A workload characterization of elliptic curve cryptography methods in embedded environments

Elliptic Curve Cryptography (ECC) is emerging as an attractive public-key system for constrained environments, because of the small key sizes and computational efficiency, while preserving the same security level as the standard methods. We have developed a set of benchmarks to compare standard and corresponding elliptic curve public-key methods. An embedded device based on the Intel XScale architecture, which utilizes an ARM processor core was modeled and used for studying the benchmark performance. Different possible variations for the memory hierarchy of such basic architecture were considered. We compared our benchmarks with MiBench/Security, another widely accepted benchmark set, to provide a reference for our evaluation. We studied operations and impact on memory of Diffie-Hellman key exchange, digital signature algorithm, ElGamal, and RSA public-key cryptosystems. Elliptic curve cryptosystems are more efficient in terms of execution time, but their impact on memory subsystem has to be taken into account when designing embedded devices in order to achieve better performance.