Significant increase of intra and inter data center traffic has been expected by the rapid spread of various network applications like SNS, IoT, mobile and cloud computing, and the needs for ultra-high speed and cost-effective short- to medium-reach optical fiber links beyond 100-Gbit/s is becoming larger and larger. Such high-speed links typically use multilevel modulation to lower signaling speed, which in turn face serious challenges in limited loss budget and waveform distortion tolerance. One of the promising techniques to overcome them is the use of advanced digital signal processing (DSP) and we review various DSP applications for short-to-medium reach applications.

The proliferation of a Colorless and Directionless and Contentionless (CD and C) architecture in metro core networks is rising up ever-greater demands on optical amplifiers to be smaller and higher integration. we overview recent advances in optical amplifier technologies, multiple EDFA arrays for compensating loss of a multicast switch and switchable gain EDFAs supporting a wide range of fiber-span loss distributions in the network and focus on the embedded passive component and pump laser in the amplifiers. We will also focus on the pluggable small form factor EDFA amplifies optical signals to enable long Hybrid Fiber Coaxial (HFC) links and amplifier for CFP-DCO/CFP2-ACO transceiver. Finally, we will discuss the feasibility of L-band amplifier and distribution Raman amplifier in a short-haul systems to realize a requisite optical signal to noise ratio (OSNR) to support high bit rate transmission beyond 100G and high capacity transmission.

Recently, the desired very high throughput of 5G wireless networks drives millimeter-wave (mm-wave) communication into practical applications. Phased array technique is required to increase the effective antenna aperture at mm-wave frequency. Integrated solutions of beam-forming/-steering are extremely attractive for practical implementations. In this talk, we review our recent progress and latest research on optical mm-wave beam steering. The remotely tunable integrated mm-wave beamformer based on optical true time delay is explored. The on-chip TTD network is realized by the architecture of arrayed waveguide grating feedback loop. The high-speed photo-diodes (>40GHz) are also integrated to reduce the package-induced power loss and cost. To allow its application in practical scenarios, the electronic integrated circuits including multiple-channel automatic power controlled trans-impedance amplifiers are designed to accommodate the electrical signal after photodiodes. Based on the integrated circuits, we have successfully demonstrated a 38-GHz beam-steered fiber-wireless system for 5G indoor coverage.

We have developed compact InP-based Mach-Zehnder modulators (MZMs) for small-form-factor pluggable coherent transceivers. In this paper, we introduce InP-based photonic integration circuit (PIC) technologies for high-performance MZMs. As the first topic, we show our design concept for a multi-quantum well (MQW) with a large refractive-index change and a low excess loss for a low-loss MZM with a low driving voltage. We fabricated dual polarization (DP) inphase and quadrature modulators (IQMs) in the form of a PIC in which we monolithically integrated a quad MZM based on the designed MQW. It operated at a half-wavelength voltage of 1.9 V with an excess loss of less than 1 dB as designed. We achieved 32-Gbaud DP-QPSK operation with a 0.2-dB penalty compared with a lithium niobate modulator. We also show another newly developed high-speed IQM for future higher-speed systems. We reduced the RF signal loss of an MZM by reducing the series resistance of the overcladding and optimizing the RF electrode structure. The fabricated high-speed IQM operated at higher than 64 Gbaud with QPSK/16QM modulations. Furthermore, we present an InP-MZM integrated with a new type of spot-size converter (SSC) fabricated with a three-dimensional semiconductor process. The SSC integrated MZM exhibited an insertion loss improvement of ~3 dB compared with our conventional InP-MZM for an optical system with a 4.5-μm mode field diameter. The low loss characteristics are beneficial in that they allow us to reduce the MZM-module cost thanks to an improved loss budget for optical assembly.

This work elaborates on: i) why the sliceable bandwidth variable transceiver (S-BVT) represents a key enabler for next-generation optical metro networks; ii) how it should be designed to take benefit of its capabilities and advanced features; and iii) which are the promising technologies to be adopted addressing the most relevant requirements and challenges. Specifically, S-BVT architectures based on multicarrier modulation and flexi-grid technologies, adopting cost-effective optoelectronic front-ends, enable flexible adaptation to dynamic traffic and variable path condition, targeting high capacity and scalability, while saving network resources and costs. Programmability and modularity are envisioned for integration in software-defined optical metro networks.

