Charsets to work on in distributed version

Moderator: BarsMonster

Post Reply
User avatar
BarsMonster
Site Admin
Posts: 1118
Joined: Wed Oct 01, 2008 7:58 pm

Charsets to work on in distributed version

Post by BarsMonster » Tue Dec 30, 2008 10:14 am

In the first distributed run over MD5, we need to decide on charsets

1) digists
2) az_ and space
3) azAZ_ and space
4) 0x00..0xFF 1-6 length, probably 7

5) Special chars: What special characters should we include?
6) Your ideas, local symbols, e.t.c

User avatar
the_drag0n
Posts: 217
Joined: Thu Oct 02, 2008 6:48 am

Re: Charsets to work on in distributed version

Post by the_drag0n » Tue Dec 30, 2008 10:18 am

id do it like that:

mixalpha 1-8 (maybe include "_-:.*!?"
numeric 1-12
all-space 1-7 (maybe even 1-8)
loweralpha + symbol 32 1-7 or 1-8

depending on how much power and time we have those 1-7 could also be 1-8.

User avatar
BarsMonster
Site Admin
Posts: 1118
Joined: Wed Oct 01, 2008 7:58 pm

Re: Charsets to work on in distributed version

Post by BarsMonster » Tue Dec 30, 2008 10:20 am

When talking about special characters, we need to include all symbols which might be easily entered on keyboard, and that should depend from country to country (My Chinese keyboard is totally crazy here :crazy:).

I am asking this because we should cover all current md5 hashes with these sets, as all these hashes would be checked together (i.e. during global keychecking you would not be able to add your custom charset), so I guess it is worth to add Cyrillic and Germal local symbols, e.t.c

As everyone working together, this allows us to check digits up to 16-17, not just 12 for example :-)

What I want to do is to have a 2-week plan for network capable to do 20-40bil keys/sec.
That means in 2 weeks we'll check 48'384 trillions (sorry do not know larger words in English :crazy:) of keys :crazy: :crazy: :crazy:

Irythros
Posts: 20
Joined: Mon Nov 10, 2008 10:00 pm

Re: Charsets to work on in distributed version

Post by Irythros » Tue Dec 30, 2008 11:08 am

48.3 Quadrillion.

As for char sets, I think leaving out asian characters would be better for speed.
German, russian, greek and latin letters would be of help to the standard:

az
AZ
10
~!@#$%^&*()_+-=[]\{}|,./<>?';:"
space

User avatar
the_drag0n
Posts: 217
Joined: Thu Oct 02, 2008 6:48 am

Re: Charsets to work on in distributed version

Post by the_drag0n » Tue Dec 30, 2008 11:09 am

how bout our lm frt keyset ?
Image
plus some chinese chars ? dunno what or how many they have.

User avatar
BarsMonster
Site Admin
Posts: 1118
Joined: Wed Oct 01, 2008 7:58 pm

Re: Charsets to work on in distributed version

Post by BarsMonster » Tue Dec 30, 2008 11:36 am

the_drag0n wrote:how bout our lm frt keyset ?
plus some chinese chars ? dunno what or how many they have.
frt looks nice.
Chinese have around 2000-5000 widely used characters, and around 50'000 total.
The same for Japanese, and Korean. Sure we'll skip asian ones :-D

User avatar
the_drag0n
Posts: 217
Joined: Thu Oct 02, 2008 6:48 am

Re: Charsets to work on in distributed version

Post by the_drag0n » Tue Dec 30, 2008 11:40 am

ok then that will do it. btw the charset is called frt-cp-850. but i have no idea where that name came from :P
ask neinbrucke he was the one who created it

edit: btw is the server side stuff already done ?

kiando
Posts: 64
Joined: Thu Oct 02, 2008 7:30 am

Re: Charsets to work on in distributed version

Post by kiando » Tue Dec 30, 2008 12:03 pm

Japanese uses about 3000 as a standard in newspaper.

User avatar
Rolf
Posts: 122
Joined: Fri Dec 26, 2008 10:48 am

Re: Charsets to work on in distributed version

Post by Rolf » Tue Dec 30, 2008 3:00 pm

My bruteforce attack charsets:
1. [0;6] mixalpha-numeric-symbol32-space
2. [9;12] numeric-space
3. [7;8] loweralpha-numeric-space

Of course, when the distributed client will be released, the maximum length bound can be raised.

Also, I encountered REALLY small amount of passwords, where cross-language characters are used( like russian+english ; or greek+english).Languages, where in addition to latin letters there are own special letters(Swedish, German etc. most languages) are an exception.

User avatar
the_drag0n
Posts: 217
Joined: Thu Oct 02, 2008 6:48 am

Re: Charsets to work on in distributed version

Post by the_drag0n » Tue Dec 30, 2008 3:28 pm

well i think those are exactly the hashes we are aiming for. bars distributed will not be to crack some loweralpha 1-6 stuff. its about bigger charsets. we are talking about a massiv amount of computing power which might depending on the number of users come to the power of a very very expensive supercomputer.

numeric 1-12 is just a matter of about 10 minutes even with my card (9600gt) so numeric 1-20 would be more realistic for bars distributed :)

asl
Posts: 28
Joined: Tue Dec 30, 2008 4:14 pm

Re: Charsets to work on in distributed version

Post by asl » Tue Dec 30, 2008 4:28 pm

abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789
!@#$%^&*()-_+=~`[]{}|\:;"'<>,.?/¡¢£¤¥¦§¨©ª«¬­®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ€

User avatar
the_drag0n
Posts: 217
Joined: Thu Oct 02, 2008 6:48 am

Re: Charsets to work on in distributed version

Post by the_drag0n » Tue Dec 30, 2008 4:32 pm

you should do a screenshot :P because depending on your pagefiles the symbols might variate.

asl
Posts: 28
Joined: Tue Dec 30, 2008 4:14 pm

Re: Charsets to work on in distributed version

Post by asl » Tue Dec 30, 2008 5:04 pm

Attachments
asl.PNG
asl.PNG (3.81 KiB) Viewed 9135 times

neinbrucke
Posts: 82
Joined: Sun Nov 02, 2008 8:53 pm

Re: Charsets to work on in distributed version

Post by neinbrucke » Tue Dec 30, 2008 11:09 pm

at least don't use the lm-frt charsets... they are meant for LM hashes, where unicode characters are first mapped to certain 'ascii' characters (actually just single byte chars). The mapping differs per OEM code page, so US english has another mapping then Latin (most western europian countries). Therefore the LM charsets contain strange bytes and lack certain often used characters, because they map to another byte with LM.

Sc00bz
Posts: 136
Joined: Fri Oct 03, 2008 8:28 am
Contact:

Re: Charsets to work on in distributed version

Post by Sc00bz » Wed Dec 31, 2008 9:31 pm

Since you're talking about MD5 we should probably do valid UTF-8 *bytes*.
00000000 to 01111111 (U+000000 to U+00007F)
10000000 to 10111111
11000010 to 11011111 (U+000080 to U+0007FF)
11100000 to 11101111 (U+000800 to U+00FFFF)
11110000 to 11110011 (U+010000 to U+0FFFFF)
11110100 (U+100000 to U+10FFFF)
128+64+30+16+4+1 = 243

Using this character set will generate invalid UTF-8 strings but it is better than 0x00 to 0xFF as it is 243 verses 256.

Corni

Re: Charsets to work on in distributed version

Post by Corni » Wed Dec 31, 2008 10:34 pm

why not just mixed-alphanumeric-space? just long enough ;)

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest