On Code Execution Tracking via Power Side-Channel

Explore a conference talk from CCS 2016 focusing on code execution tracking through power side-channel analysis. Delve into the authors' research on microcontroller unit (MCU) security, examining key extraction attacks and control flow integrity. Learn about the application of Hidden Markov Models for tracking code execution, including state definition, Viterbi algorithm implementation, and emission distribution design. Discover the evaluation setup, abnormal execution tracking techniques, and potential future developments in this critical area of embedded system security.

Intro
Microcontroller Unit (MCU) Core of embedded system
Scenario 1: Key Extraction Attack
Scenario 2: Control Flow Integrity (CFI)
Code Execution Tracking
Execution tracking via Power Side-channel
Problem Formulation . Given code & power trace of MCU
Hidden Markov Model (HMM)
Overall Framework
Naïve State Definition
Basic Block as State
Unequal-length States (Cont'd) 2. Viterbl : recover the most probable state sequence
Complexity Analysis
Emission Distribution Design
Observation Symbol Design
Signal Extraction
Calibrated likelihood 1. Sample cach instruction's average likelihood during normal execution
Evaluation Setup
Abnormal Execution Tracking
Conclusion & Future Works