Multiwavelength modelocked semiconductor lasers for photonic access network applications

The authors previously derived the condition for a non-blocking architecture in a wavelength division multiplexing (WDM) photonic switching network, i.e., the generalized parallel delta network (GPDN) [1]. GPDN is configured as a multi-stage link connection of multiple unit switches, where the hardware complexity can be adjusted by varying the number of stages according to the network size. It can provide a framework for network design of various trunk networks, access networks, and in-station networks. This paper presents an architecture for a wavelength division multiplexing photonic unit switch that is applicable to a wavelength division multiplexing photonic switching network (GPDN), and evaluates the cost. The proposed wavelength division multiplexing photonic unit switch is constructed by assuming the situation in which an optical signal is multiplexed and handled without being converted to an electrical signal and employing the wavelength division multiplexing technique. The latter is considered promising for future networks from the viewpoint of reducing hardware complexity and economizing broadband communication. The input call to the carrier is wavelength converted to the carrier corresponding to the required outlet. Then, an optical filter is used to realize spatial switching. By combining the photonic unit switches proposed here, a large-scale switching network can be constructed. © 1998 Scripta Technica. Electron Comm Jpn Pt 1, 81(8): 59–67, 1998

Recently, photonic networks based on wavelength division multiplexing (WDM) systems have developed considerably in response to the explosive growth of the Internet. Devices with novel functions are required for photonic networks, and planar lightwave circuits (PLCs) are being considered to meet this need. This paper reports recent progress on PLC- type devices with advanced functions developed for the photonic networks. We describe fabrication techniques and device applications for silica and silicone polymer based PLC technologies.

Measurement device independent quantum key distribution (MDI QKD) has attracted growing attention for its immunity to attacks at the measurement unit, but its unique structure limits the secret key rate. Utilizing the wavelength division multiplexing (WDM) technique and reducing error rates are effective strategies for enhancing the secret key rate. Reducing error rates often requires active feedback control of wavelengths using precise external references. However, for a multiwavelength laser, employing multiple references to stabilize each wavelength output places stringent demands on these references and significantly increases system complexity. Here, we demonstrate a stable, wavelength-tunable multiwavelength laser with an output wavelength ranging from 1270 to 1610 nm. Through precise temperature control and stable drive current, we passively lock the laser wavelength, achieving remarkable wavelength stability. This significantly reduce the error rate, leading to an almost doubling of the secret key rate compared to previous experiments. Furthermore, the exceptional wavelength stability offered by our multiwavelength laser, combined with the WDM technique, has further boosted the secret key rate of MDI QKD. With a wide wavelength tuning range of 5.1 nm, our multiwavelength laser facilitates flexible operation across multiple dense wavelength division multiplexing channels. Coupled with high wavelength stability and multiple wavelength outputs simultaneously, this laser offers a promising solution for a high-rate MDI QKD system.

High coherent multi-wavelength or multi-tone light source are in high demand for optical density wavelength division multiplexed (DWDM) networks as the telecommunication capacity expands exponentially. However the linewidths of commercial multi-wavelength semiconductor lasers are typically a few MHz which is not acceptable when the frequency spacing of the multi-tones is 10 GHz. In this paper, a novel and simple method to suppress the linewidths of the multi-wavelength from ~6 MHz to ~10 kHz using an all-optical approach is proposed and demonstrated. The linewidths of the multi-wavelength are suppressed by a factor of 600 and the noise level of the multi-wavelength is decreased by nearly 20 dB. Each wavelength of the multi-wavelength operates in single longitudinal mode. Finally, more than 8 wavelengths over 10 nm are suppressed simultaneously through the approach and scheme presented in this work.

The multi-wavelength optical networking (MONET) is the ultimate solution for future optical communication systems serving both commercial and specialized communication applications. Wavelength division multiplexing (WDM) schemes have been developed to fully utilize the bandwidth of optical fibers. Multi-wavelength semiconductor laser diode (LD) arrays are considered as a compact choice for WDM sources due to their capability of simultaneous transmission of several channels into an optical fiber from a single structure. This increases the total system speed and resolution without significantly increasing system size or costs. Efforts on LD arrays for WDM applications focus on LDs such as DFB/DBR and VCSEL lasers.

