Acronym · Safety & Security
TRNG (True Random Number Generator)
A hardware entropy source that harvests physical randomness such as thermal noise, metastability or oscillator jitter, then conditions and health-tests it to seed deterministic generators.
Key generation and nonces depend on its quality.
- Also known as
- Hardware RNG, Entropy source
- Type
- Acronym
- Domain
- Safety & Security, Cryptographic hardware
Notes
NIST SP 800-90B entropy assessment and continuous health tests are required for certification, not just a noisy circuit.
Overview
A TRNG, or true random number generator, is a hardware block that produces random bits from an unpredictable physical process inside the chip rather than from an algorithm. Its output seeds the deterministic generators that supply keys, nonces and other secret values, so the security of almost every cryptographic function on the device rests on it. NIST guidance calls the physical part an entropy source and treats it as one component of a complete random bit generator.
How it works
A TRNG starts with a noise source, a circuit whose output depends on effects that cannot be predicted, such as thermal noise, the timing jitter of free-running ring oscillators, or a latch settling from a metastable state. The raw bits are usually biased and correlated, so a conditioning stage compresses them into fewer bits of higher quality. Intel's on-chip generator, for example, uses thermal noise in a self-timed circuit and conditions pairs of 256-bit raw samples into one 256-bit seed with AES in CBC-MAC mode.
Health tests run continuously on the raw output to catch a failing or tampered source. NIST SP 800-90B describes tests such as the repetition count test, which flags long runs of identical values, and the adaptive proportion test, which flags bias. When a test fails, a well-designed generator stops producing output rather than giving out weak values.
Because physical sources are slow and their quality varies with voltage, temperature and process, the conditioned output is normally used to seed a deterministic random bit generator (DRBG) from SP 800-90A, which then supplies random numbers at high speed. SP 800-90C defines how the two are combined. The hard part of TRNG design is not the circuit itself but showing, with a model and measurements, how much entropy each sample really contains, and making sure nearby logic or an attacker injecting a signal cannot reduce it.
Examples
- Intel's RDSEED instruction, which returns conditioned output from the on-chip entropy source for seeding software generators, while RDRAND returns DRBG output.
- A ring oscillator TRNG in a microcontroller security subsystem, feeding a hardware CTR_DRBG used for key generation.
- An entropy source submitted by an accredited lab to NIST's Entropy Source Validation programme for an entropy validation certificate.
- A smart card controller that halts random number output when a continuous health test detects a stuck noise source.
Why it matters
Related terms
Part of the Safety & Security glossary, with every term in the topic and the lifecycles, tools and standards that use them.
Where it appears on the Hub
Used on 5 pages of the Hub. In each group, the pages that mention it most come first.
Suggest an improvement
Suggest an improvement to this entry
A clearer definition, an example, a correction or a source an editor can check. Editors review every suggestion before the entry changes.
Last reviewed . Written by The Semiconductor Hub editorial team. The full glossary