Accelerating More Secure RC4 : Implementation of Seven FPGA Designs in Stages upto 8 byte per clock
This work addresses security and performance issues in RC4 encryption for embedded systems, but it is incremental as it builds on existing ideas with hardware optimizations.
The paper tackles the problem of improving RC4 security and speed by introducing a Post-KSA Random Shuffling process and using FPGA-based parallel coprocessors, achieving a maximum throughput of 8 bytes per clock on a Xilinx Virtex-5 FPGA.
RC4 can be made more secured if an additional RC4-like Post-KSA Random Shuffing (PKRS) process is introduced between KSA and PRGA. It can also be made significantly faster if RC4 bytes are processed in a FPGA embedded system using multiple coprocessors functioning in parallel. The PKRS process is tuned to form as many S-boxes as required by particular design architectures involving multiple coprocessors, each one undertaking byte-by-byte processing. Following a ecent idea [1] [2] the speed of execution of each processor is also enhanced by another fold if the byte-by-byte processing is replaced by a scheme of processing two consecutive bytes together. Adopting some new innovative concepts, three hardware design architectures are proposed in a suitable FPGA embedded system involving 1, 2 and 4 coprocessors functioning in parallel and a study is made on accelerating RC4 by processing bytes in byte-by-byte mode achieving throughputs from 1-byte-in-1-clock to 4-bytes-in-1-clock. The hardware designs are appropriately upgraded to accelerate RC4 further by processing 2 onsecutive RC4 bytes together and it has been possible to achieve a maximum throughput of 8-bytes per clock in Xilinx Virtex-5 LX110t FPGA [3] architecture followed by secured data communication between two FPGA boards.