Testing Internet of Things crypto performance and power consumption

With the release of the SecureMark-TLS benchmark by the Embedded Microprocessor Benchmark Consortium (EEMBC) last year, it is now possible to gather information about the performance and power consumption of typically-used cryptographic functions on M-class microcontrollers (MCUs). This benchmark was developed in response to developers’ need to answer questions like “How long does it take to execute a Transport Layer Security (TLS) handshake with state-of-the-art cryptography?” and “Which MCU should I use in my product if I want to have a signature verification time of under 200 milliseconds?”

Peter Torelli and I presented the SecureMark-TLS benchmark at Arm TechCon 2018 and a recording of the presentation is available below, taken from the EEMBC YouTube channel. A short write-up motivating the use of the benchmark for company-internal research is given in this blog post.

2 thoughts on “Testing Internet of Things crypto performance and power consumption

    1. We have not been using machine learning or artificial intelligence in the construction of the performance and power consumption benchmark. I am, however, sure that you can use machine learning/AI in this context too.

Leave a Reply to Hannes Tschofenig Cancel reply

Your email address will not be published. Required fields are marked *