On Fri, April 22, 2005 9:53 am, Douglas Royds said:
> Steve Holdoway wrote:
>> Douglas Royds wrote:
>>> If the log-in mechanism allowed one log-in attempt per second, it
>>> would take almost 4 years to cover them. You might get lucky and crack
>>> it in a few months. But only if the log-in allowed one attempt per
>>> second indefinitely. So this is where Microsoft - and the open source
>>> community - can prevent brute-force attack - simply limit the rate at
>>> which attempts can be made.
>>
>> And if I'm attacking in parallel - is that still 1/sec? The login
>> routine includes an exponential increase in delay time for each
>> incorrect password, so it's pointless to try more than once.
>
> Which log-in routine, sorry?
I'd hope that all of your authentication went through pam, so any!
http://www.vsl.gifu-u.ac.jp/freeman/misc/pam-0.72/ps/pam_appl.ps
>
>> And do I need to wait until it's complete until I try again? If I'm
>> using all my (brute) force to get in, I will be doing both.
>
> Both which?
Both attacking in parallel and not waiting for a response. And, of course,
attacking on ssh, http, dns, snmp, smtp, and all the other services that
you may have running concurently as well.
>
>> (Does this footer have any legal standing?)
>
> Enough about the footer, thanks.
>
> Douglas.
>

... the average md5 password can be cracked in 30ms...
http://linuxexposed.com/Articles/Hacking/Password-Cracking-and-Time-Memory-Trade-Off.html
http://www.linuxexposed.com/Articles/Hacking/Unix-Attacking-Techniques.html
http://www.antsight.com/zsl/rainbowcrack/

Just to get started - and this is just the published stuff. I'm no expert,
but I guarantee I'll be getting *lots* more than 1 attempt / second!

( Not that I have the slightest interest in proving it )


Steve.
-- 
Windows: Where do you want to go today?
MacOS: Where do you want to be tomorrow?
Linux: Are you coming or what?

Reply via email to