Source: Carnegie Mellon University
Traditionally, techniques for computing on encrypted data have been proposed with privacy preserving applications in mind. Several current cryptosystems support a homomorphic operation, allowing simple computations to be performed using encrypted values. This is sufficient to realize several useful applications, including schemes for electronic voting and single server Private Information Retrieval (PIR). This paper introduces an alternative application for these techniques in an unexpected setting: malware. The paper points out the counterintuitive possibility of malware which renders some aspects of its behavior provably resistant to forensic analysis, even with full control over the malware code, its input, and its execution environment.