Bob, I'm very ignorant about the topic but your question lighted my steps :DÂ ! Thanks a lot. I asked my colleagues and, as I suppposed before, I think we're in the worst case (that should be secure key). What I think (and hope) is "a crypto request to the CP I'm running on should respond in micros and not millis". So I think my request, at the moment, is taken in charge to somebody far from my TCB. It would be very valuable if somebody could share some statistics performance data. What I've read on IBM white paper is that: " It is also important to realize that the application design can have a significant impact on the performance characteristics. The crypto hardware (CPACF, Crypto Express2 and Crypto Express3) is designed to process large blocks of data. Passing large blocks of data to be encrypted or decrypted will provide noticeably better performance than passing small blocks, no matter where the work is done. " that's the opposite of my case, where we encrypt a small piece of data for Millions times. Finally, and it's what I need for the most, is to be sure a millisecond is the best I can get, only to avoid getting crazy to find out solutions that don't exist (at the moment). Thanks a lot. Max
---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
