As we enter the big data and artificial intelligence (AI) era, integrating security and communication over optical fiber has become a critical challenge. This urgency is driven by the need to protect vast amounts of sensitive data, ensuring privacy security across global high-capacity optical networks. Traditional secure communication methods often struggle to maintain high-capacity transmission performance while providing robust security. Here we propose an integrated encryption and communication (IEAC) framework, designed to maximize mutual information (MI) for legal users while minimizing it for potential eavesdroppers. Enabled by end-to-end deep learning, this holistic framework trains a random number-selected geometric constellation shaping scheme to optimize encryption processes and transmission quality simultaneously. The IEAC experiment system achieves a groundbreaking single-channel transmission rate of 1 Terabit per second (Tb/s) over a 1200-km fiber link, employing a 26-channel, 3.9 THz bandwidth, full C-band wavelength division multiplexing (WDM) configuration. The MI for eavesdroppers is under 0.2 bit per symbol where the regular value is near 4.0, ensuring secure transmission. The IEAC scheme offers a scalable, promising solution to meet the escalating demand for high-throughput, secure data transmission in the face of advancing big data and AI computational technologies.
As we enter the big data and artificial intelligence (AI) era, integrating security and communication over optical fiber has become a critical challenge. This urgency is driven by the need to protect vast amounts of sensitive data, ensuring privacy security across global high-capacity optical networks. Traditional secure communication methods often struggle to maintain high-capacity transmission performance while providing robust security. Here we propose an integrated encryption and communication (IEAC) framework, designed to maximize mutual information (MI) for legal users while minimizing it for potential eavesdroppers. Enabled by end-to-end deep learning, this holistic framework trains a random number-selected geometric constellation shaping scheme to optimize encryption processes and transmission quality simultaneously. The IEAC experiment system achieves a groundbreaking single-channel transmission rate of 1 Terabit per second (Tb/s) over a 1200-km fiber link, employing a 26-channel, 3.9 THz bandwidth, full C-band wavelength division multiplexing (WDM) configuration. The MI for eavesdroppers is under 0.2 bit per symbol where the regular value is near 4.0, ensuring secure transmission. The IEAC scheme offers a scalable, promising solution to meet the escalating demand for high-throughput, secure data transmission in the face of advancing big data and AI computational technologies.
In telecommunication networks and in computer grid networks, a significant ongoing research is in a technology known as wavelength division multiplexing (WDM), which is capable to transport more than 1 Tbps aggregate traffic over a single fiber. However, when massive amounts of information are transported, as in optical networks, there are three issues that stand, cost-efficient ultra-high bandwidth delivery, security of information and data, and security of the network and of the transporting link layer. In optical networks, each fiber link consists of segments, each several kilometers long. Since a single WDM channel carries traffic, there are opportunities for sophisticated eavesdroppers to tap the fiber, copy a channel, isolate a specific payload and attempt to decrypt it if the encryption key is known or it is breakable with a fast computer. Therefore, end-to-end data security should not only rely on the client's encryption method but also on securing the transmission method on the optical link. In this paper, we present a WDM link security methodology that encrypts multiple channels and with multiple keys to render channel monitoring by eavesdroppers virtually impossible. Although our paper presents the applicability of the method on a point-to-point topology, it is applicable to a multilink and edge-to-edge network path.
In the era of large models and big data, the security of optical fiber communication backbone networks has garnered significant attention. Quantum noise stream cipher (QNSC) stands as a crucial method for safeguarding the physical layer security of optical fiber communications, yet the current schemes lag behind the rate capabilities of existing 400G optical fiber backbone networks. In this paper, we introduce deep learning into QNSC and propose an end-to-end quantum noise stream cipher (E2E-QNSC) scheme, which encrypts 16 quadrature amplitude modulation (QAM) into E2E-65536QAM/QNSC. Our experiments successfully demonstrate secure optical communication with a single-channel rate of 400 Gbps, a total capacity of 8.4 Tbps, and a transmission distance of 1520 km. Even in the most extreme scenarios, the detection failure probability (DFP) of the scheme remains at an excellent level greater than 0.9999, proving the security of the approach. The experimental results presented herein represent the highest rate-distance product record of QNSC secure transmission systems, to the best of our knowledge.
