cachedump output & rainbow tables

[shakenbake]

Originally posted here by cacosapo
Well, i've read MLF 1st poster and went to AO search facility and typed "password hash"


Veredict:
MLF was correct at first suggestion and you were very rude. Too bad for you.

Perhaps I was a little rude, I apologise for that everyone.

[morganlefay]

I know nothing....nothing I tell ya!!!

I was just pointing you to a resource.....

What I wasnt going to do is research it for you.....

Plus I had read most of you other vague posts....

Figured you just wanted to be spoon feed...like before

Full yet


MLF

[shakenbake]

>> Full yet

No ... I'd still like an informed answer from somebody!!

[cacosapo]

You see the bit I'm confussed over is that apparently cachedumps output is an MD4 hash, salted with the username.

Well, i maybe wrong (as usual) but cached passswords dumped by cachedump are current user password (hashed by MD4( password|U(username)). And this is the format that JTR use as input.

So i can imply if JTR is able to crack, a rainbow table will do the same job, right?

[sec_ware]

Hi

Certainly, cacosapo is right

Perform the following experiment:
a) get cachedump[1]

cachedump -v

extracts username:mscash.


b) get passwordspro[2]

Tools -> Hash Generator. Insert your password in "Text"
and add your username in the User name field.

Do the MScashs agree?

Now, how difficult would it be to create a table with a fixed salt, ie.
a fixed username? ...

Hence, there might be an advantage to rename the default administrator
name?

Cheers


[1] http://www.off-by-one.net/misc/cachedump-1.2.zip
[2] http://www.insidepro.com/eng/passwordspro.shtml

[Katmando]

Originally posted here by shakenbake
>> Full yet

No ... I'd still like an informed answer from somebody!!

Shakenbake,

I am of the thought that if I see a stupid post I just ignore it. It makes more since to me to spend less of my time by ignoring it than spend more time and effort cutting the poster down. But that is just me.

With that said you mentioned you read Irongeeks paper and maybe you missed this but this is what he has to say.
------------------------------------------------
Taken from http://www.irongeek.com/i.php?page=security/cachecrack

Fortunately from a security standpoint the way Microsoft hashes cached passwords is much more secure than the way they store local passwords in the SAM file. Since each cached hash has its own salt (a set of more or less random bits figured into the hash algorithm to help foil pre-computed attacks) cached passwords hashes take much longer to crack than LM (LAN Manager) hashes which don't salt the same way, are case insensitive and are split into seven character chunks.
------------------------------------------------

So it appears to me that the way MS salts this cache would make it difficult to create a precompute table (like rainbow table). What I am not sure about is if that difficulty is more or less not possible. What I mean is everything is possible but maybe beyond the time and effort required to it. One can brute force anything but it may take them 100 liftimes.


Good luck in your searches let us know if find any thing different.


Katmando

[sec_ware]

Hi

Katmando,

while it is true that MSCash is harder to crack, because one
cannot blindly use rainbow tables as for the (depreciated!)
LM-hashes[1],

it is not true, at least as far as my understanding of the MSCash[2]
goes, that the salt consists of a set of more or less random bits
(sorry irongeek). The salt is predefined, otherwise I would not have
obtained the same MSCash in my above mentioned experiment.

Hence, it is possible to create a rainbow table - however, one each for
every user. But on how many machines, the administrator is called
"administrator" in the workgroup "WORKGROUP"?

The required time to calculate twice an MD4 is not an obstacle to build
a rainbow table. The idea of the salt is to make it unfeasible to re-use
the same table over and over again, eg. for all usernames.

Cheers

[1] http://www.antionline.com/showthread...r=1#post828847
[2] http://www.antionline.com/showthread...r=1#post892906

[Katmando]

Originally posted here by sec_ware
[B]Hi

Katmando,

it is not true, at least as far as my understanding of the MSCash[2]
goes, that the salt consists of a set of more or less random bits
(sorry irongeek). The salt is predefined, otherwise I would not have
obtained the same MSCash in my above mentioned experiment.

Hence, it is possible to create a rainbow table - however, one each for
every user. But on how many machines, the administrator is called
"administrator" in the workgroup "WORKGROUP"?

Sec_ware,

Thanks for the input, very interesting.

So it seems plausible for certain usernames and domains (EX administrator/workgroup). But correct me if I am wrong grabbing the the cache is really only useful when a domain is used. Otherwise one would just grab the SAM. If it is only useful in a domain environment how many people actually use the default domain name in a production environment?

Also in Irongeeks defense he said it was harder not impossible.

But anyway it is interesting I would like to look into this further.

[shakenbake]

Thanks for the information and banter gentleman, the last few posts were much more like it.

I also apologise again for being a rude *****, yesterday was a particularly stressful one !