[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Performance: MD4 vs MD5

From: Daniel Berlin <dberlin_at_dberlin.org>
Date: 2003-04-26 04:59:49 CEST

On Friday, April 25, 2003, at 10:46 PM, Michael Price wrote:

> Daniel Berlin wrote:
>> On Friday, April 25, 2003, at 04:19 PM, Michael wrote:
>>>> Where did you read that collisions with MD4 are that frequent?
>>>> I'm aware of cryptographic attacks against the first few rounds
>>>> but do not recall reading _anything_ that states MD4's collision
>>>> rate is that high.
> >
>> "Dobbertin has shown how collisions for the full version of MD4 can
>> be found in under a minute on a typical PC. Clearly, MD4 should now
>> be considered broken."
>
> I agree. It is broken for cryptographic purposes. However, this
> discussion doesn't relate to cryptography in the slightest. We simply
> need a hash function and it our case a collision isn't really that > bad.
>
Yes it is, for the speed it runs at.

>> If you want collision rates, see
>> http://www.cs.duke.edu/~anderson/hashing/
>
> Seen it. What's your point?

That MD4 is substantially worse as a checksum than MD5, for 10-25%
speed, cryptographic or not.

Why do we need this speed again, remind me?
Have you any data that we are spending a significant amount of time
checksumming things?
Or is this just a waste of time/design masturbation?

>
>> If you do a google search on "md4 collision rate", you'll see things
>> like "(S/Key was originally just MD4, which proved to have a high
>> collision rate)"
>
> I know its broken for cryptographic purposes. This discussion doesn't
> relate to cryptography.
>
So why are you trying to imply i've brought it up?
It's high collision rate compared to MD5 makes it not a good idea for
our purposes.
Even CRC-32 is a better idea than MD4.

>> Also see http://burtleburtle.net/bob/hash/doobs.html
>
> Seen that too. What's your point?
>

It's very slow for sucking badly, too.
At least MD5 is good.

>> Note that table shows that a simple dictionary of 38740 english words
>> is enough to generate a collision in MD4.
>
> People have won the lottery on their first play. Doesn't mean a d*mn
> thing.
>
Except that it sucks as a checksum function, too.
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Apr 26 05:00:33 2003

This is an archived mail posted to the Subversion Dev mailing list.