Secret recovery from electromagnetic emissions

Electromagnetic emissions leak confidential data of cryptographic devices. The electromagnetic emission has been reported as an important side channel for cryptanalysis. Electromagnetic Analysis (EMA) exploits the external radiation of cryptographic devices during encryption to reveal secret keys. The performance of EMA depends on the acquired signals to a large extent. To protect the devices from attacks, noises are introduced in the side channel either by unintentional interference from surroundings or elaborate design from engineers. Thus the secret recovery becomes difficult and even unavailable. In this paper, we propose two signal processing techniques to counteract both of these noises. The bandpass filtering and independent component analysis are widely used in other areas. We demonstrate their applications to EMA against the encryption algorithms on application-specific integrated circuit. With these techniques, the secret keys are extracted successfully and rapidly.