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.
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.
Digital chaotic physical layer encryption plays a crucial role in safeguarding data transmission security within data center interconnections (DCIs). However, existing encryption schemes remain vulnerable to active attacks, such as illegal optical access. To address these challenges, we propose a chaotic four-dimensional (4D) modulation scheme integrated with intrusion detection, which aims at enhancing the security of DCIs. The proposed scheme utilizes chaotic permutation and substitution encryption during the 4D modulation process, along with cross-polarization permutation to protect performance gains. Additionally, we propose a cost-effective and easily implementable intrusion detection method based on constellation same set ratio (CSSR). By dynamically adjusting the confidence region through CSSR analysis, the scheme enhances its robustness against noise interference. The proposed scheme is validated through 392 Gbit/s encrypted data transmission over an 80 km standard single-mode fiber, simulating attack scenarios including optical eavesdropping and illegal optical access. These results demonstrate that the proposed scheme achieves an information entropy of 3.999956, with a key space extending to 10360. The intrusion detection mechanism provides dynamic noise tolerance without requiring additional equipment, enabling the identification of illegal signals across various optical signal-to-noise ratios. This scheme provides a novel approach for resisting illegal optical access at the physical layer, with potential application in secure DCIs.
With the development of the internet, the security of the optical data center interconnections (DCIs) has become a highly concerned topic. However, current security schemes face limitations such as the absence of viable key distribution mechanisms and the sensitivity to channel condition fluctuations. To address the above challenges, we propose a novel physical layer encryption scheme for WDM-based DCIs that combines encryption with channel conditions. Considering the need for dynamic key updates in DCIs, we first propose embedding plaintext-related keys (PRKs) into wavelength-related pilot (WP) positions for distribution. Leveraging their controllable size to combat impairments under varying channel conditions enables dynamic error propagation. Additionally, the proposed recombination mechanism supplements DNA encryption, allowing data transmission across different channels during inter-strand base exchange. The two channel 111.12 Gb/s discrete multi-tone (DMT) signal transmissions are successfully demonstrated. These results indicate that this scheme can transmit data on different paths. The key space can reach 10360and the complexity is about O(n). Furthermore, the key can be transmitted without redundancy and strongly relies on channel conditions to control security. Moreover, this scheme can alleviate the damage caused by high peak to average power ratio (PAPR), with an increase of approximately 0.5dBm at the forward error correction (FEC) limit.
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.
Physical layer security (PLS) has become a critical technique for the rapidly expanding communication network in the postquantum era, supporting superior secure communication rates with lower deployment cost. Typical PLS architecture follows the “key distribution before encrypted communication” paradigm, which causes redundant system structure and potential security issues. We propose an intrinsic physical layer secure communication architecture aided by a joint secure key distribution technique in this article. The architecture utilizes random time-varying polarization evolution induced by the fiber channel as the entropy source and estimated bit error rate information as the medium to implement an intrinsic key distribution scheme without requiring additional channels and complex devices. We realize simultaneous physical layer encrypted transmission and error-free joint secure key distribution (SKD) based on this architecture. The experimental demonstration verifies the feasibility of the scheme. We believe the highly integrated secure communication architecture with multiple guaranteed security measures may not only provide a feasible, low-cost solution for the deployment of the SKD system in a variety of communication scenarios, but also propose a “key distribution within encrypted communication” paradigm for future PLS solutions.
To address the urgent demand for secure optical fiber communication, a novel dual independent key generation and distribution scheme for Optical Network Physical Layer Security is proposed for the first time. End-to-end (E2E) learning with a Transformer algorithm featuring adaptive positional encoding is introduced to accurately model coherent fiber transmission systems, achieving an R-squared value exceeding 0.95 for waveform fitting and data distribution. The error vector phase (EVP) and amplitude noise (AN) of the E2E system are utilized to comprehensively characterize channel uniqueness, reducing system reconstruction risks while serving as sources for highly random key sequences for legitimate parties. Key consistency stems from channel reciprocity and precise E2E modeling, with feasibility analyzed across diverse link conditions. Leveraging the fact that eavesdroppers cannot access the system's full lifecycle data, keys are extracted from outputs of transfer learning (TL)-updated E2E models to enhance security. Security analysis under various eavesdropping scenarios demonstrates resistance to attacks. Experimental results show that, with FPGA-implemented post-processing, error-free key distribution is achieved over 120 km SSMF at EVP and AN key rates of 306.74 Mbps and 521.82 Mbps, respectively.
Physical Layer Security (PLS) exploits characteristics and properties of the physical layer for data encryption and supplements conventional cryptography for enhanced overall security. Most of the PLS methodologies rely on the (statistical) characteristics of the transmission channel to either generate secure encryption keys, or to exploit them together with other physical layer characteristics (i.e. advanced modulation schemes) for secure transmission. However, these approaches often lead to increased complexity and therefore become impractical for actual system implementation. Recent advancements in Quantum Key Distribution (QKD) systems allow for the utilization of ultra-secure and robust high-rate key exchange. In this work, we propose and describe practical techniques for exploiting and seamlessly integrating highrate QKD keys to encrypt modulation parameters and quantities of conventional modulation schemes like M-QAM, DMT and OFDM of communication links. Moreover, we present transmission scenarios, integrating QKD-PLS in free space optics links, together with their numerical evaluation. The main advantage of QKD exploitation to the proposed solutions comes from the seamless and transparent integration and application of high-rate keys which can either be used in their original form or feed a pseudo-random number generator, to modify the modulation properties/symbols in very high rates, such that eavesdropping and decoding of the encrypted information becomes almost impossible. Additionally, we present the architecture of a real time practical system utilizing and seamlessly integrating the QKD keys into transceiver links to form a robust and ultra-secure PLS ecosystem.