john-users - Re: Can Excessive Rounds make Password cracking Infeasable

Follow @Openwall on Twitter for new release announcements and other news

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1337883385.58625.YahooMailNeo@web120704.mail.ne1.yahoo.com>
Date: Thu, 24 May 2012 11:16:25 -0700 (PDT)
From: NeonFlash <psykosonik_frequenz@...oo.com>
To: "john-users@...ts.openwall.com" <john-users@...ts.openwall.com>
Subject: Re: Can Excessive Rounds make Password cracking Infeasable

Hello Brad,

You have started a good discussion. I believe, by number of rounds you mean number of iterations?

If so, yes, by increasing the number of iterations in a hashing algorithm, the resultant hash becomes stronger. However, the function performed by the hashing algorithm in each iteration need not be the same.

For instance, in the case of MD5-crypt, of the 1000 iterations that take place, during each iteration the function performed by the hashing algorithm is different based on the iteration counter.

Blowfish hashes provide you a way to decide the number of iterations as well.

$2a$<logarithm 2 of the number of iterations>$.........

However, what you state below, the number of iterations look extremely high (391939).

While implementing a cryptographic hashing algorithm, besides its strength the computational feasibility also needs to be kept in mind.

May I ask you, what is the distro of Linux you are using?

I hope other experienced people on this mailing list would share their ideas as well.

Thanks.

________________________________
 From: Brad Tilley <brad@...ystems.com>
To: john-users@...ts.openwall.com 
Sent: Thursday, May 24, 2012 11:36 PM
Subject: [john-users] Can Excessive Rounds make Password cracking Infeasable

This is slightly off-topic as it does not specifically relate to John use,
but I wanted to ask the opinions of others here. When do rounds make
password cracking infeasible, or do they? For example, the hash below is a
SHA-512 hash with 391939 rounds applied. You can actually feel the delay
at logon (about 2 seconds on newer machines):

test:$6$rounds=391939$UqhsyLSZ$F/K1CGpBf9yefYXCRbY5uK/LW1HzW8EiPCzdq8PMVvZ4JLhb4F464ps87MX/YwYEI0s62KIsnZBuCt45a.A4I0:1002:1002::/home/test:/bin/sh

The source code of sha512-crypt.c sets this as the maximum number of
rounds so Linux sys admins could configure this number even higher:

   /* Maximum number of rounds.  */
   #define ROUNDS_MAX 999999999

So long as the passwords are sufficiently complex and users can't select
simple words such as 'password' for their password, I would think that
these hashes are close to un-crackable (certainly not in a reasonable time
period anyway). What do other John users think?

Thanks,

Brad

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.