In this paper, flexible and highly reliable metro and access integrated networks with network virtualization and software defined networking technologies will be presented. Logical optical line terminal (L-OLT) technologies and active optical distribution networks (ODNs) are the key to introduce flexibility and high reliability into the metro and access integrated networks. In the Elastic Lambda Aggregation Network (EλAN) project which was started in 2012, a concept of the programmable optical line terminal (P-OLT) has been proposed. A role of the P-OLT is providing multiple network services that have different protocols and quality of service requirements by single OLT box. Accommodated services will be Internet access, mobile front-haul/back-haul, data-center access, and leased line. L-OLTs are configured within the P-OLT box to support the functions required for each network service. Multiple P-OLTs and programmable optical network units (P-ONUs) are connected by the active ODN. Optical access paths which have flexible capacity are set on the ODN to provide network services from L-OLT to logical ONUs (L-ONUs). The L-OLT to L-ONU path on the active ODN provides a logical connection. Therefore, introducing virtualization technologies becomes possible. One example is moving an L-OLT from one P-OLT to another P-OLT like a virtual machine. This movement is called L-OLT migration. The L-OLT migration provides flexible and reliable network functions such as energy saving by aggregating L-OLTs to a limited number of P-OLTs, and network wide optical access path restoration. Other L-OLT virtualization technologies and experimental results will be also discussed in the paper.

This paper proposes heuristics, which can maximize the number of remaining available routes and minimize the number of transceivers used in Elastic Optical Networks (EON). The aim of the proposal is to preserve the open capacity for the accommodation of future unknown dynamic demands. Case studies are carried out in order to analyze the performance of the heuristics over two networks. The results suggest that it is feasible to preserve enough open capacity to avoid blocking of future requests in EON with scarce resources.

We propose introducing aligned frequency assignment to each bandwidth channel, i.e., the semi-flexible grid, with the goal of minimizing frequency slot fragmentation under dynamic flexible-grid network expansion considering the expected need for channel capacity upgrades. In semi-flexible grid networks, for each set of channels having the same frequency bandwidth, a regular grid is defined where its spacing is the same as the required bandwidth and channels are aligned to their corresponding grids. A network expansion algorithm is developed that maximizes the efficiency of semiflexible grid assignment to achieve efficient channel bandwidth upgrading. Numerical experiments prove that the number of fibers necessary and the degree of fragmentation in the frequency domain are reduced by 15% and 80%, respectively, compared to conventional flexible grid networks accommodating the same traffic.

In this paper, we propose guard-band-shared direct-detection (GBSDD) scheme to simultaneously receive multi-band 100-Gb/s direct-detection optical signal with only one conventional 40-GHz photodiode. The modulation format of orthogonal frequency-division multiplexing based on offset quadrature amplitude modulation (OFDM/OQAM) is selected to provide signal spectrum with high side-lobe suppression ratio, which can effectively reduce the electrical subband frequency interference. In GBSDD scheme, only one guard band is required to accommodate the overlapped signal-to-signal beat interference (SSBI) induced by all the multi-band optical signals. Three different GBSDD schemes are experimentally demonstrated with total data rate of 100-Gb/s and OFDM/OQAM modulation format. The first one uses 6 subbands with transmission distance of 320 km and hybrid 32-QAM and 16-QAM formats. 6 lasers are used as the pilot carriers to beat the signal. The seconds utilizes only 2 subbands with 64-QAM format in 80-km fiber transmission, which are assigned onto two orthogonal polarizations. Only one optical pilot carrier is inserted to beat with the 2 subbands on the two polarizations. The 2 subbands are located on the one side of the optical carrier. Comparing with the first scheme, the bandwidth usage of the PD is enhanced from 1/2 to 2/3. In the third scheme, the 100-Gb/s optical signal includes 4 subbands with 32-QAM format after 880-km fiber transmission. The 4 subbands locate at the two sides of the optical pilot subcarrier. Under this condition, the bandwidth usage of the PD is further improved from 2/3 to 4/5.

Digital coherent technology is considered an attractive way of realizing both high-speed metro links and long distance transmissions. In metro areas, there is a strong demand for a smaller, faster transceiver module. This demand is mainly driven by the rapidly increasing data center interconnection traffic, where transmission capacity per faceplane is a key feature. Therefore, optical integration technology is desired. Since compensation in digital coherent technology is performed in the electrical or digital domain, users can deal with those optics performances that are not compensated for digitally. This means using a new material that cannot provide perfect characteristics but that is suitable for miniaturization and integration is possible. Silicon photonics (SiPh) is considered an attractive technology that would enable the significant miniaturization of optical circuits and be capable of optical integration with high manufacturability. While SiPh-based devices have begun to be deployed for very short or short reach links on the basis of direct detection technology, their digital coherent applications have recently been investigated in view of their integration capability. This paper describes recent progress on SiPh-based integrated optical devices for high-speed digital coherent transceivers targeting metro links. An optical modulator and receiver with related circuits have been integrated into a single SiPh chip. TEC-free operation under non-hermetic conditions and the direct attachment of optical fibers have both been realized. Very thin and small packaging with sufficient performance has been demonstrated by using the SiPh chip co-packaged with high-speed ICs.

