Kensa - Provably Secure Anti-RE
There is a large body of work devoted to program obfuscation (e.g. Obfuscator-LLVM, any DRM scheme). Provided sufficient time and effort, however, there is no reason to suspect these existing measures cannot be defeated.
Building off our work on PIR for Network Sensors, Kensa is cryptographic approach to protection of executable logic. Leveraging Private Information Retrieval (PIR) based on partial homomorphic cryptography, Kensa is capable of delivering Turing-complete software variants whose logic cannot be discovered without breaking widespread, tested cryptographic schemes (RSA & Paillier).
For example, two distrusted users / systems are running a piece of software. That software detects that there is an update available and asks the manufacturer for the update. Leveraging Kensa, the manufacturer is capable of serving two different updates, confident in the users’ inability to distinguish between the logic contained in the updates. One update may introduce Feature A, the other Feature B, or both may simply amount to a no-operation.
Kensa is not encrypted executable code. Such as scheme would be trivially defeated by simply logging the instructions executed on the CPU. Instead, Kensa retrieves and acts on data that is opaque to the user / distrusted system. The threat model is attacker-controlled hardware running Kensa-produced binaries.
This probably sounds like magic, but it’s actually not. We’re scientists, despite our rabbit-ears-in-a-hat logo.
Inquire for more.