Digital signal processing (DSP) has been widely applied in optical communication systems to mitigate various signal distortions and has become one of the key technologies that have sustained data traffic growth over the past decade. However, the strict energy budget of application-specific integrated circuit-based DSP chips has prevented the deployment of some powerful but computationally costly DSP algorithms in real applications. As a result, fiber nonlinearity-induced signal distortions impede fiber communication systems, especially in wavelength-division multiplexed (WDM) transmission systems. To solve these challenges in DSP, there has been a surge of interest in implementing neural networks-based signal processing using photonics hardware (i.e., photonic neural networks). Photonic neural networks promise to break performance limitations in electronics and gain advantages in bandwidth, latency, and power consumption in solving intellectual tasks that are unreachable by conventional digital electronic platforms. This work proposes a photonic recurrent neural network (RNN) capable of simultaneously resolving dispersion and both intra and inter-channel fiber nonlinearities in multiple WDM channels in the photonic domain, for the first time to our best knowledge. Furthermore, our photonic RNN can directly process optical WDM signals in the photonic domain, avoiding prohibitive energy consumption and speed overhead in analog to digital converters (ADCs). Our proposed photonic RNN is fully compatible with mature silicon photonic fabrications. We demonstrate in simulation that our photonic RNN can process multiple WDM channels simultaneously and achieve a reduced bit error rate compared to typical DSP algorithms for all WDM channels in a pulse-amplitude modulation 4-level (PAM4) transmission system, thanks to its unique capability to address inter-channel fiber nonlinearities. In addition to signal quality performance, the proposed system also promises to significantly reduce the power consumption and the latency compared to the state-of-the-art DSP chips, according to our power and latency analysis.

We have witnessed a rapid advancement in optical fiber telecommunications in the last decade. The bit rate of early optical transmission systems in the late 1970's was 45 Mbit/s. By the m1d4980's, the transmission speed had been increased to rates ranging from 400 Mbitls to 560Mbit/s. More recently, SONET hierarchy, from OC- 1 , OC-3, OC-12, to OC-48 at 51.84 Mbitis, 1 55.52MbitJs, 622 Mbitls to 2.488 Gbitls respectively, has been the standard chosen by CCITT. The next SONET standard rate of OC-l92 at 10 Gbitis is now under consideration. Although laser transmitters at 10 Gbitls rate are still in the research and development stage, new products will be introduced in the marketplace to fuffill systems and networking needs in the near future. The singlemode optical fiber has a potential bandwidth of over 30 THz. In order to utilize this bandwidth, high density wavelength-division-multiplexing (WDM) has been considered. WDM can offer not only the increased usage of bandwidth, but also the flexibility of routing signals in a network by means of wavelength [11. In the last few years, the transmission distances have been increased dramatically due to the use of Er-doped fiber amplifiers. High power semiconductor lasers at either 0.98 tm or 1.48 .tm wavelengths are used for the pump. The powers into the amplifier required for both pump wavelengths are about 20 mW and 50 mW respectively. On the other hand, semiconductor opticai amplifiers are attractive for monolithic integration with lasers, waveguides, modulators, wavelength multiplexers and demultiplexers. For example multi-wavelength DFB laser arrays, as many as 20 wavelengths, have been integrated monolithically with wavelength combiners (star coupler) and optical amplifiers on InP substrate [2). These integrated multi-wavelengths DFB laser arrays have the advantage of easier wavelength control and lower cost per wavelength than the discrete DFB laser transmitters for WDM applications. Such photonic integrated circuits (PlC) can increase the functionality and reduce the optical fiber-to-device interfaces that are the main factor for the high cost of optoelect.ronic components. The fiber-to-device interface is also the major concern for component reliability.

This paper compares the performance of two time-division multiplexed media access protocols designed for star-coupled wavelength-division multiplexed (WDM) photonic networks. The TDMA-C protocol is reservation based: one WDM channel is used to reserve access for data packet transmission on the remaining data channels. Control channel access arbitration is achieved through time-division multiplexing, enabling all active nodes the opportunity to transmit once every control cycle. This approach significantly reduces the long synchronization delays typical of time-division multiplexing systems since the control cycle length is proportional to the control packet size rather than the data packet size. Every node in this system has one turnable transmitter and two receivers. >

We discuss the advantages of wavelength-division-multiplexed (WDM) lightwave systems and the need for monolithically integrated multiwavelength laser arrays. For WDM light-wave systems to be cost effective, it is important to reduce the per-wavelength component cost. By photonic integration, the packaging and the control circuitry will be shared among all the wavelengths. In this paper, we review the multiwavelength laser arrays, the array coupling, and the photonic integration approach. We have demonstrated the monolithic integration of a 21-wavelength multi-quantum-well distributed-feedback laser array with a star coupler and optical amplifiers on the same chip. Operating a packaged module under CW operation, we obtained as many as 15 individual wavelengths simultaneously launched from one output port of the star coupler of this integrated chip into one lensed single-mode fiber. We also characterized the thermal and the electrical crosstalk from the neighboring lasers. This integrated chip stands a good chance of being a compact and low-cost multiwavelength laser source for future WDM lightwave systems.

