two devices gathering packet flow information on each
segment, you can get less intrusive and possibly more accurate
measures of availability.  Setup with Linux boxen and Net-SNMP,
then poll at low frequency with MRTG.  You can get routes, packet loss
via framing errors and collisions, and MRTG is already good at
graphing.  But hey, that's a topic for another list %^)
All the Best,
-Mike Slifcak

Jake Kallman wrote:
To handle the actual slowdown and availability issue, we were planning (at
least for the first implementation) to just ping each machine with one or
2 packets and then read the time off of that.  Its a pretty primitive way
of doing it, but for a one semester project it is enough to prove the
concept anyways.  The idea being, the server runs once on a network to
establish the baseline stats for speed and network map, then on each
subsequent scan it compares the data from the current scan against the
information from previous scans.  If a computer is not there, then it
would report it as an outage, or if the time was significantly higher than
the average time it would report it as a slow connection.

In essence then for each computer on the network it would run a small ping
to test the connection and speed, then it would run nmap to get port
information and OS detection.  Pretty primitive I know, but it would at
least be enough to prove the concept for the class, with later work being
focused on making the process more efficient.

The main problem that we are working on with this model for connection
availability/speed tracking is dhcp.  If a computer changes IPs then it
might not be recognized by the system.  We are working on maybe trying
some solutions based around MAC addresses or intuitive guessing by the
server when it runs the scans, but that might be something we move off to
future work also (its hard trying to fit all this into a one-semester


On Tue, 9 Mar 2004, Mike Slifcak wrote:

The one thing I don't see ntop nor nmap being useful is in
finding network slowdowns or outages, as Jake describes those.

I've seen the utility of MRTG and Cricket, backed by routers
and switches that can be monitored for these purposes.

I'm willing to be educated, so let me know how you do this.
-Mike Slifcak

Jake Kallman wrote:

Thank you all for the very fast responses and great input =)

I'll take a look at ntop also.  One thing that I didn't mention in the
initial post, and which might be of interest to Bob, is that to ease the
load on a central server (and the network on the whole) we were going to
make the server script work in a distributed and parallelized manner.
Where at the top level of your network you could have a server that just
compiles data from lower-tier machines which each run scans on the
machines connected to them.  In this way the top level server does not
have to run any scans, but rather just get the database files from the
lower level machines at the scheduled intervals.  We thought that this
would cut down a lot of scan time on larger networks, and also allow some
scalability so that as the program was installed on larger networks one
machine wouldn't be burdened with the job of handling all the scanning.

Not sure if this helps out any, but just an idea I thought I'd throw out


On Wed, 10 Mar 2004, Craig Humphrey wrote:

Hi Jake,

you may also want to look at ntop, which monitors a network through sniffing
(rather than active scanning), but integrates nmap to probe hosts.

There are a few commercial tools out there that do this kind of thing (or
similar, or sub/supersets).
I've used What's Up Gold quite a bit.
Just did a little google and https://freemap.qualys.com/ looks interesting,
and being browser based, there's a certain amount of freedom for the client

Have fun and keep us in the loop.


-----Original Message-----
From: Jake Kallman [mailto:jkallman () unr nevada edu]
Sent: Tuesday, March 09, 2004 9:04 PM
To: nmap-dev () insecure org
Subject: Nmap-based mapping/monitoring tool

I am developing a network monitoring and mapping tool based
around nmap,
which will provide a graphical representation of a network
topology and
maintain a database of information about computers in that
network.  In
essence, it will take the output of namp, run at scheduled
intervals, and
compare that data against data from previous runs to try and flag
potential security and infrastructure problems.

The idea, at a high level is fairly simple, and in fact is a
little more
complicated than it needs to be since I'm doing this as a project to
complete my undergraduate degree in CS.  I'm writing a driver program,
which will sit on a network server somewhere, and will run nmap at
scheduled intervals on all computers in the network (which
I'm going to
try and optimize somewhat by allowing for multiple nmapping
servers in the
network so as to distribute the work as much as possible).  There will
also be a client application which will allow a user to
access this data
remotely (ideally I'm trying to create this client application to
allow users to log into the server from multiple platforms,
like PDAs and
cell phones, which might not be currently available, but when
I talked to
some network engineers in my area they said that it would be a great
feature). The client will access the data from the server program, and
create a graphical map of the network, showing any potential problem
areas. Ideally, I want to be able to flag network slowdowns
and outages,
newly enabled/disabled ports on machines, newly connected
machines (with
an eye toward being able to watch for unauthorized wireless
and things of that nature.

My question is whether or not this seems like a usable idea?
If not, then
what seems unfeasible about the design?

For help using this (nmap-dev) mailing list, send a blank email to
nmap-dev-help () insecure org . List archive: http://seclists.org

For help using this (nmap-dev) mailing list, send a blank email to
nmap-dev-help () insecure org . List archive: http://seclists.org

For help using this (nmap-dev) mailing list, send a blank email to nmap-dev-help () insecure org . List archive: http://seclists.org

  By Date           By Thread  

Current thread:
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]