Weakness in Social Security Numbers Is Found

Jerry Leichter leichter at lrw.com
Thu Jul 9 22:12:44 EDT 2009

On Jul 8, 2009, at 8:46 PM, dan at geer.org wrote:
> I don't honestly think that this is new, but even
> if it is, a 9-digit random number has a 44% chance
> of being a valid SSN (442 million issued to date).
Different attack.  What they are saying is that given date and place  
of birth - not normally considered particularly sensitive - they have  
a good chance of predicting *a particular person's* SSN.

For untargetted attacks, broad statistics about the number of SSN's  
out there are fine.  But much attention these days is on targetted  
attacks against "high value" individuals.  It's in fact probably  
*easier* to find basic biographical information about date and place  
of birth of such individuals - you can often get much of it for, say,  
CEO's of public companies from their own brief bio's of their senior  
officers; scan newspapers for charity birthday events and you can get  
quite a bit more - than for a random member of the population.

Now, whether this really buys you all that much over other ways of  
getting hold of SSN's is questionable - and in fact the researchers  
are quoted as saying it's more of a demonstration of principle than  
anything practical.

BTW, 442 million SSN's have been issued, but how many are for people  
who have since died?  For many attacks, you need one for a living  
victim, which lowers the probability.
                                                         -- Jerry

The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com

More information about the cryptography mailing list