Pipelined architecture of fast modular multiplication for RSA cryptography

In this paper, a fast algorithm with its corresponding VLSI architecture is proposed to speed up the modular multiplication with a large modulus. By partitioning the operand (multiplier) into several equal-sized segments, and performing the multiplication and residue calculation of each segment in a pipelined fashion, a performance improvement can be achieved by using our algorithm compared with previous work. We also show an efficient procedure to accelerate the residue calculation and use carry-save addition to implement the architecture such that the critical path is independent of the size of the modulus. Therefore, the resulting architecture and implementation are very suitable to be applied to the high-speed RSA cryptosystem and can be easily implemented in VLSI technology.