We report a novel monolithic multi-wavelength mode-locked semiconductor laser design based on arrayed waveguide grating (AWG) for the first time. 19.6 ps pulses at a repetition rate of 2.9 GHz are obtained.

A multi-wavelength laser source is important for applications in areas such as wavelength-division multiplexing (WDM), interferometry, lidar and optical data storage. Simultaneous oscillation of multipe wavelengths have been demonstrated in several methods including spatially chirped Bragg reflector vertical cavity surface emitting laser (VCSEL) arrays,1 multichannel grating cavity (MGC) laser,2 multistripe array grating integrated (MAGIC) laser,3 and multi-wavelength DFB laser array by control selective area MOVPE.4 In those devices, arrays are required for the generation of multi-wavelength oscillation. Each array element is responsible for its individual wavelength oscillation. In this work, it is demonstrated that a single-stripe Fabry-Perot laser diode is able to oscillate at selected multiple wavelengths with a wavelength separation ≥1.3 nm, several times larger than the Fabry-Perot longitudinal mode spacing.

As the deployment of wavelength-division multiplexing (WDM) transmission facilities speeds up, the next step is to add cross-connects to route WDM signals. Electro-optical directional couplers (DCs) are well suited for the task since they can pass WDM signals. However, DCs suffer an intrinsic crosstalk problem. Removing crosstalk has been an important design issue. In this paper, we show how to take advantage of the multiple-wavelength characteristic of a WDM signal to reduce crosstalk in a DC-based photonic switching system.

A tunable multi-wavelength narrow-linewidth thulium-doped fiber laser employing two-stage cascaded Sagnac loop mirrors is proposed and experimentally demonstrated. The designed fiber laser is composed of a pump source, wavelength division multiplex, circulator, thulium-doped fiber, polarization controllers (PCs), couplers and polarization-maintaining fibers (PMFs). Two cascaded Sagnac loops are used as the cavity reflector and filter, and the proposed filter is fabricated using two sections of PMFs with 2-m and 1-m lengths, respectively. In the experiment, the laser threshold is 110 mW, and laser can emit single, double, triple, quadruple and quintuple wavelengths in the spectral range of 1873–1901 nm through the simultaneous adjustment of the two PCs. The power fluctuations and 3-dB linewidth are less than 2.1 dB and 0.2 nm, respectively, over 10 min at room temperature, and the side-mode suppression ratio is greater than 20 dB. The proposed laser will be useful in various fields, such as spectral analysis, fiber sensing and optical communication.

CMOS-compatible photonic integrated circuits (PICs) are emerging as a promising platform in artificial intelligence (AI) computing. Owing to the compact footprint of microring resonators (MRRs) and the enhanced interconnect efficiency enabled by wavelength division multiplexing (WDM), MRR-based photonic neural networks (PNNs) are particularly promising for large-scale integration. However, the scalability and energy efficiency of such systems are fundamentally limited by the MRR resonance wavelength variations induced by fabrication process variations (FPVs) and environmental fluctuations. Existing solutions use post-fabrication approaches or thermo-optic tuning, incurring high control power and additional process complexity. In this work, we introduce an online training and pruning method that addresses this challenge, adapting to FPV-induced and thermally induced shifts in MRR resonance wavelength. By incorporating a power-aware pruning term into the conventional loss function, our approach simultaneously optimizes the PNN accuracy and the total power consumption for MRR tuning. In proof-of-concept on-chip experiments on the Iris dataset, our system PNNs can adaptively train to maintain above 90 % classification accuracy in a wide temperature range of 26–40 °C while achieving a 44.7 % reduction in tuning power via pruning. Additionally, our approach reduces the power consumption by orders-of-magnitude on larger datasets. By addressing chip-to-chip variation and minimizing power requirements, our approach significantly improves the scalability and energy efficiency of MRR-based integrated analog photonic processors, paving the way for large-scale PICs to enable versatile applications including neural networks, photonic switching, LiDAR, and radio-frequency beamforming.

Development of sophisticated photonic transmission systems enabled evolution of photonic data transport networks towards cost-efficient and energy-efficient platforms capable to carry enormous traffic. Given access to technologically advanced equipment, network operator faces a series of decision problems related to how to efficiently use this technology. In this paper, we propose a mathematical model of network design problem applicable in the context of modern photonic network with wavelength division multiplexing (WDM). The mathematical models presented in the paper are given by means of Mixed-Integer-Programmes (MIPs). We first discuss a straight-forward formulation incorporating link-flow variables. The analysis of several real-size problem instances revealed that a need for a more advanced modeling approach, which will deliver problem formulations tractable by available hardware and solvers. Appropriate model was developed using Dantzing-Wolfe decomposition method. The resulting model uses link-flow based variables and incorporates significantly less constraints and variables if compared to the non-decomposed version.