Rainbow Tables, Few Considerations

Full Version: Rainbow Tables, Few Considerations

GovernmentSecurity.org > The Archives > General Network Security

Neo_

Mar 6 2004, 03:03 PM

I've tried raimbowcalc.exe to calculate time for generating rainbow tables

Well ... disappointed i am

C:>rtgen lm loweralpha-numeric 1 7 0 2400 40000000 all

I want to have abcd....z012..9, so loweralpha-numeric for lm...

Nec PIII 1.2 Ghz :
I read : "100000 of 40000000 rainbow chains generated (7 m 39 s)"

So i've calculated :
1week 3days : 1..7 char : 3,05 GB : 5 tables : 99% success
24weeks 2days : 1 .. 8 char : 48,83 GB : 80 tables : 92% success
607weeks 1day : 1 .. 9 char : 1220,7 GB : 2000 tables : 84% success

PIV 3 Ghz :
I read : "100000 of 40000000 rainbow chains generated (4 m 48 s)"
6days 16hrs : 1..7 char : 3,05 GB : 5 tables : 99% success

If i've not made any mistake, i think that i'm going to generate 5 tables with 1..7 char

mrBob

Mar 6 2004, 03:57 PM

hmm, what does that succes-rate mean?
and who would generate up to 1TB rainbow tables, lol

FiNaLBeTa

Mar 6 2004, 04:10 PM

for lm i think you onely need 1-7 cuz windows splits up passwords longer then 7.
So that would be 5 tables.

linux_dude

Mar 6 2004, 05:53 PM

That's only if we're talking LM hashes here. If someone forces their system to store only NTLM hashes in the SAM, then you're stuck with infinite length md5 options.

DaClueless

Mar 6 2004, 06:15 PM

QUOTE (linux_dude @ Mar 6 2004, 05:53 PM)

That's only if we're talking LM hashes here. If someone forces their system to store only NTLM hashes in the SAM, then you're stuck with infinite length md5 options.

MD5 is a cyclic encryption, what that means is there can be millions of ways to get the same MD5 HASH. So MD5 is finite, just big. 2^128 big.

daguilar01

Mar 6 2004, 08:44 PM

QUOTE (Neo_ @ Mar 6 2004, 08:03 AM)

PIV 3 Ghz :
I read : "100000 of 40000000 rainbow chains generated (4 m 48 s)"

hmm, your P4 takes 4min 48 secs to make one chain?
i made the alpha numberic tables on my Athlon XP 2200+ underclocked,

, @ 1.55 ghz and it took ~4min 25 secs

pretty big processer differnce, guess rgen isnt too good on P4s

Neo_

Mar 6 2004, 08:53 PM

QUOTE (daguilar01 @ Mar 6 2004, 08:44 PM)

QUOTE (Neo_ @ Mar 6 2004, 08:03 AM)

PIV 3 Ghz :
I read : "100000 of 40000000 rainbow chains generated (4 m 48 s)"

hmm, your P4 takes 4min 48 secs to make one chain?
i made the alpha numberic tables on my Athlon XP 2200+ underclocked,

, @ 1.55 ghz and it took ~4min 25 secs

pretty big processer differnce, guess rgen isnt too good on P4s

Too many things were running

saetji

Mar 6 2004, 11:46 PM

QUOTE (niemic @ Mar 6 2004, 06:15 PM)

QUOTE (linux_dude @ Mar 6 2004, 05:53 PM)

That's only if we're talking LM hashes here. If someone forces their system to store only NTLM hashes in the SAM, then you're stuck with infinite length md5 options.

MD5 is a cyclic encryption, what that means is there can be millions of ways to get the same MD5 HASH. So MD5 is finite, just big. 2^128 big.

Actually MD5 was designed such tat for each hash, there would only be a max of 5 plaintexts - its generally agreed amongst the mathematical community that the level of encryption is nearly one to one with the odd fallout here and there ... if ur interested that is

linux_dude

Mar 7 2004, 12:42 AM

Can we just agree on the fact that if it's LM hashes, then you're set with tables of 7 chars, and if it's only NTLM hashes, then that's not gonna cut it?

This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.