Hardware-Assisted Application Misbehavior Detection
ResumoProgramming is an error-prone task, which may result in application misbehavior. From the safety point of view, crashes are undesirable as they affect user experience, whereas from the security point of view, vulnerability exploitation can lead to security violations. Although fuzzing and other testing techniques help to minimize undesirable events, they do not eliminate them. As an additional “protection” layer, real-time monitoring can help in handling cases of previously unaddressed violations. However, approaches like Control Flow Integrity (CFI) are too specific to be extended to the general case. To overcome this challenge, we propose a hardware-assisted flow learning technique able to profile and detect deviations from the standard behavior, thus ensuring proper application execution.
BOTACIN, Marcus; GRÉGIO, André; GEUS, Paulo Lício de. Hardware-Assisted Application Misbehavior Detection. In: SIMPÓSIO BRASILEIRO EM SEGURANÇA DA INFORMAÇÃO E DE SISTEMAS COMPUTACIONAIS (SBSEG) , 2018 Anais do XVIII Simpósio Brasileiro em Segurança da Informação e de Sistemas Computacionais. Porto Alegre: Sociedade Brasileira de Computação, oct. 2018 . p. 253 - 266.