RE: MD5 Collisions and Evidence Integrity

> From: ¥ dosman ¥ [mailto:dosman33@hotmail.com] 
> Subject: Re: MD5 Collisions and Evidence Integrity
>
> Actually MD5 hasn't been broken... yet. A close cousin to MD5 was broken, 
> not the actual MD5 as we know it.


Not true, the earlier collisions did use different initialization values than 
the MD5 spec, but the recent collisions use true MD5 init values. See 
http://www.md5crk.com/md5col.zip for two differing 1024 bytes files that yield 
the same MD5 hash... 

The key is that, currently, while it is possible to create two files with the 
same MD5 hash, you cannot predict what the contents of those files will be. 
More importantly, when you hash a drive or file, there is currently *no way* to 
create another file that has the same hash... So, someone can't make changes to 
a file/drive in a manner that gives the same hash as the original...

Current recommendations are:
1) Continue using MD5, but understand exactly what type of collision was found 
and what it means, so you don't get caught with your pants down by opposing 
counsel... 
2) or, use two hash algorithms, such as MD5 and SHA1. I created the 2hash 
program (http://www.crossrealm.com/2hash) to do this in parallel so it doesn't 
double your hashing times... 
3) or, migrate completely away from MD5 to SHA1.


T

-----------------------------------------------------------------
This list is provided by the SecurityFocus ARIS analyzer service.
For more information on this free incident handling, management 
and tracking system please see: http://aris.securityfocus.com