Architecture of the ECC Coprocessor for Gaussian Integers

Abstract

As previously mentioned, ECC systems are computationally intensive. Software implementations of public key algorithms on small embedded systems such as flash memory controllers may be too slow for some applications [81, 82]. The ECC point multiplication dominates the complexity of such ECC schemes. Hence, coprocessors were proposed for calculating the point multiplication. Many such coprocessor designs aim to achieve high computational performance and require only a few microseconds per point multiplication [83, 84, 86, 87]. Nonetheless, for applications in small embedded systems, the computational performance is less important. For example, a latency of a few milliseconds may be acceptable for a key exchange, secure authentication, or the generation and verification of a digital signature.