CRT-Based DSP Decryption Using Montgomery Modular Multiplication on the FPGA

The main contribution of this paper is to present an efficient hardware algorithm for Chinese Remainder Theorem (CRT) based RSA decryption using Montgomery multiplication algorithm. Our hardware algorithm supporting up-to 2048-bit RSA decryption is designed to be implemented using one DSP48E1 block, one Block RAM and few logic blocks in the Xilinx Virtex-6 FPGA. The implementation results show that our RSA core for 1024-bit RSA decryption runs in 11.263ms. Quite surprisingly, the multiplier in DSP block used to compute Montgomery multiplication works in more than 95% clock cycles during the processing. Hence, our implementation is close to optimal in the sense that it has only less than 5% overhead in multiplication and no further improvement is possible as long as CRT-based Montgomery multiplication based algorithm is applied. We have also succeeded in implementing 320 RSA cores in one Xilinx Virtex-6 FPGA XC6VLX240T-1 which work in parallel. The implemented parallel 320 RSA cores achieve 26.2 Mbit/s throughput for 1024-bit RSA decryption.