This PDF file contains the front matter associated with SPIE Proceedings Volume 10560, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

To meet the increasing bandwidth demand for rapidly growing data flow from the wide spreading multimedia services, broadband Internet, and cloud services, 100G and even beyond data transmission based on compact and low-cost transceivers have been intensively studied in short-reach links, access, inter data-center-interconnections, metro and regional networks. For these systems, direct-detection is more attractive due to the lower system cost and power consumption. In this paper, we review four advanced modulation formats including pulse amplitude modulation-4 (PAM-4), carrier-less amplitude and phase modulation-16 QAM (CAP-16), discrete Fourier transform spread orthogonal frequency division multiplexing (DFT-S OFDM) and Duobinary with low-cost direct detection in 100 Gb/s/λ short-reach and metro optical transmission systems. These modulation formats are all experimentally demonstrated with similar digital signal processing algorithms. Three kinds of chromatic dispersion (CD) compensation ways named pre-CD method, single sideband (SSB) and dispersion compensating fiber (DCF) are also compared. Two types of modulators including IQ modulator and dual-drive Mach-Zehnder modulator (DDMZM) are employed to generate SSB or pre-CD signal in this experiment and their transmission performance is also evaluated. To further to improve the spectrum efficiency and reduce the cost, we also present a detailed comparison of applying three advanced modulation formats including pulse amplitude modulation-8 (PAM-8), carrier-less amplitude and phase modulation-64 QAM (CAP-64), and discrete Fourier transform spread orthogonal frequency division multiplexing (DFT-S OFDM) with a bandwidth-limited direct-detection receiver for 100 Gb/s/λ optical transmission systems. The comparison is carried out to evaluate the performance of each modulation format in terms of nonlinear equalization, received optical power and optical signal to noise ratio (OSNR).

In this article, we review our recent research progresses on the field programmable gate array (FPGA)-based real-time generation and reception of orthogonal frequency-division multiplexing (OFDM) signals for the short-reach direct-detection system and radio-over-fiber (RoF) system. With the low-complexity real-time digital signal processing (DSP) algorithms, a high-definition video signal with a data rate of 2.97 Gb/s transmission over 20 km single-mode fiber (SMF) without using forward error correction (FEC) technique is experimentally demonstrated in an intensity-modulated direct-detection optical OFDM system in the presence of sampling frequency offset (SFO). In addition, a real-time X-band OFDM-RoF system with heterodyne detection was also realized. With the help of Reed-Solomon coding with a multiplesymbol interleaving/de-interleaving scheme, the 16-ary quadrature amplitude modulation (16-QAM) OFDM-RoF signals over 10-m wireless and 2.26-km SMF-28 link was successfully achieved with a post-FEC bit error rate less than 1×10^-9.

Stokes vector receivers have attracted attention for their ability to enhance spectral efficiency due to the increase in the number of dimensions compared to conventional intensity-modulation and direct-detection (IM-DD) systems. One of the technical advantages of the Stokes receiver is that the original state-of-polarization (SOP) of a transmitted signal can be retrieved with digital signal processing (DSP) even if it is fluctuated during fiber transmission. In this paper, we review DSP technologies for SOP recovery, focusing on the modified constant modulus algorithm (CMA), and the radius directed equalization (RDE) algorithm in the Stokes space. Using computer simulations, we show that these algorithms can track random SOP fluctuations without any penalties.

Next generation wireless broadband access network requires a centralized radio access network (C-RAN) with high bandwidth and low latency of mobile fronthaul links. DWDM-PON is suitable for such fronthaul because of high bandwidth and low latency using dedicated wavelength channels. Auxiliary management and control channel (AMCC) superimposition scheme is advantageous for the management and control of DWDM-PON systems. In this paper, we review several types of superimposition and detection schemes that we have proposed and developed, and evaluate the impact of AMCC superimposition on the main signal as well as the quality of the superimposed signal.

