Software Tamper Resistance Using Program Certificates

Software tamper resistence and disabling device cryptograghically are investigated via run-time result checking of computations. The device needs to receive a new cryptograghic key after each pre-set period of time in order to continue function. The software execution integrity is checked by the authority when deciding whether or not to give out a new key. If the modification of the software execution causes an error in the computational result, it can always be detected and the device can be disabled cryptograghically by not given a new key. The detection is done via a run-time result checking using a certificate-based technique. Certificate-based techniques for run-time result verification are designed to support a secondary checker computation of a result by the authority by using the same input together with a special output, namely, the certificate, utilized by the primary computation in the device. We have designed the properties in the formal definition that the certificate design needs to satisfy in order to allow the result be checked in a provably "fool-proof" manner. The checker computation is done by the authority after the certificate is returned from the device. Since the checker algorithm is only known to the authority, not anybody else, it is hard for the hacker to forge the right information to be included in the certificate to fool the checker.