Duplication based one-to-many coding for Trojan HW detection

Osnat Keren*, Ilya Levin, Mark Karpovsky

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A functional unit having multiple errors on its output may be considered as a computational channel, in which there is no restriction on the direction and the number of bit-flips. This model characterizes the distortion of the output caused by an active Trojan hardware. In this paper, we present a Concurrent Error Detection (CED) scheme that detects an active Trojan with high probability within a given time. The suggested CED scheme consists of a duplication-based one-to-many code that protects the most probable, valuable, or vulnerable words that may appear on the output of the system.

Original languageEnglish
Title of host publicationProceedings - 2010 25th International Symposium on Defect and Fault Tolerance in VLSI Systems, DFT 2010
Pages160-166
Number of pages7
DOIs
StatePublished - 2010
Event2010 25th International Symposium on Defect and Fault Tolerance in VLSI Systems, DFT 2010 - Kyoto, Japan
Duration: 6 Oct 20108 Oct 2010

Publication series

NameProceedings - IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems
ISSN (Print)1550-5774

Conference

Conference2010 25th International Symposium on Defect and Fault Tolerance in VLSI Systems, DFT 2010
Country/TerritoryJapan
CityKyoto
Period6/10/108/10/10

Fingerprint

Dive into the research topics of 'Duplication based one-to-many coding for Trojan HW detection'. Together they form a unique fingerprint.

Cite this