Just trying to see if anyone has dealt with this software before.

This is for a ch*ld p*rn cases, where the subjects told his family the police won't
find anything.

Thank you advance for any info

Rick ******

Dan

IP:

shelly@********.com

IP:

Hmmmm... Who actually reads this board? I'd love to know :-o

shelly@********.com

IP:

One noticeable weakness is that ScramDisk's sector scrambling is based on 1024 random bytes found in the first 2 sectors of a ScramDisk volume. These sectors and the sector number are passed through a function which produces a 128 bit value unique to that volume and sector. This value is then used as the whitening and CBC IV values for the actual encryption.

Therefore, each sector in a ScramDisk volume has it's own constant IV, which is unique to that volume and sector. But it also means that no matter what data that sector contains, the IV and whitening values are the same.

ScrammDisk is also extremely vulnerable to dictionary attacks... wrote a good explanation of how ScrammDisk works, which is as follows:

1.load the 20 byte SHA-1 hash as the cipher key

2.decrypt 2048 bytes from sectors#0-3

3.load up to 32 bytes from 1024-1055 as the cipher key

4.decrypt sector#16

5.decrypt sector#17

6.if sector#16 NOT equal sector#17 then goto the last step

7.decrypt sector#18

8.if sector#18 NOT equal sector#17 then goto the last step

9.the password is correct, end

10.repeat from step 1 for all remaining ciphers

In step 1. 2048 bytes are decrypted, but at attack can take advantage of the fact that only 1056 bytes need to be decrypted. These 1056 bytes comprises a sector IV table of 1024 bytes, and the disk key of 32 bytes.

The next steps normally decrypt and compare complete sectors (512 bytes), however an attack need only decrypt one block (for most ciphers 8 bytes or 64 bits) for each step after step 1.

This results in an attacker needing to do exactly 132 decryption's per password test cycle for a 64 bit block cipher.

[This message has been edited by ggottfri (edited 10-28-2000).]

IP:

Thanks
shelly@********.com

Aha! It is as I thought,
...
ScrammDisk is also extremely vulnerable to dictionary attacks...

IP:

IP:

IP

It is taking longer and costing more to develop than I had anticipated because I am adding extra features (encrypted container recognition, password morphing, support for multi processors etc). I now anticipate that the tool will be available in mid January and I would like to get a feel for how many people would be interested in having a copy.

I normally only offer a no crack-no fee service but, because the majority of cases involve files containing CP, it makes it impossible for me to recieve evidential files via email without breaking the law.

Therefore, if enough people are interested in using the tool, I may consider licensing it to recover some of my costs.

Anyone interested, please contact me via email at shelly@********.com

IP:

Hi, I officialy represent Scramdisk (Shelly did ask who read this forum ).

We're interested by all your comments, but are amused to here that "ScrammDisk is also extremely vulnerable to dictionary attacks..." in view of the fact that greater than 100 decryptions per trial passphrase *per cipher* are necessary to test a key...Do you guys really consider >800 decryptions (SD currently supports 8 ciphers) per key to be "vulnerable"?

The situation will improve in future...We will be adding further ciphers (AES256 etc) and also additional hash functions (SHA256, RIPEMD160) - this will mean each passphrase will need to be tested against 27 cipher /hash combinations - which will mean greater than 2700 decryptions per passphrase.

Please also note that Scramdisk v3.x containers are NOT detectable from random files (unless Shelly knows something we don't - this "feature" is a problem specifically linked with SD v2.x.

We will be posting an "official" response to these points on the Scramdisk site in due course.

Regards,

Sam

IP:

....

Sam@ScramDisk - while I definately agree that 2700 decryptions is very cool, and most likely will prove a real hassle for desktops, against our Sun Enterprise 6500s (running together as one) the 2700 number probably won't be much more than a speed bump.
Please do keep us informed though of the new release!

Happy Thanksgiving All!!

farmerdude

IP:

You said "against our Sun Enterprise 6500s (running together as one) the 2700 number probably won't be much more than a speed bump."

Unfortunately, your 6500 are fighting against combinatorial mathematics

Any (reasonable...) non-dictionary passphrase with more than 10 characters or so will be totally unbreakable, no matter how many 6500 you chuck at it, the passphrase is going to last a long time.

Brute force is not the way to try to break Scramdisk....And we don't know of any backdoor or shortcut mechanism either. Still, most users probably aren't all that good about selecting passphrases

Regards,

Sam

IP