Optimization of Pseudo Random Number Generator

Abstract

There has been a phenomenal growth in wire-less technology and mobile communication technology in last decade. Mobile devices are being increasingly used to access corporate networks to transmit business data over a diversity of links. Security stands out as a major concern in the design of such networks, especially in applications such as mobile commerce and corporate affairs that characterized by sensitive and time-critical information transfers. One of the most common security standards on mobile uses the GSM (Global System for Mobile Communications) stream cipher hardware level encryption based on the A5/1 algorithm. The circuitry for this is embedded into the Mobile Equipment as well as in the base station for secure communication. A5/1 algorithm uses LFSR (Linear Feedback Shift Register) to produce key stream in order to encrypt the voice data transmitted between subscriber’s cellular phone and the base station. Although it is strong encryption algorithm among all cryptographic algorithms used in GSM recent research studies have shown that A5/1 cipher is cryptographically weak and this in turn to a number of attacks. It has feeble clocking mechanism and the output key stream of A5/1 has low rate of linear complexity .To overcome these problems, we proposed novel security enhancement mechanism for GSM stream cipher. The novelty of our approach is the introduction of feedback tap mechanism enhanced by variable taps and four nonlinear combination functions. The hypothesis of proposed algorithm is supported by mathematical calculation provides high algebraic degree correlation immunity against correlation attack, algebraic attack, linear approximation attack and bereleykamp-Massey attack. The work is intended to enhance the linear complexity of preceding algorithm by design modification, implementation and evaluation for optimization to make it more secure and robust on a real test bed. Method is code in VHDL and synthesize on FPGA-SPARTAN 6 device. NIST Statistical test package is used to measure linear complexity of output key stream.