Re: distributed nmap?

[D.R.Tzeck <drt () ailis de>]

"Aaron D. Turner" <aturner () pobox com> writes:

> On Sat, 18 Mar 2000, Arturo Busleiman wrote:
>
> > hi!
> >
> > Why not adding a --agent x.x.x.x [port] parameter?
> > It would turn nmap into an agent, and the 'boss' client would be running
> > at x.x.x.x (port [port] if specified)
> >
> > It would be easier, the boss client would be nmap --boss n
> > it then would sit there waiting till 'n' agents connect, then allowing to
> > enter scan options/targets, send them to each agent (of course, it would
> > distribute the port range among them!!). 
> >
> > what do you think of this?

> Personally, I would prefer a simple client/daemon wrapper for nmap.  That
> would provide an easy to maintain layer of abstraction between nmap and
> the means of communication.  One could write such an animal in a few hours
> with Perl which would be almost as portable as a C app. 

You *can* do that in a few minutes using ucspi-tcp by Dan Bernstein
and the "commandline input" mode of nmap 2.13ß. This is certainly
better than adding the bloat of "--agent x.x.x.x [port]" to nmap.
The question here is: Do you want to do it?

For maximum effect you have to run nmap as root. Giving the Network access
to a programm running as root is generally a bad idea. nmap is not
designed to sit on a security boundary. Nobody wants to see messages
like "root exploit for nmap 2.4.5" on Bugtraq.

Distributed Portscanning is a nice Idea. Something like
distributed.net showing open ports on the Whole Internet (in
realtime?) would be nifty. People (including me )are working on this -
but you can't do that by yust adding another feature to nmap.

Putting all features you might ever need for anything into a single
tool is not the way unix works.

drt

-- 
finger drt () ailis de for OpenPGP Key 0x3E7222DD - http://rc23.cx/

> You appear to be absolutely incapable of realising that there are
> people in this world who can see more than one side to a question...
On the contrary. I see both sides, and I have evaluated both sides,
and I have found that one side is vastly superior to the other. This
may seem ruthless, but that's how engineering works.
                                 Daniel J. Bernstein, comp.security.unix