Demo: Generating formally proven low-level parsers with EverParse

100
Опубликовано 8 февраля 2022, 17:11
Speaker: Aseem Rastogi, Principal Researcher, Microsoft Research India

DARPA and MITRE estimate that 80 percent of software security vulnerabilities have incorrect input validation as their root cause. In such scenarios, attackers provide malformed input, which, when not properly rejected, causes various misbehaviors such as buffer overruns or integer overflows, which ultimately lead to giving the attacker full control of the system. Thus, hardening critical software systems by systematically replacing their input validation code with formally proven message parsers can make a radical difference. This talk is the third of three research talks that will present ongoing and future research and engineering efforts to this end, demonstrating how projects such as Microsoft Research EverParse and DARPA SafeDocs harden input validation for various applications, ranging from network communication protocols to document formats. See talks by Sergey Bratus, DARPA, and Tahina Ramananandro, Microsoft Research Redmond, for more information.

Learn more about the 2021 Microsoft Research Summit: Aka.ms/researchsummit
автотехномузыкадетское