As the demand for broadband applications continues rising, low cost and high capacity PON system has attracted more attention to keep up with the increasing demand in the future access network. Recently, the IEEE 802.3 Ethernet Working Group has already sponsored the discussion of next generation Ethernet passive optical network (NG-EPON) to provide 25-Gb/s per wavelength. In order to upgrade current 10-Gb/s PON to realize the capacity of 100 Gb/s PON with 25 Gb/s per wavelength, a variety of experimental demonstrations have shown the feasibility of 25-Gb/s per wavelength using the modulation schemes of four-level pulse amplitude modulation (PAM-4), electrical duobinary (EDB), optical duobinary (ODB) or non-return-to-zero on-off-keying (NRZ-OOK) for high speed transmission. In order to achieve higher performance, these transmission schemes are always combined with the advanced digital signal processing (DSP) which increases the technical complexity and the cost as well. Most of the previous demonstration are based on the off-line processing, therefore requires more time for the practical deployment. In this paper, we demonstrate the first field trial a real-time 100Gb/s TWDM-PON system with 4×25-Gb/s downstream and 4×10-Gb/s upstream transmission using 10G-class directly-modulated lasers (DMLs) and APD/PIN receivers. A single delay-interferometer (DI) is used to achieve frequency equalization as well as chirp management to increase the high frequency components of the system response and combat the chromatic dispersion (CD) during the fiber transmission. Note that there is no DSP applied for the whole system. Electrical clock/data recovery (CDR) chips are integrated on the main board for data generation, recovery and real-time bit error rate (BER) measurement. We obtained a power budget of 33 dB with 0-40km of standard single mode fiber based on NRZ-OOK modulation format for the downstream. The system stability is also verified using deployed 40-km fiber infrastructure over 67-hour real-time measurement.

The nonlinear compensation algorithm based on Volterra series has been proved effective in low order modulation OFDM system, such as QPSK/16QAM. In this paper, we demonstrate a 64QAM/ 128QAM DFT-S-OFDM signal generation with DML with some advanced algorithms such as DD-LMS, ISFA, DFT-S and nonlinear compensation to improve the signal performance. For the first time we demonstrate that the nonlinear compensation algorithm based on Volterra series can improve the performance of the high-order modulation DFT-S-OFDM signal such as 64QAM and 128QAM. In this experiment we have realized 19.1/11.2Gb/s 64/128QAM signal transmission over 15km fiber at 1307nm. For 64QAM case, the receiver sensitivity can be improved about 1dB when all the algorithms mentioned in this paper are adopted. And the BER can be improved from 4.7x10-3 to 2.8x10-3 at 7.0dBm for 128QAM signal, which reaches the HD-FEC threshold of 3.8x10-3.

Nonlinearity mitigation (NLM) is required to extend optical reach of spectrally efficient superchannel transmission with higher order modulation format. NLM by spectral inversion (SI) has been believed to be impractical because of required link symmetry with respect to SI. We show that SI still can mitigate significant nonlinearity in non-symmetric link configurations. We also developed an analytical model to estimate NLM by SI in arbitrary link configurations and SI placement. Especially, the impact of link symmetry on inter-subcarrier and intra-subcarrier NLM by SI in superchannel transmission is evaluated by simulation and explained by the analytical model.

Access networks based on vertical cavity surface emitting laser (VCSEL) transmitters offer alternative solution in delivering different high bandwidth, cost effective services to the customer premises. Clock and reference frequency distribution is critical for applications such as Coordinated Universal Time (UTC), GPS, banking and big data science projects. Simultaneous distribution of both data and timing signals over shared infrastructure is thus desirable. In this paper, we propose and experimentally demonstrate a novel, cost-effective technique for multi-signal modulation on a single VCSEL transmitter. Two signal types, an intensity modulated 10 Gbps data signal and a polarization-based pulse per second (PPS) clock signal are directly modulated onto a single VCSEL carrier at 1310 nm. Spectral efficiency is maximized by exploiting inherent orthogonal polarization switching of the VCSEL with changing bias in transmission of the PPS signal. A 10 Gbps VCSEL transmission with PPS over 11 km of G.652 fibre introduced a transmission penalty of 0.52 dB. The contribution of PPS to this penalty was found to be 0.08 dB.

Precise and accurate timing signals distributed between a centralized location and several end-users are widely used in both metro-access and speciality networks for Coordinated Universal Time (UTC), GPS satellite systems, banking, very long baseline interferometry and science projects such as SKA radio telescope. Such systems utilize time and frequency technology to ensure phase coherence among data signals distributed across an optical fibre network. For accurate timing requirements, precise time intervals should be measured between successive pulses. In this paper we describe a novel, all optical method for quantifying one-way propagation times and phase perturbations in the fibre length, using pulse-persecond (PPS) signals. The approach utilizes side mode injection of a 1550nm 10Gbps vertical cavity surface emitting laser (VCSEL) at the remote end. A 125 μs one-way time of flight was accurately measured for 25 km G655 fibre. Since the approach is all-optical, it avoids measurement inaccuracies introduced by electro-optical conversion phase delays. Furthermore, the implementation uses cost effective VCSEL technology and suited to a flexible range of network architectures, supporting a number of end-users conducting measurements at the remote end.