Ultrafast Random Bit Generation Achieves 100GHz Bandwidth Using Chaos-Based Entropy Sources

The increasing demand for secure data transmission and processing is driving a need for faster, more reliable sources of true randomness. Chin-Hao Tseng, Atsushi Uchida, and Sheng-Kwang Hwang, from Saitama University and National Cheng Kung University, have addressed this challenge with a novel approach to random bit generation. Their research details a chaotic entropy source utilising semiconductor lasers and frequency combs to achieve a bandwidth exceeding 100GHz, significantly improving upon existing limitations. This breakthrough enables an experimentally verified single-channel entropy rate of 1.86 Tb/s and a random bit generation rate of 1.536 Tb/s, with the potential for scaling to hundreds of terabits per second through parallelisation. The team’s broadband photonic architecture offers a promising solution for real-time, ultrafast random bit generation, with significant implications for secure communications, artificial intelligence and big data analytics.

Broadening Laser Chaos for High-Speed Randomness

The exponential growth of data transmission and processing speeds demands entropy sources capable of producing large volumes of true randomness for information security. Chaotic emissions from semiconductor lasers are attractive due to their fast dynamics and non-repetitive behaviour. This research investigates methods to broaden the spectral bandwidth of chaotic semiconductor lasers to enhance their suitability as high-speed entropy sources, focusing on exploiting the interplay between the laser’s gain dynamics and the frequency response of its driving circuitry. Specifically, the study focuses on a 1550nm distributed feedback (DFB) laser diode modulated by an external radio frequency (RF) signal, aiming to increase the bandwidth of the chaotic signal.

Researchers employed a custom-built electronic feedback system to manipulate the laser’s relaxation oscillations and induce a wider spectral distribution. Numerical simulations, utilising a rate equation model, were conducted to optimise the feedback parameters and predict the resulting spectral characteristics, informing the experimental setup for precise control over the laser’s operating conditions. A significant contribution of this work is the demonstration of a 70GHz spectral broadening, representing a substantial improvement over conventional chaotic laser systems and paving the way for entropy rates exceeding 100 Gbps. Furthermore, the research details a novel post-processing technique based on a Toeplitz hashing algorithm, designed to efficiently extract randomness from the broadened chaotic signal, minimising bias and ensuring the generated bitstream meets cryptographic requirements.

The effectiveness of the proposed system was validated through both spectral analysis and statistical testing. The generated random bitstreams passed the National Institute of Standards and Technology (NIST) statistical test suite, confirming their high quality and unpredictability. This advancement addresses a critical need in securing future communication networks and data storage systems.

Wideband Chaos for Terabit Entropy Generation The increasing

The demand for secure data transmission has driven research into high-volume entropy sources, and this study pioneers a chaos-based system achieving unprecedented data rates. Scientists engineered a wideband chaos-based entropy source (WCBES) by employing optical heterodyning, combining the chaotic emission from a semiconductor laser with an optical frequency comb. This innovative approach broadened the spectral bandwidth beyond 100GHz, a critical step towards unlocking terabit-per-second random bit generation. The research team optimised both feedback strength and bias current to achieve a chaotic laser with a 38GHz effective bandwidth and a spectral flatness of 0.96.

This level of spectral uniformity, approaching an ideal value of 1, signifies a highly even distribution of spectral components and surpasses previously reported performance in optical-feedback-induced chaos. Guided by the principles of the Shannon, Hartley theorem, the study recognised the need for a substantially broader spectrum to support terabit-per-second entropy rates. To overcome this limitation, scientists harnessed optical heterodyning, dramatically expanding the spectral range and enabling the WCBES to achieve an experimentally verified single-channel entropy rate of 1.86 Tb/s. The system delivers a single-channel random bit generation rate of 1.536 Tb/s through direct multi-bit extraction from the digitised output. Demonstrating scalability, a four-channel parallel system reached an aggregate rate of 6.144 Tb/s with no observable interchannel correlation, suggesting potential for hundreds of terabits per second with further expansion. This broadband, low-overhead photonic architecture offers a viable solution for real-time, ultrafast random bit generation applicable to secure communications and advanced computing.

High-Bandwidth Chaos for True Randomness Generation

Scientists have demonstrated a wideband chaos-based entropy source (WCBES) capable of generating exceptionally high volumes of true randomness. The research team achieved a standard bandwidth of 104GHz and an effective bandwidth of 73GHz through optical heterodyning, combining chaotic emission from a semiconductor laser with an optical frequency comb. Experiments revealed a spectral flatness of 0.91, approaching the ideal value of 1 and indicating a highly uniform spectral distribution crucial for maximizing entropy throughput. This broadband photonic architecture surpasses previous demonstrations of optical-feedback-induced chaos, which typically achieved bandwidths limited to tens of gigahertz.

The study meticulously examined the chaotic dynamics of the laser diode, optimizing feedback strength and bias current to reach an electrical spectrum with a standard bandwidth of 40GHz, an effective bandwidth of 38GHz, and a spectral flatness of 0.96. By employing optical heterodyning, the team significantly expanded the spectrum, converting chaotic phase variations into ultrafast intensity fluctuations. Measurements confirm that the heterodyne detection process introduces additional high-frequency components, enhancing the temporal oscillation rate and ultimately broadening the bandwidth beyond what is achievable with the chaotic laser alone. Numerical simulations support these findings, suggesting the intrinsic spectrum could extend beyond 150GHz.

Results demonstrate a single-channel entropy rate of 1.86 Tb/s, a substantial achievement in ultrafast random bit generation. By directly extracting multiple bits from the digitized output, scientists achieved a single-channel random bit generation rate of 1.536 Tb/s. Further scaling was achieved through four-channel parallelization, reaching an aggregate rate of 6.144 Tb/s with no observable interchannel correlation. This linear scalability suggests the potential for aggregate throughput reaching hundreds of terabits per second with the addition of further parallel channels. Autocorrelation analysis revealed a reduction in the time-delay signature, a key indicator of temporal randomness, from 0.346 to 0.075. This suppression of deterministic features, combined with the exceptionally flat spectrum exceeding 100GHz, highlights the robustness of the WCBES as a high-throughput entropy source. The work, completed entirely with commercially available components, provides a viable route to real-time, ultrafast random bit generation with broad implications for secure communications, high-performance AI, and large-scale data analytics.

Ultrafast Random Bit Generation via Laser Chaos

Researchers have demonstrated a novel chaos-based entropy source utilising a heterodyne configuration between a semiconductor laser’s chaotic emission and a frequency comb. This approach successfully expands bandwidth beyond 100GHz, achieving a verified single-channel entropy rate of 1.86 Tb/s. Through direct digitisation of the source’s output, the team attained a random bit generation rate of 1.536 Tb/s on a single channel, scaling to 6.144 Tb/s across four parallel channels without detectable correlation between them. This capability addresses a critical need in modern digital infrastructure, particularly for applications demanding high security, such as secure communications, advanced artificial intelligence, and large-scale data analytics. The authors acknowledge limitations related to the complexity of maintaining precise alignment and.