Re: Reduce zenmap topology noise (patch)

[Colin L Rice <ricec2 () rpi edu>]

On 11-12-29 06:26 AM, Patrik Karlsson wrote:
> On Wed, Dec 14, 2011 at 1:24 PM, Brahim Sakka<brahim.sakka () gmail com>wrote:
>
> > I've stumbled upon this several times and was looking for a way to
> > reduce the noise you described. Thank you for this patch and I hope it
> > gets committed to Nmap.
> >
> >
> > 2011/12/14, Anders Sundman<anders () 4zm org>:
> > > Hi,
> > >
> > > When running many traceroutes that traverse the same "anonymous" hop
> > > between two known hosts, zenmap creates a new node in the graph for each
> > > traceroute. This introduces a lot of unnecessary clutter in the graph.
> > > 200 traceroutes passing 2 (physical) anonymous devices currently
> > > generates 400 nodes in the topology graph.
> > >
> > > Sure, it might be a new physical anonymous device each time, but it
> > > most likely isn't. Even if it is, I don't think it makes sense to
> > > represent them as two distinct nodes (since there is no way to tell them
> > > apart).
> > >
> > > I suggest combining nodes in the graph that represent anonymous hops
> > > between two known devices.
> > >
> > >            / anon_1 \
> > > ..->  ip_a          ip_b ->  ..  =>  .. ->  ip_a ->  anon_1 ->  ip_b ->  ..
> > >            \ anon_2 /
> > >
> > >
> > > Here is a patch implementing this behavior:
> > >
> > > http://www.4zm.org/files/2011/integration.py.diff
> > >
> > > Regards,
> > > Anders
> > > _______________________________________________
> > > Sent through the nmap-dev mailing list
> > > http://cgi.insecure.org/mailman/listinfo/nmap-dev
> > > Archived at http://seclists.org/nmap-dev/
> > _______________________________________________
> > Sent through the nmap-dev mailing list
> > http://cgi.insecure.org/mailman/listinfo/nmap-dev
> > Archived at http://seclists.org/nmap-dev/
> Thank's for the patch Anders. I don't use Zenmap myself, but based on your
> explanation it sounds like a usable patch.
> Could some Zenmap guru take a look at this patch so that we can make a
> decision whether to commit it or not?
>
> Thanks,
> Patrik
I'm no guru but I've done some zenmap patch work. The obvious thing it 
is missing is some testcases. I don't remember if we have built them 
into radialnet but we have them in zenmap and It'd be nice if we kept 
them maintained.

The logic looks sound. I think we can use
post_hop = find_hop_by_ttl(hops, ttl+1) if ttl < max(ttls) else None
instead of
post_hop = find_hop_by_ttl(hops, ttl+1) if ttl < len(hops) - 1 else None
just so it is prettier.

-Colin

_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/