Performance Analysis of Parallel Implementation of Advanced Encryption Standard (AES) Over Serial Implementation

Abstract

Cryptography is the study of mathematical techniques related to aspects of information security such as confidentiality, data integrity, entity authentication, and data origin authentication. Most cryptographic algorithms function more efficiently when implemented in hardware than in software running on single processor. However, systems that use hardware implementations have significant drawbacks: they are unable to respond to flaws discovered in the implemented algorithm or to changes in standards. As an alternative, it is possible to implement cryptographic algorithms in software running on multiple processors. However, most of the cryptographic algorithms like DES (Data Encryption Standard) or 3DES have some drawbacks when implemented in software: DES is no longer secure as computers get more powerful while 3DES is relatively sluggish in software. AES (Advanced Encryption Standard), which is rapidly being adopted worldwide, provides a better combination of performance and enhanced network security than DES or 3DES by being computationally more efficient than these earlier standards. Furthermore, by supporting large key sizes of 128, 192, and 256 bits, AES offers higher security against brute-force attacks. In this paper, AES has been implemented with single processor. Then the result has been compared with parallel implementations of AES with 2 varying different parameters such as key size, number of rounds and extended key size, and show how parallel implementation of the AES offers better performance yet flexible enough for cryptographic algorithms.