Re: PGSQL encryption functions - Mailing list pgsql-sql

From Tom Lane
Subject Re: PGSQL encryption functions
Date
Msg-id 5332.1130942392@sss.pgh.pa.us
Whole thread Raw
In response to Re: PGSQL encryption functions  ("Mark R. Dingee" <mark.dingee@cox.net>)
List pgsql-sql
"Mark R. Dingee" <mark.dingee@cox.net> writes:
> The script I'm using to "break" md5 presumes that the cracker knows the 3 
> elements being concatenated together to form the plain-text sting which is 
> then passed into md5.  The method I'm using then begins running through 
> various permutations.  Do you believe that the methodology is appropriate or 
> that I'm being a bit paranoid?

Well, this is a fundamentally insecure way of using *any* crypto hash method.
You're blaming MD5 for the fact that you're misusing it.

There has to be some component of the hash input that the attacker
doesn't know and can't trivially guess.  Adding a randomly chosen "salt"
string is one common way to do that.
        regards, tom lane


pgsql-sql by date:

Previous
From: Moritz Bayer
Date:
Subject: Re: function, that uses different table(names)
Next
From: "Robert Blixt"
Date:
Subject: Function with dynamic command (EXECUTE) not working