Posted by Jerry Stuckle on 12/13/06 02:37

Norman Peelman wrote:
> "Toby Inkster" <usenet200611@tobyinkster.co.uk> wrote in message
> news:l22354-d0o.ln1@ophelia.g5n.co.uk...
>
>>Norman Peelman wrote:
>>
>>
>>>Well, I knew i'd get some flak for that. :) I was assuming they would
>
> be
>
>>>unique enough for the situation.
>>
>>There are varying degrees of uniqueness?
>>
>>--
>>Toby A Inkster BSc (Hons) ARCS
>>Contact Me ~ http://tobyinkster.co.uk/contact
>>
>
>
> The poster is converting strings not files to md5. md5 is not a method
> of encryption per se but a more complicated method of CRC to 'help' validate
> data... It is convienent to use md5 to 'one-way encrypt' small strings used
> as passwords. I would assume that the poster is also storing the matching
> plaintext along with the md5 so if a match was found he could then see if
> the plaintext matched as well but, according to the definitions i've read on
> the net; the odds are extremely high, paraphrasing:
>
> Taken from
> http://www.google.com/search?hl=en&lr=&defl=en&q=define:MD5&sa=X&oi=glossary_definition&ct=title
>
> The odds of any two strings having the same md5 are on the order of 2^64
> operations.
> The odds of finding a string having a given md5 is on the order of 2^128
> operations.
>
> ...so yes, i'd say that there are variable degrees of uniqueness. So maybe
> an easier way than my first post would be:
>
> Using textmd5 as PRIMARY KEY
>
> $db_link = mysql_connect(...,...,...); //etc.
> $query = "INSERT INTO logs_full VALUES ($textmd5, $textplain)";
> $resone = mysql_query($query, $dbc);
> if (!mysql_errno($db_link))
> {
> // match found (duplicate entry for key error 1062 returned from MySQL)
> // could then double check textplain here although odds of being here
> are very high
> // do other stuff here
> }
> else
> {
> // no match found
> // new entry has just been inserted
> }
>
> Of course this assumes he is also storing new md5's to check against. I
> gueuss we don't fully know what the poster completely intends on doing
> unless he pops back in with a little more info.
>
> Norm
>
>

And when you're comparing unknown strings, the odds go down rapidly.

The odds of anyone having the same birthday as you is 1 in 365.25.
However, in a group of 24 people, the odds of any two of them having the
same birthday are 1 in 2.

I don't have the time (or inkling) to determine how many md5 hashes you
would have to generate to get a reasonable chance of hitting a
duplicate. But it is much less than 2^64.

--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attglobal.net
==================

[Back to original message]