Skip to content
LOVE

Entropy generated and verified inside your device.

Locally Verified Entropy. Three words describing one idea: the source of randomness sits inside your system, and it tells you how it is behaving while it runs. Every Quside product is built around it.

The gap

Cryptography depends on entropy. Entropy comes from hardware.

Every key, nonce, seed and token depends on unpredictability, and that unpredictability has to come from a physical process somewhere. An entropy source validated once cannot report what happened to it afterwards — that is a property of how it was built, not of how carefully it was chosen.

The usual model

Validated once, then assumed

  • An entropy source is tested at a point in time, before deployment.
  • The certificate travels with the product for the life of the device.
  • Temperature, voltage and ageing move the source; nothing reports it.
  • The first sign of a problem is usually downstream, in a key.
What LOVE describes

Measured while it runs

  • The physical process has a model, so it can be measured against something.
  • On-device metrology tracks the parameters that describe its behaviour.
  • The device reports whether it is still inside its validated operating range.
  • That record is available to the host, so operations and audit can inspect it.

For a demonstration of the gap rather than a description of it, the inspection bench puts a healthy source and a failed one side by side, and shows what each method can see.

What it means

What has to be true before entropy is worth relying on.

LOVE stands for Locally Verified Entropy. The source sits inside your trust boundary; the device measures the physical process against a model while it runs; and what it produces is entropy from that process. All three have to hold at once.

L

Locally

Inside your trust boundary

The source sits inside your device and inside your trust boundary. Nothing crosses a network to reach the point of use.

Trust boundary
The entropy is generated on the same board as the cryptography that consumes it. There is no third party between the physical process and the key, and no external dependency to add to your threat model.
Availability
A source inside the device is available whenever the device is powered. It does not fail with a link, a peer, a subscription or an outage somewhere else.
Latency
Entropy arrives over a local bus rather than a network path, so seeding and reseeding do not have to be scheduled around round trips.
Sovereignty and audit
The physical source sits in a place you control and an auditor can point at. Randomness arriving over a network cannot be verified where it is used, which is why Quside does not offer entropy as a service.
V

Verified

Measured against a model, in service

The device measures the physical process against a defined physical and statistical model while it runs, and reports whether it remains inside its validated operating conditions.

Measured against a model
A quantum-optical process has a physical model that can be written down, and a model is something a device can be measured against.
Not output statistics alone
Statistics computed on an output stream cannot establish that a source is unpredictable — output from a well-seeded deterministic generator looks equally random. Verification here compares measured physical behaviour to the model.
A health state, continuously
The device reports whether it is still inside the operating conditions its model was validated against, rather than presenting a certificate issued once at procurement.
What it gives you
Continuous visibility into entropy-source health: a signal operations can alarm on and compliance can read, alongside key management, protocol design and implementation quality.
E

Entropy

The resource cryptography consumes

The security resource being delivered: the unpredictability every key, nonce, seed and token depends on.

The resource itself
Entropy is not a feature of an algorithm. It is a physical input that cryptography consumes, and its quality sets a ceiling on everything built above it.
A component, not a finished generator
A validated entropy source is one part of a random bit generator. In practice it seeds an approved deterministic mechanism, which is what your application actually calls.
Where the standards fit
SP 800-90B sets requirements for entropy sources, SP 800-90A specifies approved deterministic mechanisms, and SP 800-90C — final in September 2025 — describes the constructions that connect them.
Why the seeding matters
Approved DRBGs depend on strong seeding, correct construction and appropriate reseeding from trusted entropy sources. The deterministic layer inherits the quality of what seeds it.
At runtime

What happens while the device is running.

Health is measured against a defined physical and statistical model of the source. Statistics on the output alone cannot establish that a source is unpredictable, so they are not what the device reports.

  1. Step 01
    Physical process

    A quantum-optical process with a known physical model generates the raw signal.

  2. Step 02
    Measurement

    On-device metrology samples the parameters that describe how that process is behaving.

  3. Step 03
    Entropy estimate

    Entropy is quantified within the defined physical model, rather than inferred from output statistics alone.

  4. Step 04
    Health state

    The device reports whether it remains inside its validated operating conditions.

  5. Step 05
    Audit output

    That health record reaches the host system, where operations and compliance can inspect it.

The architecture

A validated source is a component, not a finished generator.

In a real system the entropy source seeds an approved deterministic mechanism, and that mechanism is what the consuming application calls. SP 800-90C, final in September 2025, describes the constructions that connect the two.

  1. SP 800-90B

    The entropy source

    The physical process, its model, the entropy quantified within that model, and the continuous health tests the standard requires. This is where Quside’s certificate E178 sits, scoped to Quside PCIe One and FMC One.

  2. Conditioning

    From raw output to seed material

    Raw output from a physical source is conditioned before it is used as seed material. Conditioning does not create entropy; it concentrates what the source already produced into a form the layer above can consume.

  3. SP 800-90A

    The deterministic mechanism

    An approved DRBG expands seed material into the stream applications consume. It is a sound construction that depends on strong seeding, correct construction and appropriate reseeding from a trusted entropy source.

This is why runtime verification matters at the bottom of the stack rather than the top. The deterministic layer inherits the quality of what seeds it, and nothing in the algorithm layer will report a source that has quietly drifted. See the full standards map.

Evidence

What an independent party has established.

The held certificates, with the scope of each. The full record, including the space qualification evidence, is on the certifications page.

NIST SP 800-90B

Entropy certificate E178
Held
Scope
Quside PCIe One and FMC One.
Establishes
The entropy source has been validated against the SP 800-90B requirements for entropy sources.
What it does not cover
The complete random bit generator. E178 covers the entropy source itself; SP 800-90C sets out how a validated source combines with an approved deterministic mechanism to form one, and that assembly is yours to make.

ENS

Alta
Held
Scope
Quside PCIe One (Garnet™ Series) and FMC One (Nellite™ Series), version 1.0, under one CCN-CERT catalogue entry.
Establishes
Qualification for use in Spanish public-sector systems categorised up to Alta — the highest of the three ENS categories, which therefore also covers Media and Básica.
What it does not cover
Other parts of the range, which need their own qualification: this one covers Quside PCIe One and FMC One at version 1.0, and runs to 31 December 2026, when it comes up for renewal. CCN’s Aprobado track, for systems handling classified material, is a separate qualification route.

The question worth asking your suppliers.

Where is your entropy generated, how is its quality verified, and who can observe that verification? Put it to any vendor, us included.