Although, the technological breakthroughs such as WDM had allowed the capacity per fiber to be increased around tenfold every four years in the past decade, however, the capacity of the optical communication systems based on these transmission technologies is slowly becoming saturated. To satisfy the exponential growth of the Internet traffic, for the next generation short reach systems, including data center transmission and optical interconnect (OI) applications, the space-division multiplexing (SDM) can be a way forward. The SDM technology based on the multicore fiber (MCF) has recently attracted much attention as a potential approach. In this paper, design strategy of computer-compatible 8-core trench-assisted MCF (TA-MCF) is presented to reduce the intercore crosstalk. Moreover, the influence of butt-coupled TA-MCF OI on coupling loss is also discussed. On the other hand, another alternative approach, the mode division multiplexing (MDM) is also showing promise and mode (de)multiplexer is one of the key devices in such a MDM system. Designs of mode splitters using asymmetric directional couplers for the fundamental quasi-TE (TM) mode with the higher order quasi-TE (TM) modes (de)multiplexer including the H_y²¹ (H_x²¹), H_y³¹(H_x³¹, H_y⁴¹(H_x⁴¹, and H_y⁵¹(H_x⁵¹) modes are optimized by using a full-vectorial H-field finite element method.

Electro-absorption modulator (EAM) integrated distributed feedback (EADFB) laser are widely used for 10-, 40- and over 40-km optical communications. In a certain power dissipation condition, there is intrinsic difficulty in increasing the modulated light output power of EADFB lasers, because large insertion loss of the EAM deteriorates the power conversion efficiency. In this study, we investigated an SOA integrated with a EADFB laser to improve the power conversion efficiency of the EADFB laser. The device is called an SOA assisted extended reach EADFB laser, or AXEL for short. For a transmission with a 1.3-μm wavelength, the transmission distance was limited to 40 km because of the large fiber loss of 0.3 dB/km. To overcome this kind of transmission distance limit, we demonstrated a 1.3-μm AXEL with significantly increased power conversion efficiency. In addition, a 25-Gbit/s 80-km transmission by using AXEL and APD-ROSA was firstly achieved beyond the limitation of transmission distance. In contrast, with respect to an L-band wavelength range, a large chromatic dispersion severely limits the transmission distance. Then, we also investigated the AXEL with 1.57 μm wavelength range, and found that the SOA can operate as both optical booster and chirp compensator. Furthermore, the extension of 10-Gbit/s transmission to 80-km and enhancement of modulated light output power to 9.0 dBm were simultaneously achieved by taking advantages of its chirp compensation effect and high power conversion efficiency. The presented results indicate that the AXEL is an attractive candidate for a high-efficiency modulated light source with any wavelength range.

In this work optical networks in mesh topology with OPS/OBS photonic switching are analysed and evaluated. Analytical and computer simulation methods are utilised for evaluation of average number of hops (ANH), packet loss fraction (PLF) and transmission delay (latency). These OPS/OBS networks show that under high input traffic, networks with nodes of higher interconnection grade are more efficient than just larger networks (increased number of nodes). For both optical packest and bursts results indicate high throughput, low ANH, low latency and very low packet loss fraction at the optical layer in our networks.

In this paper, the impact of the number of channels on the performance of elastic optical networks (EONs) is examined considering a multilevel modulation format and coherent transmission. Network design parameters such as spectral bandwidth and channel symbol error rate (SER), are analysed. We simulated the transmission of quadrature phase shift-keying (QPSK) signals, modulated at 56 and 100 Gbps, to evaluate a proposed flexible spectral allocation method in order to evaluate the effect of number of channels and the required total spectral bandwidth.

Many high performance computers (HPC) and cloud computing applications rely on distributing tasks among large numbers of virtual and real servers. This implies that advancements in performance of data centers and HPCs is increasingly dependent on connectivity. In order to insure high degree of connectivity at increasing bit rates and distances the demand for large bandwidth-distance product connections is increasing. These can almost exclusively be provided using optical interconnects. Traditionally optical-interconnect come in the form of pluggable transceivers. However the increases in number of connections and bit-rate poses a limit to further scaling (the front-plate bottleneck). A shift towards mid-board optics is in the making but requires solutions which are compact, power efficient and low cost for manufacturing. In this talk we will present our most recent demonstrations of high density optical interconnect solutions as well as high density switches. First some details about the design aspects and advantages of compact electronic switches employing mid-board optical engines will be discussed. Then, for addressing the challenge of low cost optical interconnects, we will give details on our recent work targeting high channel count VCSELs based sub-modules. Results based on 2.5D and 3D assembly on high resistivity silicon will be discussed as well as the use of direct die attach to flexible PCBs for making high density interconnects.

