EMFIF2 is a modular framework for automated Electro-Magnetic Fault Injection (EMFI) fuzzing on embedded systems. The framework combines CNC-machine precision with an EMP generator to automate fault injection, enabling researchers and engineers to run sophisticated hardware attacks with minimal manual intervention. Its modular, scriptable design lets users define complex fault scenarios and observe device responses in real time. Through a series of case studies, the talk demonstrated EMFIF2’s effectiveness in uncovering flaws in popular embedded systems, covering the framework’s architecture, the integration of CNC and EMP technologies, and the challenges of reproducing physical attack conditions in a controlled environment.

The paper presents a solution for securing password-based authentication using Shamir’s $(k,n)$ threshold scheme, where $n$ password-derived secrets (shares) are created, and $k\leq n$ shares are necessary and sufficient for reconstructing the password. The solution is information-theoretically secure, with each share stored on a different host (Shareholder), requiring an attacker to compromise $k$ Shareholders to reconstruct the secret. To resist compromising the coordinating server (Dealer), the authors define a variant of Shamir’s scheme where the abscissas are unknown to the Dealer and Shareholders, making reconstruction impossible even if they are compromised. The authors have designed protocols for registration and authentication, analyzed scenarios with partially/totally compromised Dealer and/or Shareholders, and developed a prototype demonstrating the correct, effective, and efficient operation of the proposed method, providing a feasibility study for future cloud-based “authentication-as-a-service” implementations.