Parllel hashing algorithms for security and Forensic Applicatons

Abstract

Modern computing is now using parallelism to achieve higher performance as opposed to scaling frequency in the past Security algorithms are compute intensive and can benefit significantly in their performance from parallel implementations Parallel computing can also be used for more energy efficient implementation of security algorithm making them more usable in areas like mobile computing personal computing and cloud computing Cryptographic hashing algorithms are considered to be one of the important requirements for the digital data security and digital forensics Many secure hashing algorithms are available but most of them are serial in nature Not much work has been done to parallelize these algorithms to exploit parallel computing resources that are available in modern computing devices newlineThe design of new secure and fast hashing algorithms is important for many applications in today s information based world In this thesis we broadly explore the analysis and design of key security primitives cryptographic hash functions and digital signatures used for assuring authentication and data integrity A novel hashing transformation has been proposed and the security of the proposed algorithms have been analyzed using some well known statistical tests in the field Several analyses and computer simulations are performed to verify their characteristics confirming the satisfaction of the characteristics and conditions of the cryptographic hashing algorithms In addition to this tests have been also performed in some of the application areas of hash functions such as the digital signatures and the digital forensic applications newlineWe have also compared the performance of algorithms based on proposed transformation with the well known hashing algorithms Our comparison of results show that the proposed algorithms have a better performance than other existing algorithms Apart from the speedup gained due to parallel implementation the energy efficiency of the algorithms has also been