The KL82 series have a "low power" trusted cryptography (LTC) module, however, the datasheet explicitly states that there are "no [electrical] specifications necessary for the device's security and integrity modules."
How can we know that this module is truly "low power", and what does that even mean?
What kind of computational performance does this module offer when processing the various cryptographic functions?
I'm presently doing research into elliptic curve cryptography for IoT, so the specifics of the unit's performance are relevant to my work.
Does anyone know where I can find some solid, published data on this?
Thanks in advance.
The LTC module is part of our Cryptographic IP’s and acquires the low power naming because it provides better performance, reduces power consumption and code than other of our IPs in certain applications. Please refer to the wolfssl article about the K8x / KL8 LTC NXP Kinetis K8X LTC support for PKI (RSA/ECC) with #TLS13
Generally, electrical characteristics are not provided in the datasheets. For example, the current consumption of the Module itself, it is difficult to characterize because one cannot separate with ease the CPU and LTC module current consumption. From this post KL8x current draw for LP Trusted Crypto (LTC) you can get an idea of a test being made on a KL81Z128 @96MHz
Regarding, performance, you can check our FRDM-KL82z SDK example wolfssl_benchmark. In the case you were not able to run the demo you will find my results my results.
I have some KL82 LTC performance comparisons (with SW and other crypto modules) at https://www.utasker.com/docs/uTasker/uTasker_Cryptography.pdf
But these are only for AES and SHA and I don't think the module can help with elliptic curves.
[uTasker project developer for Kinetis and i.MX RT]
Contact me by personal message or on the uTasker web site to discuss professional training, solutions to problems or product development requirements
Thanks for the heads up - I haven't used Elliptic curves and since the MMCAU doesn't have support for public key acceleration I was assuming that the LTC would be the same, which was wrong...