Date of Award

3-10-2010

Document Type

Thesis

Degree Name

Master of Science in Electrical Engineering

Department

Department of Electrical and Computer Engineering

First Advisor

Jeffrey T. McDonald, PhD

Abstract

Protecting software from adversarial attacks is extremely important for DoD technologies. When systems are compromised, the possibility exists for recovery costing millions of dollars and countless labor hours. Circuits implemented on embedded systems utilizing FPGA technology are the result of downloading software for instantiating circuits with specific functions or components. We consider the problem of component hiding a form of software protection. Component identification is a well studied problem. However, we use component identification as a metric for driving the cost of reverse engineering to an unreasonable level. We contribute to protection of software and circuitry by implementing a Java based component identification tool. With this tool, we can characterize time required for carrying out adversarial attacks on unaltered boolean circuitry. To counter component identification methods we utilize boundary blurring techniques which are either semantic preserving or semantic changing in order to prevent component identification methods. Furthermore, we will show these techniques can drive adversarial cost to unreasonable levels preventing compromise of critical systems.

AFIT Designator

AFIT-GE-ENG-10-22

DTIC Accession Number

ADA517004

Download

Share

COinS