Over the last decade, datacenter networks have leveraged CMOS scaling alongside extensible architectures to provide steady gains in performance. Now, the end of CMOS scaling threatens to dramatically increase energy in next-generation datacenters. A scalable, efficient photonic switching platform that enables fast network reconfiguration may shatter conventional electronic switch power-versus-performance scaling trends and decouple the ties between bandwidth and lithography node. The key metric is the nanosecond-scale switching time. However, the rapid reconfiguration also introduces challenges into the switch and network design. Nevertheless, several key technical innovations may help fast photonic switching technologies make a substantial impact on future datacenters.

An Optical Network on Chip (ONoC) relies on a switch that is capable of routing information from one element of the chip to another one, presenting both high data rate and reduced latency during transmission. One efficient way to achieve such functionality, is to combine elementary 4x4 switches to form a general YxY switch. We propose here to enhance this elementary block into a reconfigurable non-blocking 4x4 switch that can take advantage of Wavelength Division Multiplexing to improve its global data rate. For added flexibility, we ensure wavelength granularity so that each connection adapts the number of wavelength taken to its needs. On the one hand, in a single wavelength configuration, we achieve the first fully non-blocking 4x4 switch with double-rings that is compared to state of the art achievements. On the other hand, in a n-wavelengths WDM configuration, we demonstrate the behavior of our reconfigurable nonblocking granular WDM switch with double or single rings. Each ring (or double ring) is thermally tuned during a setup step, to place the resonant peak exactly where expected, to ensure stability of the switch. This architecture is scalable at will, and can be used in a YxY n-wavelengths switch for reconfigurable architectures, that could then benefit from WDM’s increased data rate for applications in High Performance Computing.

Within the European Project TERABOARD, a photonic integration platforms including electronic-photonic integration is developed to demonstrate high bandwidth high-density modules and to demonstrate cost and energy cost target objectives. Large count high bandwidth density EO interfaces for onboard and intra-data center interconnection are reported. For onboard large count interconnections a novel concept based on optical-TSV interconnection platform with no intersections and no WDM multiplexing is reported. All input/output coupler arrays based on a pluggable silica platform are reported as well.

We propose a flexible, software-defined optical switching fabric for cloud data centers, enabling multi-petabit per second network capacities. Our design is based on the cyclic interconnection pattern of arrayed waveguide grating (AWG) devices, whose routing functionality is complemented with recirculation fibers. Unlike traditional optical data center network proposals that rely on two independent fabrics for supporting mice and elephants, our design enables the support of flows of various sizes and requirements using a single AWG-based fabric and yields bandwidth flexibility by integrating wavelength and subwavelength switching granularities. There are two sets of connections paths in our design: dedicated paths between each pair of AWG input and output ports, and shared paths that are set up by multiple recirculation fibers. The recirculation fibers enable the statistical multiplexing of mice. As well, they provide for flexible, on-demand circuit provisioning between input and output ports. Applying Birkhoff-von Neumann matrix decomposition on a residual traffic matrix comprising the demands that cannot be supported through the dedicated paths, we come up with a weighted sum of permutation matrices that get mapped onto the set of available recirculation fibers. The calculated coefficients determine the proportion of a timeframe that the permutation matrices are serviced by distinct fibers. The matrix decomposition requires the combined scheduling of wavelength and time domains so that the AWG can operate as an adaptive flow switching device. Enhancing the functions of our wavelength-routing design with space switching using an optical MEMS switch results in extreme network scales, spanning millions of processing cores.

