Development of an LFSR based Generic Framework for Cryptographically Secure PRNG using Parallel Computing to Enhance the Security of Network Applications

Abstract

Pseudo-Random numbers are at the core of any network security application. Also, security of Satellite Phones and Cellular Phones, depends heavily on the pseudo-random numbers generated. In the network security domain its use is particularly in key-generation, re-keying, authentication, smart-phone security etc. Also, current research shows that satellite based telephony system, having GMR-1 and GMR-2 algorithms for secret key generation are prone to attacks. The algorithm A5/1 used in GSM technology is also cryptographically poor. Hence generation of strong sets of pseudo-random number is needed. These random numbers are produced through a Pseudo-Random Number Gen- erator(PRNG). This generator in general terms is called a CIPHER.Hence, if there is a

aw or the PRNG produces predictable sets of random numbers, then the entire appli- cation would be prone to attacks. Therefore, development of a generic framework for generating strong sets of pseudo-random numbers is proposed. Hardware implementa- tion for GSM stream cipher has already been implemented, under a particular segment of mobile communication. Progress was made in all four folds, namely, vulnerability assessment, protocol design, implementation on both software and hardware, and eval- uation.The project has explored many research problems in this area, which opened up excellent research opportunities. The proposal aims to build an in-general framework and a uni ed model for enhanced security speci cally for LFSR based stream ciphers.The proposed generic model, uses results from the above case study. For the hardware deploy- ment, Spartan-6 FPGA toolkit is used and for the software part, a parallel computing platform namely CUDA is used. The model is aimed at development of a framework which generates strong sets of pseudo-random numbers for its use in various network security, satellite and cellular applications.