Overview: the role of TRNGs and the problem of bias
True random number generators (TRNGs) are valued for their ability to produce entropy from physical sources such as thermal noise, radioactive decay, or clock jitter. These sources are inherently non-deterministic, making them candidates for applications where unpredictability is essential.
However, the output of physical entropy sources is often imperfect. Raw entropy may contain bias, exhibit uneven bit distributions, or arrive in irregular bursts. These behaviours reduce the usefulness of the output unless “conditioned” or “whitened” in post-processing of raw entropy using a suitable PRNG.
The aim of post-processing of raw entropy from a physical source is to remove bias and intermittency inherent in the raw entropy.
Traditional whitening methods often reduce throughput or discard a significant amount of entropy. They may also introduce additional complexity or assumptions that are hard to verify.
The goal is to retain the intrinsic uncertainty of the original entropy while producing output that is statistically sound and operationally reliable. Mesinja offers a method to support this.
Mesinja’s role in entropy conditioning
Mesinja RNGs provide a structurally robust method for post-processing raw entropy. Their architecture is well suited to the needs of TRNG systems that require bias reduction and elimination of intermittency without compromising throughput or introducing opaque logic.
In a typical configuration, the Mesinja RNG produces two coordinated streams of digits. One stream is combined with raw entropy using a mixing operation such as bitwise exclusive-or or modular arithmetic. This mixed stream is then used as input for the next sequence of mathematical approximations within the generator. The second stream becomes the post-processed output of the TRNG system.
The result is a system that can ingest variable-quality entropy and still produce a uniform, independent bit stream. This makes Mesinja an effective and transparent whitening TRNG method while preserving the unpredictability of the physical source.
Advantages over conventional post-processing methods
Many traditional debiasing techniques reduce the usable entropy of a TRNG system. Methods such as the von Neumann extractor, cryptographic hash functions, or other resilience functions are designed to correct bias or intermittency in raw entropy, but often come with trade-offs. They often reduce throughput, depend on detailed characterisation of the source, or introduce opaque logic that is difficult to verify.
Mesinja RNGs offer an alternative as they preserve the unpredictability of the original source. The ability of a Mesinja RNG to mix entropy into its structural input allows it to absorb intermittent or biased data while still producing uniform and independent output.
This approach eliminates the need for complex heuristics or dynamic tuning. Output quality remains high even when the raw entropy varies in availability or statistical profile. Mesinja thus provides an inspectable method of conditioning TRNG output, without the overhead or information loss common in conventional systems.
Compliance and entropy utilisation
Mesinja RNGs support the post-processing requirements set out in standards such as NIST SP 800-90B, which govern how entropy from physical sources should be conditioned (although certification of the Mesinja RNGs under the NIST standard has not yet been sought).
The Mesinja output is reproducible under test conditions and traceable for audit. At the same time, the system resists forward or reverse inference. This supports secure use in embedded, regulated, or high-assurance environments. The structure preserves the unpredictability of the original source without discarding entropy or introducing opaque transformations.
The result is efficient entropy conditioning with reliable and uniform output. Transparency and performance are maintained without trade-offs.
Use case summary and contact
Mesinja provides a clear and efficient post-processing method for true random number generators. It enhances entropy without reducing throughput, preserves the unpredictability of the source, and offers a transparent structure.
The design is well suited to embedded TRNG modules, hardware security modules, and custom cryptographic processors and because it supports deterministic validation, it integrates smoothly into high-assurance systems.
If you would like to discuss deployment options or explore implementation in your environment, you are welcome to get in touch.