Please note: the training ticket does not include access to the conference. Similarly, the conference ticket does not grant access to the trainings. If you have any questions, reach out to us.

Training Objectives:

Modern cryptography has produced a multitude of ciphers that protect our daily lives including secure authentication, electronic transactions etc.

However, once the cipher is implemented on a physical device (microprocessor, FPGA, ASIC etc.) it becomes vulnerable to physical attacks such as side-channel analysis and fault injection.

With side-channel attacks the attackers monitor closely the power consumption or electromagnetic emission of a cryptographic device and they are able to extract the secret key using statistical techniques.

In this training we will provide extensive overview of side-channel analysis (SCA) of crypto implementations, showcasing the core techniques for key recovery. During the training the students will get the chance to develop several basic side-channel analysis tools in Python. They will also learn to perform hands-on physical attacks on real hardware. Subsequently they will use the tools to perform attacks on real-world datasets aiming at the secret key extraction.

We will also discuss the ways Machine learning and AI changed the side-channel analysis landscape and attackers’ capabilities in particular. First, we learn how AI can improve side-channel attacks and leakage exploitation in general. Second, we demonstrate the way side-channel analysis threatens implementations of AI on embedded devices and other platforms like GPUs, FPGAs etc. Finally, we discuss recent trends and other impacts of AI on hardware security.

The training will cover passive side-channel attacks on crypto implementations and countermeasures, including template attacks and leakage evaluation techniques. The training participants will learn to prepare the attacks including the attack set-up and execution on real hardware.

What to Expect? | Key Learning Objectives:

Each training module will start with a tutorial identifying the main concepts and the theory behind physical attacks. After that, several assignments will be given to master the content and learn about key insights and identify practical challenges.

Training Detailed Description:

A detailed description of the course structure and content, including an outline (day-wise agenda) of theory and practical exercises. Hands-on & interactive approaches are strongly encouraged.

Day 1: Intro to Side-channel attacks

Side-channel attacks on crypto implementations and countermeasures:

• Basic concepts and principles
• Side-channel leakage modeling
• Simple power analysis (SPA)
• Differential power analysis (DPA) and higher-order attacks

Assignments 1-2: Hands-on DPA attacks on software and hardware implementations

Day 2: Advanced attacks

Profiling attacks:

• Template attacks
• Step-by-step guide to profile and exploit a device's side-channel leakage

Assignment 3: Template attack on an ECC implementation on ARM Cortex micro-controller

Higher order attacks:

• Introduction to side-channel countermeasures (masking, shuffling, etc.)
• Higher-order side-channel attacks

Assignment 4: 2nd order attacks on a DPA-protected implementation

Leakage evaluation:

• TVLA and alternatives
• Leakage simulators

Assignment 5: TVLA evaluation of (un)protected implementations

Day 3: Interplay of AI and SCA

AI-assisted SCA

Side-channel analysis of AI implementations

Assignment 6: Deep learning SCA attacks on ECC implementations

Assignment 7: CUDA implementations of SCA

Who Should Attend? | Target Audience:

• Researchers and students who want to learn the core techniques of side-channel analysis
• Researchers interested in the security of AI-enabled systems
• Pen testers, government officers, auditors and evaluators of secure embedded devices
• Developers of secure IoT products
• Chip designers
• Any embedded security enthusiast

What to Bring? | Software and Hardware Requirements:

Own laptop running Windows / Linux / macOS

Installed Python version 3

Extra pen drive

What to Bring? | Prerequisite Knowledge and Skills:

• Basic programming
• Basic knowledge of statistics
• Familiarity with every-day cryptography, such as AES, PKC etc.

Resources Provided at the Training | Deliverables:

• Lecture slides and assignments
• Python code examples for analysis
• Side-channel datasets captured from AVR/ARM processors and FPGA implementations
• GPU dataset for reverse engineering neural nets
• Detailed description of set-ups used in training

ABOUT THE TRAINERS

Lejla Batina is a full professor at Radboud University. She got her professional doctorate in engineering from Eindhoven University of Technology and her PhD in Cryptography from KU Leuven, Belgium (2005). Prior to joining Radboud University she spent 3 years in industry as a cryptographer at Pijnenburg Securealink (later SafeNet B.V.) in The Netherlands. Her research interests include cryptographic implementations and physical attacks and countermeasures. She has coauthored more than 160 refereed articles on topics from secure cryptographic implementations and embedded systems security. Her current research interests are in intersection of AI and hardware security. She leads a group of 10+ researchers at Radboud University and 12 PhD students have so far graduated under her supervision.

Łukasz Chmielewski holds the position of Assistant Professor at Masaryk University in Brno, Czech Republic. His primary area of expertise revolves around side-channel analysis (SCA) of public-key cryptosystems. In general, he is also interested in hardware attacks, including fault injection, on real-world devices. Currently, he is actively involved in enhancing the capabilities of the side-channel CRoCS lab. Moreover, in recent years, he has worked on the applications of deep learning to SCA, targeting both symmetric and asymmetric schemes. In the past, he obtained his PhD and was a postdoctoral researcher in the Digital Security Group at Radboud University Nijmegen. He also has significant commercial experience in SCA, FI, and software-security evaluations of embedded devices. His overall practical experience in physical attacks spans the last 12 years.

Peter Horvath has a Master's in Computer Science with a focus on Artificial Intelligence and Cybersecurity. He is currently a PhD candidate at Radboud University advised by Prof. Lejla Batina. His research interests include reverse engineering neural network implementations on GPU and estimating side-channel leakage of crypto algorithms pre-silicon.