Thanks for your nice reactions, I hear a lot of interesting ideas! I have a
few follow-up questions:
If you realy have connected a Win2k DC to the internet, without
firewall, system lockdown or disabling netbios/file sharing <snip>
Even worse, I've enabled Terminal services so anyone can try to log in - and
a some people try.
It's a sacrificial machine, when I get bored of the exercise I'll just
reinstall it. In the meantime I learn a lot about logging and log-analysis.
I've enabled any (Win2K native) log-option that I know of and I still don't
feel quite satisfied. I have the auditing options set according to MS
guidelines, I'm logging web-access, FTP-access and even the DNS-debug
options. I have enabled "failed" file access to my whole C drive (doesn't
seem to generate a lot of entries though ...). I have the network-monitor
running all the time, so I can see what people are trying to do.
I don't have the feeling that someone is inside the machine (yet). No weird
log entries. No new files or strange processes. No unexplainable network
traffic.
Question 1: What would be the signs of a succesful hack?
Question 2: I know about "snort" and such tools, but first I want to see
what a Win2K machine can log on it's own. Anything else I could enable?
I may be wrong, but I have the feeling that an up to date, patched Win2K
machine is quite resilient on it's own. I've run nessus against the box and
I've still 7 holes and 25 warnings to go, but most of them are just DoS
vulnerabilities. Nothing someone could hack me with (I hope).
Question 3: Am I totally wrong?