As performance execution is an important issue to encryption we need to know time performance relation between AES 256 with OFB operation mode and the random number tests implemented in DiceLock. We’ve got cipher and decipher time relation with stream length.

To perform the analysis we’ve used our future product DiceLock cipher and CounterTime class (counterTime.h and counterTime.cpp files available at our Knowledge site) of DiceLockTime DLL

We’ve checked that a linear regression formula with great conffidence has been computed which will allow us to extract DiceLock components relations.

We have computed the linear regression between AES 256 with OFB operation mode cipher time and plaintext stream length. Once computed we’ve got the following cipherTime and plaintext stream length plot:

The linear regression executed with RapidMiner shows that:

that is:

cipherTime = – 0.000115096427166367 + 0.000000163201822481 * length,

Coefficient = 0.998140067270979300

After that, we’ve computed the linear regression between AES 256 with OFB operation mode decipher time and ciphertext stream length. Once computed we’ve got the following decipherTime and ciphertext stream length plot:

The linear regression executed with RapidMiner shows that:

That is:

decipherTime = – 0.000310405294140850 + 0.000000163601687315 * length,

Coefficient = 0.997187435292768000

These data has been checked making use of Excel linear regression analysis and confirms the results. All detailled data can be found at AES 256 OFB of Time Analysis section.

*So, DiceLock research continues on … *

DiceLock is a trademark or registered trademark of his owner where applies. DiceLock research is the research to implement “Self-corrector randomisation-encryption and method”, european patent EP1182777 granted 08/21/2003 and US patent US7508945 granted 03/24/2009.

## Leave a Reply

You must be logged in to post a comment.