Re: Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?
From: David Wagner (daw_at_taverner.cs.berkeley.edu)
Date: 01/20/05
- Next message: Bill Unruh: "Re: Factoring problem, solved"
- Previous message: Lits O'Hate: "Re: Factoring problem, solved"
- Maybe in reply to: Andy: "Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Next in thread: Paul Rubin: "Re: Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Reply: Paul Rubin: "Re: Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Wed, 19 Jan 2005 23:46:22 +0000 (UTC)
a c wrote:
>The problem I have is the limited capabilities of the HSP. It's designed for
>banking use and HMAC is not part of the capabilities -and we cannot access
>the secret keys either for use with our own application, only Ansi Key Block
>cryptograms which are used as input to encryption and MAC functions (eg:
>3DES CBC).
As Paul Rubin says, you can use 3DES-CBC-MAC instead of SHA1-HMAC.
(I slightly prefer 3DES-OMAC or AES-OMAC, but the difference is unimportant
if the messages you are hashing are of a constant, fixed length.)
>Looking at the response from David Wagner (thanks) this has the advantage
>that it uses standard (for our equipment) crypto operations, but we have to
>update old records (update the velocity checking data with fixed IV after a
>fixed period).
Yes. Most databases should allow you to do this, I would have thought.
>The other issue is the number of transactions processed within 48 hours -
>more than 100,000 which would mean at any one time there would be 100000
>records encrypted in this less than optimum manner.
Yes. At 20 bytes a record, that is 2 MB, not even a penny worth of disk.
At 100,000 encryptions per 2 days, that is 0.5 encryptions per second, which
any HSP should be able to handle trivially. Hard for me to imagine that this
would be a bottleneck, unless I'm missing something.
>i've search, but can't find, a description of A/B switchover. Could I have a
>pointer please.
I have two sets, set A and set B. At the start of odd-numbered days,
I delete everything from set A and re-set it to the empty state. At the
start of even-numbered days, I delete everything from set B and re-set
it to the empty state. When I have a transaction, I add a timestamped
record of that transaction to both set A and set B. At any point,
I can retrieve the transactions that have occurred within the past 24
hours by looking in one or the other of these sets (on odd-numbered days,
I look in set B; on even-numbered days, I look in set A).
In your application, set A would have a key kA and an IV ivA associated
with it; set B would have kB and ivB. When transaction T occurs, do this:
add (T, E(kA, ivA, CC #), timestamp) to set A, and
add (T, E(kB, ivB, CC #), timestamp) to set B.
When a set S is re-set, delete everything from S, securely delete kS and
ivS, and pick a new key and IV. To count the number of transactions
with a particular credit card number over the past 24 hours, pick the
appropriate set to look at, say set S, compute the value E(kS, ivS,
CC #), extract all records from set S that have this same ciphertext
value and have a timestamp within the past 24 hours, and count.
If you want to count over a 48-hour period, then just double the time
scale above.
Note: E(k, iv, x) could be a deterministic encryption algorithm, or it
could just be a keyed hash (a PRF) like SHA1-HMAC, 3DES-CBC-MAC, etc.
It doesn't really matter, as you never
- Next message: Bill Unruh: "Re: Factoring problem, solved"
- Previous message: Lits O'Hate: "Re: Factoring problem, solved"
- Maybe in reply to: Andy: "Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Next in thread: Paul Rubin: "Re: Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Reply: Paul Rubin: "Re: Combine Secure 3DES Encryption with ability to count occurence of known plaintext - how to accomodate both aims?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
