DIALECT: Communication Protocols Customization via Feature DIAgnosis, Lacing, Elimination, Cross-grafting, and Trimming
Abstract
Communication protocols in computing environments enable software systems to communicate messages and boostsecurity during such com"munication. An adverse side-effect of certain communication protocols, especially thatsupport multiple user-desired features in leg""acy systems, is the unnecessary software bloat because of having toimplement a variety of features that may be rarely used (and may" actually be unnecessary from functionalitystandpoint). Such rarely used protocol features could be exploited by malicious parties" as back-door entries to gainaccess to the sensitive information. Also, most legacy systems including naval assets no longer have a""ccess to theoriginal source code and exist in binary only form. To boost the security of such legacy systems, automated techniques"that would let the protocol implementation to adapt and transform would be critical to achieve secure communication.In this proposa"l, we have two main objectives: 1. Decipher and deconstruct the communication protocol used by anapplication binary through underst"anding the application s control fol w patterns and mapping them to protocol states.2. Perform Protocol FeaturCustomization through transforming the protocol states and creating dialects of theunderlying communication protocol to improve the overall system security. Protocol Features are identified usingsubset of protocol states that may be optionally removed or modified without affecting other aspects of the protocolfunctionality. Protocols are then customized toprevent adverse security exploits. We will explore tech"niques such as Feature Elimination, Feature Trimming, FeatureLacing, Feature Cross-grafting and Feature Blackhole to create dialect"s of the communication protocol that will besubstantially difficult to exploit from the adversary s perspective. The customized fea"ture is then validated, tested andverified to make sure that it does not break the existing protocol or create additional (unintend"ed) securityvulnerabilities.Relevance to ONR: ONR has keen interest in Late Stage Software Customization and Complexity Reduction" to protectthe naval assets, and has prioritized Cyber Security and Complex Software Systems research as a program withdesignated"" thrust areas that emphasize key aspects such as information security and assurance, secure informationmanagement and interaction," and determining the security properties of software systems. This proposalseeks to explore more practical approaches for de-constructing communication protocols in software and customizingthem to protect the ONR s software assets against security attacks. Due t"o the unavailability of source code in legacysystems, our solution approach performs analysis on application binary to decipher its" behavioral patterns beforecustomizing them. The project will explore novel methodologies to relate the application s behavior to r"egions of binarycode, and a feedback-driven, autonomous approach will be adopted to make sure that customization does notcompromis""e system security. If successful, this project can provide a more robust autonomous planning engine toprotect the US Navy s softwar"e and other assets that are controlled by such systems.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 01, 2017
- Source ID
- N000141712786
Entities
People
- Tian Lan
Organizations
- George Washington University
- Office of Naval Research
- United States Navy