Source author record

Zain Ul Abideen

Zain Ul Abideen appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Cryptography and Security Hardware Architecture Machine Learning

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Obfuscating the Hierarchy of a Digital IP

Numerous security threats are emerging from untrusted players in the integrated circuit (IC) ecosystem. Among them, reverse engineering practices with the intent to counterfeit, overproduce, or modify an IC are worrying. In recent years, various techniques have been proposed to mitigate the aforementioned threats but no technique seems to be adequate to hide the hierarchy of a design. Such ability to obfuscate the hierarchy is particularly important for designs that contain repeated modules. In this paper, we propose a novel way to obfuscate such designs by leveraging conventional logic synthesis. We exploit multiple optimizations that are available in the synthesis tool to create design diversity. Our security analysis, performed by using the DANA reverse engineering tool, confirms the significant impact of these optimizations on obfuscation. Among the many considered obfuscated design instances, users can find options that incur very small overheads while still confusing the work of a reverse engineer.

preprint2022arXiv

Preventing Distillation-based Attacks on Neural Network IP

Neural networks (NNs) are already deployed in hardware today, becoming valuable intellectual property (IP) as many hours are invested in their training and optimization. Therefore, attackers may be interested in copying, reverse engineering, or even modifying this IP. The current practices in hardware obfuscation, including the widely studied logic locking technique, are insufficient to protect the actual IP of a well-trained NN: its weights. Simply hiding the weights behind a key-based scheme is inefficient (resource-hungry) and inadequate (attackers can exploit knowledge distillation). This paper proposes an intuitive method to poison the predictions that prevent distillation-based attacks; this is the first work to consider such a poisoning approach in hardware-implemented NNs. The proposed technique obfuscates a NN so an attacker cannot train the NN entirely or accurately. We elaborate a threat model which highlights the difference between random logic obfuscation and the obfuscation of NN IP. Based on this threat model, our security analysis shows that the poisoning successfully and significantly reduces the accuracy of the stolen NN model on various representative datasets. Moreover, the accuracy and prediction distributions are maintained, no functionality is disturbed, nor are high overheads incurred. Finally, we highlight that our proposed approach is flexible and does not require manipulation of the NN toolchain.

preprint2020arXiv

A Systematic Study of Lattice-based NIST PQC Algorithms: from Reference Implementations to Hardware Accelerators

Security of currently deployed public key cryptography algorithms is foreseen to be vulnerable against quantum computer attacks. Hence, a community effort exists to develop post-quantum cryptography (PQC) algorithms, i.e., algorithms that are resistant to quantum attacks. In this work, we have investigated how lattice-based candidate algorithms from the NIST PQC standardization competition fare when conceived as hardware accelerators. To achieve this, we have assessed the reference implementations of selected algorithms with the goal of identifying what are their basic building blocks. We assume the hardware accelerators will be implemented in application specific integrated circuit (ASIC) and the targeted technology in our experiments is a commercial 65nm node. In order to estimate the characteristics of each algorithm, we have assessed their memory requirements, use of multipliers, and how each algorithm employs hashing functions. Furthermore, for these building blocks, we have collected area and power figures for 12 candidate algorithms. For memories, we make use of a commercial memory compiler. For logic, we make use of a standard cell library. In order to compare the candidate algorithms fairly, we select a reference frequency of operation of 500MHz. Our results reveal that our area and power numbers are comparable to the state of the art, despite targeting a higher frequency of operation and a higher security level in our experiments. The comprehensive investigation of lattice-based NIST PQC algorithms performed in this paper can be used for guiding ASIC designers when selecting an appropriate algorithm while respecting requirements and design constraints.