An analysis of the AAḃ asymmetric encryption scheme on embedded devices for IoT environment / Syed Farid Syed Adnan

Lightweight cryptography offers energy-efficient cryptographic capabilities on low powered devices such as those commonly found in the Internet of Things (IoT). One such lightweight scheme is the AA-Beta (AAḃ) asymmetric cryptographic scheme whose algorithm consists of only basic arithmetic operations of addition and subtraction for both the encryption and decryption processes. These features resulted in faster runtime compared to the more established RS A asymmetric encryption scheme, making AAp a potential alternative for IoT security. At the time of writing this thesis, AAḃ algorithm still exists as a mathematical concept and proven in a mathematical based software. To date, this research found no known practical implementation of the AAḃ algorithm to prove or to validate its efficiency on a real-world computing platform. There has been no analysis of the AAḃ performance on any resource-constrained platform although previous mathematical simulations showed that it would perform well in resource-constrained platforms. It is also not known how the algorithm would perform against the widely used RSA on resource-constrained platforms. This thesis seeks to study the AAḃ design philosophy and the specifications of the AAḃ asymmetric encryption scheme, develop the AAḃ encryption scheme and evaluate the computational speed, power consumption and feasibility of AAḃ encryption scheme on an embedded system in the practical domain. The results from the study are being compared to the mathematical simulation, and experimentally, to the RSA. This investigation takes the form of an IoT environment, beginning with an in-depth examination of the AAḃ encryption scheme design, and continuing into the development and real-world application of AAḃ from its mathematical origin. The experimental analysis focused on the AAḃ algorithm's performance on embedded platforms, namely, the Raspberry Pi microcomputer and microcontroller (ARM Cortex-M7) platforms. A feasibility assessment for an AAḃ cryptosystem for sensor nodes including a client to server testbed with wireless communications was carried out in the final stage. In this research work, the performance analysis of the AAḃ scheme produced remarkable timing improvements for the encryption and decryption of messages when compared to previous trials on a numeric computing environment. The research goes on to compare the energy consumptions for encryption and decryption using the AAḃ AAp scheme with similar processes using the Textbook RSA scheme on the aforesaid embedded platforms. The AAḃ encryption process demonstrates a significantly lower energy consumption compared to RSA, where as much as three times less energy was used by AAḃ when encrypting messages while considerable energy savings were also seen during AAḃ message decryption on the Raspberry Pi 2 and ARM Cortex-M7 device. A conclusion can thus be made that the AAḃ encryption scheme is a cryptographic scheme with a great potential for deployment on low-powered devices especially at the encryption side, offering fast and energy-efficient asymmetric cryptographic capabilities to all devices.