Entropy in HSMs and secure elements
An HSM or secure element generates keys internally, so its entropy source is the root of everything it protects. What the module's certification covers, and whether the entropy claim is evidenced within its boundary, decide how much the certificate actually tells you.
The entropy source is inside the boundary for a reason
A certified module has a defined cryptographic boundary, and entropy must be accounted for within it. That is not only a security preference: an entropy source outside the boundary raises questions an evaluator has to resolve about how the material crosses and what protects it.
This is a large part of why HSMs generate entropy locally, and it is a concrete instance of the general argument about trust boundaries.
What certification does and does not cover
A FIPS 140-3 or Common Criteria certificate is scoped. It may cover the module as a whole, and the entropy source may be evaluated under SP 800-90B or AIS 31 as part of that, or separately, or on a different version.
The question worth asking is not whether the module is certified but which document, which version, and whether the entropy source was evaluated in the configuration you are buying. The verification questions apply to module vendors exactly as they apply to component vendors.
The gap that remains after certification
Certification establishes what was true at evaluation. It does not establish that the source in a specific module, three years into deployment at elevated temperature, still matches the model it was evaluated against.
For a module whose keys protect a certificate hierarchy or a payment estate, that gap is worth closing with runtime evidence rather than inference — the observability distinction applies here with the highest stakes on the site.
Secure elements and constrained devices
A secure element in a phone, a car or a meter faces the same problem with less margin. There is less power, less area and no operator, and the device still needs its own unpredictable identity at first provisioning.
That is the case where an entropy source small enough to integrate matters most, and it is the doorway from this pillar into cryptographic hardware and root of trust.
Keep reading
- Where entropy breaks
CWE-330 in the wild, breaches caused by weak randomness
CWE-330, "Use of Insufficiently Random Values", is the class covering every failure where predictable values were used where unpredictable ones were required. It is the root of a family that reaches key generation, session tokens, initialisation vectors, nonces and password resets.
- Where entropy breaks
Entropy starvation at boot, in virtual machines and containers
Entropy starvation happens when a system needs randomness before its sources have produced enough — most commonly at first boot on embedded devices, in freshly cloned virtual machines, and in containers that inherit a host state. The usual workarounds trade the problem for a weaker one.
- Where entropy breaks
Famous RNG failures, from Debian OpenSSL to shared RSA keys
The best-known randomness failures — the Debian OpenSSL key generation defect, Android wallet key collisions, and the large-scale surveys that found shared RSA factors across internet hosts — share a pattern. The algorithms were sound and the material fed into them was not.