Re: Order preserving encryption of numeric data

["Håkon Hallingstad" <hakon@ion.no>]

The ordering breaks the strength of any cipher.

Let us assume we know crypt(a), and want to find 'a'. Then we could do
a binary search for 'a' starting with a = M/2., where M :=
max(crypt(a)). And so we could find 'a' after lg(M) iterations.

If this cipher would have comparable strength to today's ciphers, this
would mean lg(M) ~ 2^90., or M ~ 10^(3 10^26), unrepresentable by
todays computers.

Regards,
Håkon Hallingstad

On Tue, May 17, 2005 at 03:05:52PM +0530, Anindya Mozumdar wrote:
> Hi,
>     Dont know whether this is an appropriate question to ask in this
> list, but someone experienced in dealing with such a situation could
> help.
> 
>     I would like an algorithm which will encrypt/decrypt numeric data
> using a key, such that after encryption their orders will be
> preserved,i.e, if crypt(a) denotes the encrypted form of a, and a < b,
> then crypt(a) < crypt(b). ( Note that this assumes that the algorithm
> will produce crypted data which can actually be ordered ). Of course,
> something like a -> ma + n, where m and n are constants would work,
> but I would like something more nontrivial.
> 
>     The order preserving property is necessary as the crypted data
> will be stored in a database, and I would like to issue database
> queries on them. ( The other solution is to retrieve all values from
> the database, decrypt them, and find the appropriate subset, but that
> would be too expensive ).
> 
> Thanks.
> Anindya.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-c-programming" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-
To unsubscribe from this list: send the line "unsubscribe linux-c-programming" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html