Abstract Optical encryption is an emerging technique that uses optical systems to securely encode and decode information for data transmission. In this study, we present an integrated optical encryption platform that consists of a scattering-augmented ptychographic encryption system, a transmission-efficient compressive sampling strategy, and a high-capacity decryption algorithm based on a hybrid model-data driven approach. Our encryption system leverages random light scattering as a physical unclonable function, enabling the encoding of complex object wavefronts with high optical throughput. In contrast to traditional encoders that offer insecure pseudo-randomness, the light scattering process ensures true randomness in the reported platform. To optimize transmission efficiency, we implement a compressive sampling strategy that reduces data transmission volume by at least one order of magnitude. Moreover, we devise a decryption algorithm that merges model-driven optimization with data-driven deep learning, effectively addressing the highly underdetermined retrieval task and providing robust decryption against various measurement noise and communication interference. The efficacy of our system is corroborated by an integrated optical prototype capable of encrypting ten million of pixels within minutes. Communication experiments confirm the resilience of our decryption algorithm, which retrieves high-fidelity results even under extreme transmission conditions characterized by a 20% bit error rate. Our innovative encryption technique presents a comprehensive solution for high-capacity data transmission, fortified by scattering-augmented security.
Protecting the security of optical network is a major worldwide challenge. The conventional schemes for securing communications are based on cryptography at the MAC layer and its higher layers, which are facing a great threat with the development of quantum computers. Optical encryption is a promising way for building security on the physical layer of optical communications. In this work, we propose a novel optical encryption scheme for wavelength division multiplexing (WDM) fiber-optic communication systems, by utilizing the private chaotic phase scrambling. Secure transmission of 50 Gbps signal over 50-km standard single-mode fiber is experimentally and numerically demonstrated. The results show that the WDM signal is efficiently encrypted into a noise-like signal, due to the spectral broadening effect of chaotic phase scrambling. Since the processes of encryption and decryption are totally implemented in the optical domain, the proposed scheme can encrypt the entire network traffic with low latency and high speed. Moreover, the proposed encryption scheme is compatible with the existing WDM optical networks, and only one pair of encryption and decryption devices is sufficient to encrypt all WDM channels, thus the scheme can be easily implemented at low hardware cost.
Navigating the era of artificial intelligence and massive cloud computing presents a critical challenge: securing the explosive growth of data traffic across optical data center interconnects (DCIs). However, conventional physical-layer security (PLS) methods often treat encryption and authentication as disjointed modules, lacking the intelligent adaptability for dynamic channels. Here, we propose an intelligent intrinsic PLS framework to jointly execute identity authentication and data encryption in short-reach DCIs. Enabled by end-to-end deep learning, this scheme integrates a mixture density network (MDN) for dynamic channel condition with a constellation autoencoder. Utilizing chaotic index modulation, the system nonlinearly learns highly distinguishable geometric constellations and embeds watermarks into subcarrier indices for redundancy-free transmission. Through this intelligent mapping, security attributes are natively fused into the physical signals. We achieve a secure transmission using discrete multi-tone (DMT) signaling over a 10-km standard single-mode fiber (SSMF). The results demonstrate that our scheme dynamically optimizes watermark embedding, strongly thwarting unauthorized eavesdropping and robustly resisting impersonation attacks. This avant-garde paradigm offers a highly scalable and intelligent solution for secure next-generation optical DCI networks.
Data proliferation in the digital age necessitates robust encryption techniques to protect information privacy. Optical encryption leverages the multiple degrees of freedom inherent in light waves to encode information with parallel processing and enhanced security features. However, implementations of large-scale, high-security optical encryption have largely remained theoretical or limited to digital simulations due to hardware constraints, signal-to-noise ratio challenges, and precision fabrication of encoding elements. Here, we present an optical encryption platform utilizing scattering multiplexing ptychography, simultaneously enhancing security and throughput. Unlike optical encoders which rely on computer-generated randomness, our approach leverages the inherent complexity of light scattering as a natural unclonable function. This enables multi-dimensional encoding with superior randomness. Furthermore, the ptychographic configuration expands encryption throughput beyond hardware limitations through spatial multiplexing of different scatterer regions. We propose a hybrid decryption algorithm integrating model- and data-driven strategies, ensuring robust decryption against various sources of measurement noise and communication interference. We achieved optical encryption at a scale of ten-megapixel pixels with 1.23 µm resolution. Communication experiments validate the resilience of our decryption algorithm, yielding high-fidelity results even under extreme transmission conditions characterized by a 20% bit error rate. Our encryption platform offers a holistic solution for large-scale, high-security, and cost-effective cryptography.