Data is being encrypted in a database. A message digest is being created for deduplication purposes. Is it ok to use the message digest that will be the key of the table that houses the encrypted data as the initialization vector? Since the data is not being transmitted, just archived and retrieved locally, is this a no-no? Since the message digest should theoretically be unique for every record, I'd expect the concern of using the same IV would not be a problem.
4
-
6$\begingroup$ Very related $\endgroup$– mikeazoCommented Sep 10, 2015 at 18:44
-
$\begingroup$ Thanks for the pointer. It was informative, but it made me consider something else. I don't believe the IV is something that can be extracted from the encrypted data, but I could be wrong. If it could be extracted, then I can see how you're leaking information. If not, then access to the row of data in the database would have the message digest so it would already be exposed, but the encrypted packet alone would not be leaking any info. Keeping the database table secure is another part of the puzzle, may be related, but perhaps can be handled in another way. $\endgroup$– TFullCommented Sep 10, 2015 at 20:10
-
$\begingroup$ Typically the IV is transmitted in the clear, so we assume the attacker has access to it. $\endgroup$– mikeazoCommented Sep 10, 2015 at 21:00
-
$\begingroup$ I think the answer to the linked question answers this too. $\endgroup$– otusCommented Sep 11, 2015 at 6:09
Add a comment
|