Date of Award

3-14-2014

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Electrical and Computer Engineering

First Advisor

Robert F. Mills, PhD.

Abstract

Incorporating security in supervisory control and data acquisition (SCADA) systems and sensor networks has proven to be a pervasive problem due to the constraints and demands placed on these systems. Both attackers and security professionals seek to uncover the inherent roots of trust in a system to achieve opposing goals. With SCADA systems, a battle is being fought at the cyber -- physical level, specifically the programmable logic controller (PLC). The Stuxnet worm, which became increasingly apparent in the summer of 2010, has shown that modifications to a SCADA system can be discovered on infected engineering workstations on the network, to include the ladder logic found in the PLC. However, certain firmware modifications made to a PLC can go undetected due to the lack of effective techniques available for detecting them. Current software auditing tools give an analyst a singular view of assembly code, and binary difference programs can only show simple differences between assembly codes. Additionally, there appears to be no comprehensive software tool that aids an analyst with evaluating a PLC firmware file for modifications and displaying the resulting effects. Manual analysis is time consuming and error prone. Furthermore, there are not enough talented individuals available in the industrial control system (ICS) community with an in-depth knowledge of assembly language and the inner workings of PLC firmware. This research presents a novel analysis technique that compares a suspected-altered firmware to a known good firmware of a specific PLC and performs a static analysis of differences. This technique includes multiple tests to compare both firmware versions, detect differences in size, and code differences such as removing, adding, or modifying existing functions in the original firmware. A proof-of-concept experiment demonstrates the functionality of the analysis tool using different firmware versions from an Allen-Bradley ControlLogix L61 PLC.

AFIT Designator

AFIT-ENG-14-M-32

DTIC Accession Number

ADA599675

Download

Share

COinS