Non-Linear Collision Analysis

Xin Ye, Cong Chen, Thomas Eisenbarth

Abstract

As an unsolved issue for embedded crypto solutions, side channel attacks are challenging the security of the Internet of things. Due to the advancement of chip technology, the nature of side channel leakage becomes hard to characterize with a fixed leakage model. In this work, a new non-linear collision attack is proposed in the pursuit of the side channel distinguishers with minimal assumption of leakage behavior. The attack relies on a weaker assumption than classical DPA: it does not require a specific leakage model. The mechanism of collision generation enables independent recovery of partial keys so that for the first time the collision attack can be fairly compared with other standard side channel distinguishers. The efficiency of this attack has been verified by experiments on an unprotected microcontroller implementation of AES. Its immunity to modeling errors is confirmed through simulation of a broad range of leakage functions.
Original languageEnglish
Title of host publicationRadio Frequency Identification: Security and Privacy Issues
EditorsNitesh Saxena, Ahmad-Reza Sadeghi
Number of pages17
Volume8651
Place of PublicationCham
PublisherSpringer International Publishing
Publication date15.11.2014
Pages198-214
ISBN (Print)978-3-319-13065-1
ISBN (Electronic)978-3-319-13066-8
DOIs
Publication statusPublished - 15.11.2014
Event10th International Workshop on Radio Frequency Identification: Security and Privacy Issues - Oxford, United Kingdom
Duration: 21.07.201423.07.2014

Fingerprint

Dive into the research topics of 'Non-Linear Collision Analysis'. Together they form a unique fingerprint.

Cite this