Threat Model

The threat model in Threat Model also applies to Intel SGX. In this section, the threat model will be extended and made more specific to SGX. An internal threat is able to use a high-resolution microscope to read the data stored in the fuses. Fuses are one-time programmable read-only memories that are large and easily read by this type of attack. Intel stores SGX-specific key material in fuses inside the CPU, but to make it more difficult for an attacker, the key material is encrypted in the fuses. The encryption key used is hard-coded into the circuitry of the CPU and the encryption algorithm is AES-128. Intel SGX does not protect against cache timing attacks, software side-channel attacks using performance counters, and power analysis attacks. Detailed documentation of the SGX hardware implementation is not publicly available at the time of this writing. This makes it difficult to analyze potential side-channel attacks.