We analyze the maximum transmission distance and hop count of M-QAM signals, where link and node transmission characteristics are jointly considered. With the modulation format optimally determined by the analyses, spectral efficiencies of ultra-dense wavelength-division-multiplexing (WDM) networks are maximized.

We review the recent advancement in the system and device technologies for coherent optical communications. One major topic is high-dimensional modulation, and in particular the nonlinearity-tolerant modulation format family, based on four-dimensional 2A8PSK. This family, covering 5, 6, 7 bits/4D symbol, outperforms most known corresponding modulation formats in the linear and nonlinear region. We also review our recent progress on forward error correction including polar codes, and monolithic narrow linewidth semiconductor lasers.

The growing bandwidth demands of advanced driver assistance systems (ADAS) and infotainment technologies make Gigabit Ethernet over plastic optical fiber (POF) a natural choice for next-generation automotive data networks, especially in light of the recent approval of the IEEE 802.3bv standard for Gigabit Ethernet transmission over POF. POF-based transmission provides the advantages of low cost, light weight, easy termination, durability, and immunity to electromagnetic interference (EMI), while Gigabit Ethernet extends the current maximum data rate of 150 Mb/s provided by Media Oriented Systems Transport (MOST). Thus, we examine important design choices that impact the performance of POF-based automotive data links for data rates up to and beyond 1 Gb/s and different choices of modulation format, including NRZ and PAM-n. Because simulation is an efficient and cost-effective solution for studying the complex interplay of multiple design choices without requiring physical prototypes, we base our analysis on a comprehensive modeling framework for optical communication systems incorporating large-core step-index fiber and fiber-to-fiber connectors. We study anticipated system performance in terms of bandwidth and BER for different choices of link length and connector count, including the IEEE 802.3bv targets of 15 meters with four connectors and 40 meters with no connectors. In addition, we consider the impact of connector misalignments (both lateral and longitudinal) and source launch profile (measured in terms of its encircled angular flux, or EAF), which also directly affect link bandwidth.

VLC systems using commercial light emitting diode (LED) offer provision of simultaneous data transmission and illumination. However, limited modulation bandwidth and non-linear distortions are the main challenges faced by VLC designs in achieving high transmission data rates. Wavelength-division multiplexing in VLC (WDM-VLC) can improve the data rates to meet gigabit/s data standards. In this paper, we study the VLC data transmission using a commercially available red-green-blue-amber (RGBA) LED model. With quadrature amplitude modulation based discrete multi-tone (QAM-DMT) transmission, we demonstrate that proposed VLC system can achieve aggregate transmission rate of 2.1 Gb/s at bit error rate (BER) well below 2×10^-3, which is typical hard-decision forward error correction (FEC) threshold.

Optical signal processing using coherent optical frequency combs could have various potential applications for optical communications. In the beginning, an approach to achieve a tunable optical high-order QAM generation based on multichannel aggregation is discussed. Then, an all-optical pilot-tone based self-homodyne-detection scheme is demonstrated under two scenarios: (i) multiple WDM channels with sufficient power of pilot tones, and (ii) a single channel with a low-power pilot tone. Finally, a fragmented bandwidth allocation enabled by reconfigurable channel slicing and stitching is presented, in which a single optical channel could be redistributed into fragmented frequency slots and ultimately recovered at the receiver side.

Conference Presentation for "Ultra-low-noise amplification on a 100-micron-scale chip"

Digital subcarrier multiplexing (DSCM) is a frequency division multiplexing (FDM) technique which makes use of multiple digitally created subcarriers on each wavelength. This circuit-based approach combined with elastic optic networking has the potential to provide high bandwidth efficiency, sub-wavelength level flexible data-rate granularity, and electronic compensation of transmission impairments. We present our study and testbed development about a DSCM crossconnect which allows dynamic bandwidth, frequency and power level assignment of each individual digital subcarrier channel in order to achieve the desired data rate granularity as well as the quality of transmission.

