Network-based Protocol Implementation Engineering and A Study on MachineLearning based Binary Code Analysis

Network-based protocol implementations play a significant role in the communication area, but the feature creep and increasing vulnerabilities have become serious problems. To deal with these problems, we proposed TOSS, an online server systems tailoring tool through binary rewriting. TOSS minify binary programs with just-enough features to reduce the potential attack surface and deal with the debloating problem. Besides, this paper also gives a comprehensive study of machine learning based binary code analysis to find the relationship between binary code and machine learning algorithm. For network-based protocol implementations whose source code is available, we proposed a packet generation and analysis functions identification tool through slicing, and test it with several network-based protocol applications.

Author

• Sun, Shaowen

Language

• en

Keyword

• Communication
• Machine Learning
• Packets

Date created

• 2019

Type of Work

• Master's Thesis

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Electrical Engineering

Degree

• M.S.

Advisor

• Dehghanian, Payman

Committee Member(s)

• Lan, Tian
• Venkataramani, Guru

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/jh343t03r