Publication Date

5-2022

Date of Final Oral Examination (Defense)

2-1-2022

Type of Culminating Activity

Dissertation

Degree Title

Doctor of Philosophy in Computing, Cyber Security

Department

Mathematics

Major Advisor

Liljana Babinkostova, Ph.D.

Advisor

Robert Erbes, M.S.

Advisor

Sin Ming Loo, Ph.D.

Advisor

Marion Scheepers, Ph.D.

Abstract

Symmetric key cryptographic primitives are essential to encrypt data and protect communication between parties. Due to resource constraints, some modern devices are not capable of executing traditional cryptographic algorithms. This fact necessitates new lightweight cryptographic algorithms. Current research into lightweight cryptology is vast, in part due to the National Institute of Standards and Technology's (NIST) lightweight cryptographic standardization process.

There is not much research into the vulnerability to a power analysis attack created by the choice of parameters of lightweight symmetric ciphers. This dissertation develops and demonstrates white box and black box cryptanalysis models for power analysis attacks on lightweight cryptographic primitives.

The white box cryptanalysis targets the GIFT-COFB family of lightweight ciphers that include NIST lightweight standard finalists, and examines the security of their substitution layers in the power analysis setting. Findings include: When deployed over fields of characteristic 2, the most used platform, the non-linearity metric provides the best prediction of susceptibility to power analysis attacks. When deployed over fields of characteristic 3, substitution boxes display a wide range of vulnerability to power analysis attacks, leading to a classification of substitution boxes into weak and strong categories.

The black box cryptanalysis focuses on a proprietary cryptosystem acting between two embedded systems which require a lightweight cipher. The results of the black box cryptanalysis include a model for the decryption process of the proprietary system, and a software implementation of a prediction algorithm that predicts the plaintext giving rise to given ciphertext values.

These research results shed a new light on the resilience of lightweight cryptographic protocols against side-channel and black-box attacks and help in bridging the gap between theory and practice.

DOI

https://doi.org/10.18122/td.1957.boisestate

Available for download on Wednesday, May 01, 2024

Share

COinS