In order to meet the recent vast demands of ICT bandwidth, high speed transmission systems, such as 100G, 200G and 400G, have been developed and installed with extremely high pace. De -fact standardization bodies such as IEEE802.3/OIF have played important roles in the industry for leading the proper concepts/designs of the leading edge high speed transceivers. Reviewing the recent activities of these standardization bodies, the future migration towards small form factor transceivers and its supporting integration technologies are examined. Latest MSA activities are also reviewed.

The traffic volume processed within the datacenter increases exponentially. In a typical datacenter, top-of-rack (ToR) switches connected to servers are interconnected via multi-stage electrical switches. The electrical switches require power-consuming optical-to-electrical and electrical-to-optical conversion. To resolve the problem, a single optical switch needs to be introduced to offload large-capacity flows. The optical switch in the datacenter must have a large number of input/output ports to support many ToR switches. The combination of delivery-and-coupling (DC) switches and wavelength-routing (WR) switches comprised of 1xN non-cyclic arrayed-waveguide gratings (AWGs) can attain high-port-count switches. To further increase the port count, the system loss must be reduced or higher-power transmitters must be used. To overcome this difficulty, we propose novel optical-switch architecture in which nxN uniform-loss and cyclic-frequency (ULCF) AWGs are utilized for the WR-switch part, where the system loss can be reduced by the factor of n. To confirm the effectiveness of our proposal, 12x48 ULCF AWGs were newly fabricated with planar-lightwave-circuit (PLC) technology. Part of a 1,536x1,536 optical switch was constructed, and good transmission performance was experimentally confirmed by bit-error-ratio measurements in 96-wavelength 32-Gbaud DP-QPSK signals in the full C-band. The throughput was 153.6 Tbps.

We propose and experimentally demonstrate a potentially simple and cost-effective optical multi-carrier source based on electro-absorption modulated laser (EML) cascaded with phase modulator (PM) driven by synchronous sinusoidal radio-frequency (RF) signal, which can generate 10 subcarriers with frequency spacing of 12.5 GHz and power difference less than 3 dB. We further experimentally demonstrate a novel W-band radio-over-fiber (RoF) system based on EML cascaded with PM, which can realize the generation and transmission of 25-Gb/s quadrature-phase-shift-keying (QPSK) modulated millimeter-wave (mm-wave) signal at 100 GHz adopting optical remote heterodyning and heterodyne coherent detection. The bit-error ratio (BER) after 80-km single-mode fiber-28 (SMF-28) transmission and 2-m wireless delivery can be less than the forward-error-correction (FEC) threshold of 3.8×10^-3.

In this paper, the challenges and problems for the in-band and off-band adaptive coding will be studied. To provide the agile and fast in-band adaptive coding, the information of the channel is inserted as the preamble into the each frame. Pseudorandom noise is utilized as the preamble and it is distributed and updated by the SDN controller to the transmitters and the receivers. Deploying the parallel preamble mapping distributed by the SDN controller is not only for the channel synchronization and the clock recovery but also for serving the in-band adaptive coding. The proposed scheme for the cross-layer design of the centralized SDN controller for the adaptive coding system is demonstrated and discussed. The proposed adaptive coding system is implemented and tested in the 100Gbps DP-QPSK fiber-optics transmission system. The performances of the system with different codes and adaptive scheme in the system are examined and plotted. Also, the time-domain analysis for the LDPCs changing is performed to demonstrate its operational features based on the different channel conditions signal-to-noise ratio (SNR). By deploying the pseudonoise preamble detection, the system enables the rapidly in-band adaptive coding service with low overhead and short guard time. The cross-layer SDN controlling scheme, for updating the mapping between preamble and codes, provides the agile and robust service in the fiber-optics network.

