Analysis and improvement of S-Box in Rijndael- AES algorithm

Abstract

The internet has become a part of everyday life and is used as a communication tool, a way to bank, invest, shop and an educational and entertainment medium. As the importance and popularity of the internet has grown over the years, so has the number of threats from hackers on the internet which has necessitated the need for the encryption of confidential data. Various methods of data encryption have been used over time, with developments being made to improve these techniques as hackers develop improved ways of attacking the algorithms used for encryption. This process of continued improvement of cryptographic security brought about the development and acceptance of the Advanced Encryption Standard (AES), which is a National Institute of Standards and Technology specification for the encryption of electronic data including financial, telecommunications, and government data. The Rijndael algorithm was selected as the encryption algorithm for AES in October 2001 and is currently used by government agencies and the private sector to secure sensitive unclassified information. Research has shown that Rijndael is susceptible to differential/ linear cryptanalysis for 7 and 8-round Rijndael, saturation attacks, algebraic attacks and side channel attacks on reduced versions of Rijndael, which could pave the way for a full-blown attack on the Rijndael algorithm in the future. This research investigates the weaknesses present in the Rijndael algorithm using various custom-made testing tools and then using the results of this investigation to improve the security of the algorithm. The improvement is provided in the form a technique of generating highly non-linear output using a non-linear random number generator which uses the recursive inverse congruential method. The research will comprise of three phases; literature review, analysis of the Rijndael algorithm using custom-made tools and development of an improvement whose performance will be evaluated in comparison to the current algorithm.