QRNG form factors, from PCIe card to silicon IP
Quside supplies entropy sources as PCIe cards, FMC modules, embedded modules and licensable silicon IP. The choice is governed by where entropy needs to be generated relative to what consumes it, and by how many units you are building.
The four shapes
- PCIe card — for servers and appliances with a slot free. The straightforward path for data centre and HSM deployments, and usually where evaluation starts.
- FMC module — for FPGA development and integration work, where the entropy source has to sit alongside custom logic.
- Embedded module — for devices where a card is not an option: constrained power, constrained space, and volumes where cost per unit matters.
- Silicon IP — for customers integrating the source into their own die, where the entropy source becomes part of a chip they are already building.
How to choose
The question is not which is fastest. It is where entropy has to be generated. If the consumer is a server application, a card in that server is the shortest path. If the consumer is a secure element inside a device shipping in millions, an entropy source in a data centre is not a solution at any throughput — the entropy would have to cross the boundary the secure element exists to defend.
Volume then decides the rest. Below a few thousand units, a module is almost always cheaper in total cost than an integration programme. Above that, integration starts to pay for itself, and above a certain scale IP licensing is the only shape that makes sense.
The path most customers take
Evaluation on a card, integration work on a module, then a decision about silicon once volume is real. Each step is a smaller change than it looks, because the entropy source and its model stay the same across form factors — what changes is the packaging and the interface.
The route to integration is described in the hardware pillar. If you are choosing between these for a specific system, the buyer's guide has the questions worth asking first.
Keep reading
- Deployment and integration
QRNG for HSMs and PKI
An HSM generates keys that protect everything else, so the entropy source inside it is the root of the whole hierarchy. What matters for an HSM or a certificate authority is not only that the source is physical, but that its entropy claim is evidenced within the module boundary and can be demonstrated during evaluation and audit.
- Deployment and integration
Integrating a QRNG: C, Python, Rust and the Linux entropy path
On Linux a hardware entropy source is normally exposed through the kernel random subsystem, so applications keep calling getrandom() and inherit the better seed without changing. Direct access through a device interface or a library is available where an application needs entropy that has not passed through the kernel pool.
- Deployment and integration
QRNG throughput explained, why 1 Gbit/s matters and when it does not
Entropy throughput matters only in relation to how fast a system consumes it. The sizing calculation is reseed interval multiplied by concurrent DRBG instances, plus any bulk consumption; for most workloads that lands far below a gigabit per second, and for a few it lands above it.