I worked in a finantial institution where I had around 65 000 000 of potential
IP adresses. I mean "potential" IP adresses: many of them where empty, others
with just a couple of machines, and very few had all the IP adresses given to
physical machines.
To scan this (there wer actually just classes C), it took me 5 days & nights
for covering all the IP adresses, and we were just nmaping (simple nmap -sS)
all of these IP adresses.
Just notice that this tuming depends a lot on the traffic on your subnets: the
more traffic there is, the slower will be your scans ... ;)
-----Message d'origine-----
De : Tim [mailto:tim-pentest@sentinelchicken.org]
Envoyé : jeudi 10 février 2005 03:38
À : Moonen, Ralph
Cc : pen-test@securityfocus.com
Objet : Re: Mapping Class A network ( any easy trick?)
Apart from the timing issues (I agree totally) I still think that you
cannot call nmap+nbtstat or nmap+nessus or whatever combinatioin
'penetration testing'. To me that is Vulnerability Analysis. VA on a
class A is tricky enough. Pentesting (ie attempted exploitation of all
discovered vulenrabilities) on a full class A is extremely difficult.
Impossible for 1 man and his laptop, given average population of
networks.
Of course not. I was merely trying to provide some information on what it
takes to actually scan that many IPs. Our scan only consisted of OS
fingerprinting, about 20 TCP and 20 UDP ports scanned, and a quick nbtscan
query. This would be enough information to map a network where one truly
doesn't know what is out there, but by no means is it a "penetration test".
Just the first step (in some situations).
tim