Recent growth of optical networks requires the installation of new and upgrade of existing components to support increased capacity demands. The task of designing optical transport systems and selecting equipment requires taking into account installed equipment as well as transceivers and (R)OADMs available at a moment. Often real OADMs, MDXs and transceivers operate at particular wavelengths, i.e. equipment is wavelength specific. Eventually, when designing a network, an engineer should take into account topology constraints such as optical channel path budget and equipment operating parameters such as transceiver reach and sensitivity. As WDM optical communication systems evolve from simple point-to-point links to complex network architectures, the optimized design of add/drop equipment and effective wavelength utilization becomes a challenging task when considering equipment constraints. We demonstrate an algorithmic approach, that offers an efficient method for the wavelength assignment and add/drop equipment allocation problem. This approach, which could be part of a comprehensive planning environment, allows optical network engineers to design cost effective add/drop equipment configurations with the most compact wavelength plan, resulting in the effective utilization of the DWDM grid. Our approach accounts for future channel loading, optical channel path parameters, e.g. its length, traffic demand parameters such as signal rate, equipment parameters such as add/drop port capability, operating wavelengths, transceiver bandwidth, etc. We will demonstrate how this approach will provide an easy way to design an optical network from scratch or upgrade the configuration of the existing ones. Different node architectures, including WSS based configurations or cascaded OADMs can be accommodated.

The IP traffic in datacenter has rapidly increased over the last couple of years because of the wide deployment of cloud computing services. Hence, multimode fiber (MMF) links are already deployed in the rack-to-rack networks particularly in large-scale datacenters to sustain the rapid growth of data traffic. In recent years, the network topology in datacenters shifts from the legacy tree-type to fabric-type: the leaf and spine structure is requiring higher-bandwidth distant product for the MMF links to maintain low latency. However, the modal dispersion has been a large issue to realize high bandwidth links.To address the problem, we focus on the restricted mode launch (RML) approach, by which optical power is coupled to limited numbers of propagating modes in MMFs. Key parameters for the RML are the numerical aperture (NA) and the spot size on the MMF launch end: both of them should be small to reduce the modal dispersion. These are difficult to achieve with small-size launch systems composed of microlenses. Thus, in this paper, a new compact RML device without using microlenses is designed and fabricated: an axially tapered graded-index (GI) core polymer optical waveguide is applied to the RML exciter. Firstly, the launch condition for a 1000-m long OM3 fiber is investigated by calculating the effective modal bandwidth (EMBc). To realize OM4-grade EMBc, we find that the NA and the spot size of the launching beam should be lower than 0.118 and 33.2 μm, respectively. Also, the polymer waveguides are designed to meet the above conditions. The optimum refractive index profile of the tapered core is calculated by solving the Fick’s diffusion equation, and then using the beam propagation method (BPM), the spot sizes and NAs from the tapered polymer waveguides are simulated. Finally, we fabricate GI core tapered waveguides using the imprint method.

The diversity traffic requirements, reliability communication infrastructure, and the real-time end-to-end (E2E) latencies are some of the major communication challenges to support a diverse set of emerging Internet of Things (IoT) applications include Smart Grid (SG) applications. For instance, using point-to-point fibers between each device and the controller has been reported, previously, as one of the solutions to address the E2E latency requirements. However, even with the fiber capacity, utilizing the technique was limited due to its excessive cost. Hence, using a commercial multiservice cellular network such as Long-Term Evolution (LTE) and Long-Term Evolution-Advanced (LTE-A) is a considerable solution due to the high-performance metrics: high throughput, low latency, higher reliability, and large bandwidth.

In this paper, we propose an uplink LTE Cascaded Priority-based scheduling algorithm (CPb) that supports a diverse set of Smart Grid (SG) applications, and improves the performance metrics compared to other two well-known schedulers, Proportional Fairness (PF) and Round Robin (RR). The proposed CPb algorithm uses a differentiation technique, applying the Time Domain Scheduler (TDS) and the Frequency Domain Scheduler (FDS), to meet the various SG traffic requirements and types for massive Machin-to-Machine (M2M) devices. Four SG traffic types for each M2M device are used in this study: (1) SG delay sensitive event-driven traffic is used as a SG Distribution Automation (DA), (2) Time driven traffic is used for the other SG types of traffic, including video surveillance, (3) Power quality data, and (4) Periodic Advanced Meter Infrastructure (AMI) data. The CPb results show a significant improvement in the performance metrics compared to the PF and RR schedulers, according to the LTE QoS Class Identifier (QCI) parameters.