Binary Traceability and Validation
In many operational contexts, deployed binaries are the only consistently available artifacts for analysis, auditing, and compliance verification. Protection, auditing, and traceability must therefore include explicit checks on final artifacts rather than relying solely on upstream source-level assumptions.
Software supply-chain attacks exploit this systematically: by compromising third-party components, build pipelines or distribution channels, an adversary can inject malicious or vulnerable code that reaches deployed systems. The 2024 XZ Utils backdoor, the 2020 SolarWinds compromise and the 2023 3CX attack all illustrate this: in every case, the shipped binary was the last - and sometimes only - layer where the modification was objectively present and detectable.
BINTRACE develops methods and tools to incrementally validate binary updates, starting from a trusted first version and establishing a certification path up to the current version. The project aims to combine automated construction of semantic "deltas" between two binary versions, semantic reconstruction, and vulnerability/backdoor detection, with a strong focus on industrial-scale performance.
- Binary diffing
- Supply chain security
- Reverse engineering
- Vulnerability detection
- Semantic analysis
- Software traceability
Objectives
- Design an extensible and efficient data structure to represent semantic differences between binaries
- Detect and interpret binary-level changes, quantify their functional consequences and impact
- Classify modifications, reporting new security issues
Funding
- Funding: French National Research Agency (ANR)
- Duration: 48 months
Contact
Coordinator
Dylan Marinho, Associate Professor
LIP6, Sorbonne Université
Contact
bintrace@listes.lip6.fr