Acoustic Cryptanalysis Attempts on CPU and Keyboard
Acoustic sound is a natural phenomenon which can be emitted from interactions among physical objects or materials with its specific acoustic signature in frequency, amplitude and signal pattern or spectra. The emission occurs when a small surface displacement of a material is produced due to a rapid release of energy in a material, or on its surface. Acoustic sound which has specific spectra signature could be a powerful tool to reveal important information from the object being observed. The electronic system such as computer equipment and its peripheral produces a wide range of emanations in the terms of electromagnetic signal, as well as acoustic signal. Some research has showed that electromagnetic emanations can reveal information about the operations being performed by a machine , and recent experiments have shown that acoustic emanations can do the same. Experiments deal with acoustic emissions from Personal Computer (PC) conducted by Shamir and Tromer, as well as acoustic research for keyboard or keypad sounds by Asonov and Agrawal are discussed in this paper. The analogy and hypothesis of acoustic source are described also. The specific pattern in spectra observed from PC could be used to indentify or distinguish the running computation or tasks on the CPU. Furthermore, both Neural networks system and an acoustic triangulation approach are explained in this paper, those techniques are adapted to distinguish keystroke of a keyboard. The acoustic leakages could divulge about important information. The leakages make the secure oriented system is vulnerable from attackers or eavesdropper, and this paper recommends the countermeasures to prevent such leakages potentially used by eavesdropper.
Eindhoven University of Technology
Eindhoven, the Netherlands