A Parallel Arithmetic Architecture Extension for Elliptic Curve Cryptography over GF[2(m as power)]

Elliptic Curve Cryptography (ECC) is becoming increasingly the promising public-key method as it uses smaller keys at a security level equivalent to other public-key algorithms like Rivest-Shamir-Adleman (RSA). The operation that dictates the execution time of an ECC protocol is the scalar point multiplication. Scalar point multiplication operation boils down to modular addition, multiplication, and inversion operations. In this dissertation, a new modular and parallel ECC architecture is proposed which performs these operations on GF(2m) operands. The proposed architecture offers key size selection feature and saves power for small key operands by deactivating the unused modules. Furthermore, the proposed architecture handles multiple different operations in parallel in order to increase the overall performance and throughput of the system. A unique feature of the proposed architecture is that it is not restricted to a specific key size for parallel requests. In contrast, it is capable of handling different key sizes depending on the executed parallel operations without the need to explicitly reconfigure the hardware parameters. A simulator tool is also developed for this parallel architecture which helps finding optimum configuration setting for the architecture regarding the input data pattern. The experimental results show significant improvement in the timing, throughput, and energy performances with a slight overhead in the circuit area.