Dissertation/Thesis Abstract

Dynamic Malware Analysis of GPU-Assisted Cryptoviruses Using Contained AES Side-Channels
by Espinoza, Jesus Pastor, D.Sc., Capitol Technology University, 2018, 503; 13853787
Abstract (Summary)

GPUs may be used to enhance cryptoviruses with an AES cryptosystem for encipherment to cause a denial of resources (DoR). A GPU-assisted payload with an implementation of AES can be used to bypass security products, accelerate the encryption of targeted files, unpack concealed code to deter malware analysis, and continuously change the execution space to prevent reverse engineering. In this quantitative study of GPU-assisted cryptoviruses, the extent to which input data size via execution time and GPU performance may be manipulated for encryption key recovery during dynamic malware analysis is measured using a malware probing laboratory. The treatments applied to GPU-assisted AES payloads comprise a plurality of side-channel attacks for key recovery. Contained timing and access attacks include an application of contained correlations of timing and unique cache line accesses for GPU-assisted payloads with T-Tables in global memory and the practical use of contained differences between means of timing sorted by bank conflicts for GPU- assisted payloads with T-Tables in shared memory. Tracing attacks over GPU-assisted AES payloads demonstrates the capabilities of contained tracing of GPU registers as an effective gray-box model for cryptanalysis. The experimental results establish a statistically significant effect for variable data size on execution time and memory access in a GPU-assisted payload for AES key recovery. By comparison, there is a nonsignificant effect for variable data size on memory trace in a GPU-assisted payload for AES key extraction.

Indexing (document details)
Advisor:
Commitee:
School: Capitol Technology University
Department: Cybersecurity
School Location: United States -- Maryland
Source: DAI-B 80/07(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer Engineering, Computer science
Keywords: Advanced encryption standard, Cryptanalysis, Cryptovirology, Cuda, Cybersecurity, Ransomware
Publication Number: 13853787
ISBN: 9780438989191
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest