The Optical Networking and Communication
Conference & Exhibition

San Diego Convention Center,
San Diego, California, USA

Network Automation Presentations and Events

Network Automation and Analytics Schedule Sunday, 11 March Monday, 12 March Tuesday, 13 March Wednesday, 14 March Thursday, 15 March
Workshop: AI-assisted Automated Network Operation: Getting Matured or Not? 16:00 - 18:30        
Machine Learning and Performance Monitoring (M2F)   10:30 - 12:30      
Probabilistic Shaping I (M3C)   14:00 - 16:00      
Telemetry and Survivability (M3A)   14:00 - 16:00      
Modelling, Disaggregation & Networking Automation (M4A)   16:30 - 18:30      
Deployments of Optical Access and Front Haul (M4I)   16:30 - 18:30      
Probabilistic Shaping II (M4E)   16:30 - 18:30      
Practical Aspects in Fiber Transmission (M4B)   16:30 - 18:30      
Capacity Planning (Tu2F)     14:00 - 16:00    
Panel: Machine Learning and SDN: Towards Intelligent Data Centers     14:00 - 16:00    
SDN/NFV Demonstration Zone (Tu3D)     16:30 - 18:30    
Transport Network Design (Tu3E)     16:30 - 18:30    
Machine Learning and Network Availability (W1D)       8:00 - 10:00  
Performance Monitoring and Nonlinear Transmission (W1G)       8:15 - 10:00  
Data Center Traffic Management (W4I)       16:30 - 18:30  
Deployable Transport Networks (W4A)       16:30 - 18:30  
Machine Learning for Network Control and Management (W4F)       16:30 - 18:30  
5G Transport (Th1B)         8:30 - 10:00
Application Awareness and Online Optimization (Th1D)         8:00 - 10:00

Machine Learning and Performance Monitoring (M2F)
Monday, March 12, 10:30 - 12:30

Integration of Multivariate Gaussian Mixture Model for Enhanced PAM-4 Decoding Employing Basis Expansion (M2F.1)
10:30 - 10:45     Contributed Submission
Presenter: Feng Lu, Georgia Institute of Technology, United States
For the first time, we introduced Gaussian-mixture-model (GMM) and basis expansion in the PAM-4 decoder after equalization to estimate and eliminate residual linear/nonlinear impairments. With 100-Gbps testbed, we experimentally demonstrated 1 to 1.5-dB sensitivity improvement.

Transmission of 4×50-Gb/s PAM-4 Signal over 80-km Single Mode Fiber using Neural Network (M2F.2)
10:45 - 11:00     Contributed Submission
Presenter: Xiang Li, WRI, China
We apply neural network algorithms with two kinds of input features as the nonlinear DSP method, and realize 4×50-Gb/s PAM-4 IM/DD transmission for 80-km fiber with ~2 dB power sensitivity improvement over conventional nonlinear algorithms.

Optical Performance Monitoring in Fiber-Optic Networks Enabled by Machine Learning Techniques (M2F.3)
11:00 - 11:30     Invited Submission
Presenter: Faisal Khan, The Hong Kong Polytechnic University, Hong Kong
We review applications of machine learning (ML) in various aspects of optical communications including optical performance monitoring, fiber nonlinearity compensation, and software-defined networking. The future role of ML in optical communications is also discussed.

Joint Estimation of Linear and Nonlinear Signal-to-Noise Ratio based on Neural Networks (M2F.4)
11:30 - 11:45     Contributed Submission
Presenter: Francisco Javier Vaquero Caballero, University of Cambridge, Spain
A novel technique estimating ASE and non-linear SNR is presented. Our method is evaluated by simulations obtaining a std error of 0.23 dB for both ASE and non-linear SNR

Blind and Fast Modulation Format Identification by Frequency-offset Loading for Hitless Flexible Transceiver (M2F.5)
11:45 - 12:00     Contributed Submission
Presenter: Lu Jianing, School of Optical and Electronic Information, Huazhong University of Science and Technology, Department of Information Engineering, The Chinese University of Hong Kong, United States
We propose a blind and fast modulation format identification enabled by frequency-offset loading technique for hitless flexible transceiver. Arbitrary format switching among 28-Gbaud DP-4/8/16/64QAM, hybrid-4/16QAM, and SP-128QAM can be accurately identified with only 2048 symbols.

An Accurate and Robust PDL Monitor by Digital Signal Processing in Coherent Receiver (M2F.6)
12:00 - 12:15     Contributed Submission
Presenter: Shoichiro Oda, Fujitsu Laboratories Ltd., Japan
An accurate and robust PDL monitor based on receiver side DSP is proposed and experimentally verified. The standard deviation of monitor error is 0.1 dB and the method tolerates various imperfections, such as unsynchronized clock.

An Accurate Algorithm to Quantitatively Identify the Performance Degradation Caused by Linear Crosstalk (M2F.7)
12:15 - 12:30     Contributed Submission
Presenter: Xiaofei Su, Fujitsu R&D center, China
A new algorithm to estimate the signal impairment by linear crosstalk is proposed. Experiments demonstrate the linear relationship between the Q penalty and crosstalk indicator with R2=0.98 under various transmission distances and conditions.


Probabilistic Shaping I (M3C)
Monday, March 12, 14:00 - 16:00

Probabilistic Constellation Shaping: Challenges and Opportunities for Forward Error Correction (M3C.1)
14:00 - 14:30     Invited Submission
Presenter: Laurent Schmalen, Nokia Bell Labs, Germany
We discuss the impact of probabilistic amplitude shaping on the forward error correction stage in future optical transceivers and highlight some potential pitfalls, as well as challenges and opportunities for future research that may arise.

Balancing Probabilistic Shaping and Forward Error Correction for Optimal System Performance (M3C.2)
14:30 - 15:00     Invited Submission
Presenter: Junho Cho, Nokia Bell Labs, United States
We review probabilistic shaping (PS) and forward error correction (FEC) and study the interplay between the two to jointly maximize the overall information rate.

Combining Probabilistic Shaping and Nonlinear Mitigation: Potential Gains and Challenges (M3C.3)
15:00 - 15:15     Contributed Submission
Presenter: Fernando Guiomar, Politecnico di Torino, Portugal
We experimentally compare different options for transmission at 200G net bit-rate and demonstrate that the benefits of probabilistic shaping and nonlinear mitigation via SRO and/or DBP can be effectively combined to enable propagation reach enhancement of >40%.

A Simple Nonlinearity-Tailored Probabilistic Shaping Distribution for Square QAM (M3C.4)
15:15 - 15:30     Contributed Submission
Presenter: Eric Sillekens, University College London, United Kingdom
A new probabilistic shaping distribution that outperforms Maxwell-Boltzmann is studied for the nonlinear fiber channel. Additional gains of 0.1 bit/symbol MI or 0.2 dB SNR for both DP-256QAM and DP-1024QAM are reported after 200 km nonlinear fiber transmission.

Experimental and Numerical Comparison of Probabilistically-Shaped 4096 QAM and Uniformly-Shaped 1024 QAM in All-Raman Amplified 160 km Transmission (M3C.5)
15:30 - 15:45     Contributed Submission
Presenter: Seiji Okamoto, Tohoku University, Nippon Telegraph & Telephone Corp, Japan
We experimentally and numerically compared probabilistically-shaped 4096 QAM with uniformly-shaped 1024 QAM with a spectral efficiency of 15.3 bit/s/Hz after an all-Raman 160 km transmission. The 1.4-dB power margin was improved by constellation shaping.

Residual Non-Linear Phase Noise in Probabilistically Shaped 64-QAM Optical Links (M3C.6)
15:45 - 16:00     Contributed Submission
Presenter: Dario Pilori, Politecnico di Torino, Italy
We show that the performance penalty for probabilistically-shaped constellations induced by residual non-linear phase noise after standard CPE is significantly higher in low-symbol rate systems, canceling the potential gain achievable through symbol rate optimization.


Telemetry and Survivability (M3A)
Monday, March 12, 14:00 - 16:00

Experimental Demonstration of Active and Passive Optical Networks Telemetry (M3A.1)
14:00 - 14:15     Contributed Submission
Presenter: Luis Velasco, Universitat Politecnica de Catalunya, Spain
A distributed architecture enabling active and passive optical network telemetry is presented; a YANG data model is used for remotely configuring monitoring devices for telemetry purposes. Experimental demonstration is carried out for failure localization.

Field Trial of Monitoring On-Demand at Intermediate-Nodes Through Bayesian Optimization (M3A.2)
14:15 - 14:30     Contributed Submission
Presenter: Fanchao Meng, University of Bristol, United Kingdom
We demonstrate an intelligent monitoring on-demand switching strategy at network nodes using Bayesian optimization. It is shown our proposed method achieves identical monitoring capability as complete system exploration while saving a lot of data.

Cognitive Tool for Estimating the QoT of New Lightpaths (M3A.3)
14:30 - 14:45     Contributed Submission
Presenter: Christine Tremblay, École de technologie supérieure, Canada
We propose a cognitive tool which takes into account linear and nonlinear impairments for QoT estimation and we perform a comparative analysis of three machine-learning techniques for lightpath classification using synthetic BER data.

Softwarized, elastic and agile optical networks for dynamic environmental change and failure recovery (M3A.4)
14:45 - 15:15     Invited Submission
Presenter: Hiroaki Harai, National Inst of Information & Comm Tech, Japan
We present elasticity and agility in softwarized optical network construction and service continuation, where existing services are kept transient quality against sudden traffic changes and failures. We show proper optical power management in network protection.

Machine-Learning-Based Soft-Failure Detection and Identification in Optical Networks (M3A.5)
15:15 - 15:30     Contributed Submission
Presenter: Massimo Tornatore, Politecnico di Milano, Italy
We develop and test several machine-learning methods to perform detection and identification of equipment failures in optical networks. Results, obtained over real BER traces, show above 98% accuracy in most cases with reasonable algorithm complexity.

Experimental Validation of Transport SDN Restoration of Signal-Degraded Connections in Flexi-Grid Networks (M3A.6)
15:30 - 15:45     Contributed Submission
Presenter: Ricardo Martínez, Ctr Tecnologic de Telecoms de Catalunya, Spain
We validate experimentally the integration of a PCE-based T-SDN controller with an OAM Handler to restore signal-degraded flexi-grid connections. A distributed monitoring system is adopted where a RSMA algorithm exploits benefits of elastic optical networks.

Observe-Decide-Act: Experimental Demonstration of a Self-Healing Network (M3A.7)
15:45 - 16:00     Contributed Submission
Presenter: Konstantinos Christodoulopoulos, Computer Technology Institute and Press, Greece
We experimentally demonstrate a self-healing network, following the observe-decide-act paradigm: monitoring reveals degradation, decision is taken, network is reconfigured to restore service. We also demonstrate, for the first time, pre-programmed resilience in integrated data/control testbed.


Modelling, Disaggregation & Networking Automation (M4A)
Monday, March 12, 16:30 - 18:30

Software Defined Optical Network from the Perspective of a Software Developer (M4A.1)
16:30 - 17:00     Invited Submission
Presenter: Lihua Yuan, Microsoft Corp, United States
Azure invests heavily in software automation to help the network evolving towards the desired scalability and reliability. This talk shares our learnings through this evolution and discuss the impact to optical infrastructure.

Node Internal Modeling for Network Recovery with Emergency Optical Systems (M4A.2)
17:00 - 17:15     Contributed Submission
Presenter: Sugang Xu, NICT, Japan
We propose a ROADM internal model and demonstrate network recovery with the control of a lightweight emergency optical system, which has a highly customizable internal structure to meet the different requirements in disaster recovery.

Experimental demonstration of fully disaggregated white box including different types of transponders and monitors, controlled by NETCONF and YANG (M4A.3)
17:15 - 17:30     Contributed Submission
Presenter: Nicola Sambo, Sant' Anna di Pisa, Italy
We experimentally demonstrated a fully disaggregated white box composed of two different types of transponders, monitors (including filtering effect parameters), add-drop multiplexers, and switches. NETCONF and YANG control the hardware.

OpenROADM over ONOS (M4A.4)
17:30 - 18:00     Invited Submission
Presenter: Marc De Leenheer, Open Networking Foundation, United States
We discuss the integration of OpenROADM data models into ONOS, the network OS built for high performance, scale and high availability. We review experimental data, deployment considerations, lessons learned, and next steps.

P4-based Multi-Layer Traffic Engineering Encompassing Cyber Security (M4A.5)
18:00 - 18:15     Contributed Submission
Presenter: Filippo Cugini, CNIT, Italy
The design and experimental validation of effective P4-based solutions for dynamic optical bypass, traffic offload and cyber security without involving SDN controller are proposed. P4 switch latency scalability versus installed flow entry size is demonstrated.

Fast and Accurate Lightpath Validation for SDN Controllers (M4A.6)
18:15 - 18:30     Contributed Submission
Presenter: Rui Morais, Coriant Portugal, Instituto De Telecomunicacoes, United States
We present a method to quickly and accurately evaluate the feasibility of an optical path in online routing scenarios. The method is based on performance metrics that compress the set of all feasible paths.


Deployments of Optical Access and Front Haul (M4I)
Monday, March 12, 16:30 - 18:30

Edge Compute and AT&T’s PON Deployment Vision (M4I.1)
16:30 - 17:30     Tutorial Submission
Presenter: Edward Walter, AT&T, United States
This tutorial talk will review the Open Mobile Edge Cloud (OMEC) as a unified design that encompass aspects of various initiatives’ already undertaken in academia/industry and Passive Optical Networks (PON) in support of consumer, business, and infrastructure deployments.

Experimental Demonstration of 100 Gb/s Optical Network Transport and Aggregation for Ethernet Fronthaul with Low and Bounded Delay (M4I.2)
17:30 - 17:45     Contributed Submission
Presenter: Raimena Veisllari, TransPacket, Norway
3-node integrated packet/circuit network experiment demonstrates 100Gb/s transport and aggregation of five 10Gb/s links with low and bounded delay. 3.4µs maximum end-to-end delay is achieved, even when combining with less delay-sensitive traffic, reaching 98% utilization.

Coordinating Multi-access Edge Computing with Mobile Fronthaul for Optimizing 5G End-to-End Latency (M4I.3)
17:45 - 18:00     Contributed Submission
Presenter: Wei Wang, Beijing Univ. of Posts and Telecomm., University of California, Davis, China
In 5G, latency-sensitive traffic might be processed directly at central-offices by Multi-access Edge Computing (MEC) right after being transported through Mobile Fronthaul (MFH). We investigate how to optimize end-to-end latency by coordinating MEC with MFH.

Scenarios and Economic Analysis of Fronthaul (M4I.4)
18:00 - 18:30     Invited Submission
Presenter: Andrea Di Giglio, Telecom Italia Lab,
This work presents a comparison between C-RAN and emerging DA-RAN approaches for 5G networks. Main results show that DA-RAN and splitting alternative to CPRI bring economical and architectural advantages.


Probabilistic Shaping II (M4E)
Monday, March 12, 16:30 - 18:30

On Joint Design of Probabilistic Shaping and Forward Error Correction for Optical Systems (M4E.1)
16:30 - 17:30     Tutorial Submission
Presenter: Georg Böcherer, Huawei Technologies, Germany
Probabilistic shaping for rate/reach increase and flexibility in optical systems is treated. Practical design tools are developed for distribution matching algorithms, demapping strategies, and FEC codes, accounting for the performance-complexity trade-off.

Low-Complexity Variable-Length Output Distribution Matching with Periodical Distribution Uniformalization (M4E.2)
17:30 - 17:45     Contributed Submission
Presenter: Tsuyoshi Yoshida, Chalmers University of Technology, Mitsubishi Electric Corporation, Japan
Run-length code based distribution matching (DM) for probabilistic shaping is combined with a uniformalizer to realize low-complexity fixed-length DM. The proposed method is 0.4 dB better than previous low-complexity DM methods.

Experimental Verification of Rate Flexibility and Probabilistic Shaping by 4D Signaling (M4E.3)
17:45 - 18:00     Contributed Submission
Presenter: Fabian Steiner, Technical University of Munich, Germany
The rate flexibility and probabilistic shaping gain of 4-dimensional signaling is experimentally tested for short-reach, unrepeated transmission. A rate granularity of 0.5 bits/QAM symbol is achieved with a distribution matcher based on a simple look-up table.

Universal Hybrid Probabilistic-geometric Shaping Based on Two-dimensional Distribution Matchers (M4E.4)
18:00 - 18:15     Contributed Submission
Presenter: Zhen Qu, NEC Laboratories America, University of Arizona, United States
We propose universal distribution matchers applicable to any two-dimensional signal constellation. We experimentally demonstrate that the performance of 32-ary QAM, based on hybrid probabilistic-geometric shaping, is superior to probabilistically shaped 32QAM and regular 32QAM.

Efficient Offline Evaluation of FEC Codes Based on Captured Data with Probabilistic Shaping (M4E.5)
18:15 - 18:30     Contributed Submission
Presenter: Tsuyoshi Yoshida, Chalmers University of Technology, Mitsubishi Electric Corporation, Japan
We propose a tool for reusing experimental or simulation data of probabilistically shaped signals with different FEC codes. A single recorded histogram of log-likelihood ratios is sufficient to examine arbitrary coding at low BERs.


Practical Aspects in Fiber Transmission (M4B)
Monday, March 12, 16:30 - 18:30

Ultra-low Loss Silica Core Fiber (M4B.1)
16:30 - 17:00     Invited Submission
Presenter: Yoshiaki Tamura, sumitomo electric industry, Japan
We review recent advances in optical fiber technologies that realize ultra-low losses as low as 0.142 dB/km including pure-silica core and reduction of microscopic density fluctuation.

Observation of Guided Acoustic-Wave Brillouin Scattering and its Digital Compensation in Coherent QAM Transmission (M4B.2)
17:00 - 17:15     Contributed Submission
Presenter: Masataka Nakazawa, Tohoku University, Japan
Guided acoustic-wave Brillouin scattering (GAWBS) noise was observed in a digital coherent QAM transmission, which deteriorated the transmission performance. A new digital signal processing method is presented that compensates for the GAWBS noise.

Impact of Spontaneous Guided Acoustic-Wave Brillouin Scattering on Long-haul Transmission (M4B.3)
17:15 - 17:30     Contributed Submission
Presenter: Maxim Bolshtyansky, TE SubCom, United States
We measure forward optical scattering coefficient of thermally excited acoustic modes of modern submarine transmission fiber to be -32 dB/Mm. We estimate 0.6 dB of Q-factor penalty in our testbed due to this effect.

Proactive Detection of Optical Cable Failure Caused by Water Freezing Using 1-μm-band Mode-Detection OTDR (M4B.4)
17:30 - 17:45     Contributed Submission
Presenter: Atsushi Nakamura, NTT Corporation, Japan
An experiment shows that information allowing the prediction of optical loss increases due to icing in cable conduits can be extracted by using OTDR measurements in the 2-LP mode region of the fibers under test.

Laboratory Measurements of SOP Transients due to Lightning Strikes on OPGW Cables (M4B.5)
17:45 - 18:00     Contributed Submission
Presenter: Fabio Pittalà, Huawei Technologies Duesseldorf GmbH, Germany
We report laboratory experiments in which lightning current impulses of magnitudes up to and above 150kA were applied to OPGW cables of length up to 18m. Unprecedentedly, fast SOP speeds exceeding 8Mrad/s were measured.

Requirements for simulation-aided design of SDM systems (M4B.6)
18:00 - 18:30     Invited Submission
Presenter: Igor Koltchanov, VPIphotonics, Germany
We discuss general requirements and implementation options for an efficient simulation framework supporting the design of SDM systems, as well as aspects regarding the analysis, comparison and optimization of underlying technology and component options.


Capacity Planning (Tu2F)
Tuesday, March 13, 14:00 - 16:00

The Software-Defined Flexible Optical Network (Tu2F.1)
14:00 - 15:00     Tutorial Submission
Presenter: António Eira, Coriant, Instituto de Telecomunicações, Portugal
This tutorial discusses how hardware and software flexibility affect the design of optical networks. It covers the interworking between multi-rate/flow interfaces, universal OTN switching, FlexE enabled clients, and how SDN controllers can leverage this flexibility effectively.

Predeployment of Transceivers for Dynamic Lightpath Provisioning in Translucent Flexgrid Optical Networks (Tu2F.2)
15:00 - 15:15     Contributed Submission
Presenter: Krzysztof Walkowiak, Wroclaw University of Science and Technology, Poland
We focus on translucent flexgrid optical networks with flexible back-to-back regeneration and analyze the impact of the number of available transceivers on network performance for several proposed transceiver location methods and dynamic routing algorithms.

Benefit of progressive deployment of regenerators along with traffic growth in WDM elastic networks (Tu2F.3)
15:15 - 15:30     Contributed Submission
Presenter: Thierry Zami, Nokia Corporation, France
We present a new strategy of regenerator placement along with traffic growth in the elastic WDM networks, by introducing additional regenerators on the already allocated connections only from when needed to accommodate extra demand of capacity

Long-Term Capacity Planning in Facebook Network (Tu2F.4)
15:30 - 16:00     Invited Submission
Presenter: Yuri Smirnov, Facebook Inc., United States
This paper discusses future traffic prediction and design of resilient communication networks as part of long-term capacity planning initiaitves at Facebook.


SDN/NFV Demonstration Zone (Tu3D)
Tuesday, March 13, 16:30 - 18:30

Integrated Optical-Wireless Resource Slicing Management for 5G Service-based Architecture and Multi-level RAN (Tu3D.1)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Rentao Gu, Beijing University of Posts and Telecommunications, Beijing University of Technology, China
This demo provides an integrated end-to-end optical-wireless slicing cross latest SBA based core network, backhaul network and two-level fronthaul networks, presenting consistent heterogeneous coordination capabilities on different multi-vendor transport equipment and computing resources.

Demonstration of Routing and Spectrum Assignment Automation in a Transport SDN Framework (Tu3D.2)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Srivatsan Balasubramanian, facebook inc, United States
We demonstrate a use case of transport SDN which can help with automation of routing and spectrum assignment thereby removing an error and delay prone manual planning phase.

High Performance Streaming Telemetry in Optical Transport Networks (Tu3D.3)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Abhinava Sadasivarao, Infinera Corporation, United States
We demonstrate streaming telemetry capabilities for optical networks implemented as modular software service. The telemetry system is capable of user-defined configurable streaming to an external collector at very high frequencies of all critical optical performance metrics.

Demonstration of Real Time VNF Implementation of OLT with Virtual DBA for Sliceable Multi-Tenant PONs (Tu3D.4)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Frank Slyne, Trinity College Dublin, Ireland
We demonstrate the VNF implementation of a sliceable PON architecture enabling true multi-tenancy, giving Virtual Network Operators full control over capacity scheduling. We analyze resource sharing efficiency and latency performance for different NFV co-location scenarios.

CASTOR: An Architecture to Bring Cognition to Transport Networks (Tu3D.5)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Luis Velasco, Universitat Politecnica de Catalunya, Spain
CASTOR architecture to enable cognitive networking is demonstrated. Extended nodes make local decisions, whilst a centralized system beside the network controller makes network-wide decisions. Interaction with ONOS, Net2Plan, and passive monitoring devices is exhibited.

Flow/Application Triggered SDN control in Hybrid Data-center Network “HOLST” (Tu3D.6)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Yukihiro Imakiire, Keio University, Japan
We have proposed a new data-center network architecture “HOLST” with dynamic and adaptive network configuration for reducing power consumption. This demonstration shows the method of detecting and assigning flows from traffic and application trigger.

Topology Description Generation and Path Computation Framework for Dynamic Optical Path Network with Heterogeneous Switches (Tu3D.7)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Kiyo Ishii, AIST, Japan
A topology description scheme, a topology generation system, and a path computation system for dynamic optical path networks which support dynamic generation of wide area performance-guaranteed network slices will be demonstrated.

Towards IP & Transport Network Transformation Using Standardized Transport NorthBound Interfaces (Tu3D.8)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Ricard Vilalta, CTTC, Spain
This demo proposes the usage of standardized YANG data models for multi-vendor and multi-layer optical control interoperability. L2/L3 network service establishment will be demonstrated as part of the network transformation strategy for SDN/NFV.

Automated Management and Control of a Multi-Vendor Disaggregated Network at the L0 Layer (Tu3D.9)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Omer Yilmaz, Infinera Corporation, United States
We propose to demonstrate automated service management and automated optical power controls over Infinera and Lumentum Open Optical Line Systems using an Infinera Layer-0 SDN controller.

Joint Optimal Service Chain Allocation, VNF instantiation and Metro Network Resource Management Demonstration (Tu3D.10)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Francisco-Javier Moreno-Muro, Politechnical University of Cartagena, Spain
In a metro network with VIMs orchestrated by an ETSI-OSM instance, and an optical transport controller, we demonstrate optimized service chain provisioning using the open-source Net2Plan tool with interfaces to OSM (new) and transport controller

Network Slicing Resource Allocation and Monitoring over Multiple Clouds and Networks (Tu3D.11)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Ricardo Martínez, Ctr Tecnologic de Telecoms de Catalunya, Spain
This demo presents an in-operation network slice resource allocator, which is able to consider networking and cloud infrastructure. Monitoring cloud and network resources allows enhancing the (re-)allocation of network slices, while accommodating novel slice requests.

Fully Disaggregated ROADM White Box with NETCONF/YANG control, Telemetry, and Machine Learning-based Monitoring (Tu3D.12)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Filippo Cugini, CNIT, Italy
A first demonstration of ROADM White Box augmented with machine learning capabilities is demonstrated. The white box includes various level of disaggregation, NETCONF/YANG control, telemetry and spectrum-based advanced monitoring functionalities.

O2CMF: Experiment-as-a-Service for Agile Fed4Fire Deployment of Programmable NFV (Tu3D.13)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Isabella Ceravolo, Federal University of Espirito Santo (UFES), Brazil
An open platform over OpenStack for control and management of experiments (O2CMF) for merging and adapting wireless and optical federated testbeds with proper cloud infrastructure is presented. TOSCA-enabled orchestration provides programmability for NFV experiments.

Demonstration of NFV for Mobile Edge Computing on an Optically Disaggregated Datacentre in a Box (Tu3D.14)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Michael Enrico, HUBER+SUHNER Polatis Ltd, United Kingdom
This demonstrator showcases the hardware and software integration achieved by the dReDBox project [1] towards realization of a novel architecture using dynamically-reconfigurable optical interconnects to create a flexible, scalable and efficient disaggregated datacentre infrastructure.

Network Orchestration for Dynamic Network Slicing for Fixed and Mobile Vertical Services (Tu3D.15)
16:30 - 18:30     SDN & NFV Demo Submission
Presenter: Rodolfo Alvizu, SWAN networks, Italy
We demonstrate how a hybrid and hierarchical transport-SDN control plane based on a network orchestrator and an SDN controller can provide dynamic network slicing for enterprise-networking services and mobile metro-core networks.


Transport Network Design (Tu3E)
Tuesday, March 13, 16:30 - 18:30

Networking Benefit of Multi-Subcarrier Transceivers (Tu3E.1)
16:30 - 16:45     Contributed Submission
Presenter: Mattia Cantono, Politecnico di Torino, Italy
We analyze the benefit of multi-subcarrier transceivers on two network topologies. We describe nonlinearities modeling challenges and demonstrate the existence of an optimal subcarrier symbol rate at network-level yielding average OSNR increase up to 0.7dB.

Remote Abstraction of an Installed Dark Fiber Network using Noise to Signal Ratio (Tu3E.2)
16:45 - 17:00     Contributed Submission
Presenter: David Ives, University of Cambridge, United Kingdom
A dark fiber network was partially abstracted utilizing network monitors, prior information and a single probe channel. Validation employed 13x200Gb/s DWDM signals transmitted at the optimum launch power, with measured performance 0.2dB better than abstracted.

Data analytics based optical performance monitoring technique for optical transport networks. (Tu3E.3)
17:00 - 17:30     Invited Submission
Presenter: Takahito Tanimura, Fujitsu Laboratories Ltd., The University of Tokyo, Japan
We experimentally demonstrate that a convolutional neural network (CNN) can acquire an accurate OSNR estimation functionality from asynchronously sampled data right after intradyne coherent detection, while the feature selection before the CNN can be omitted.

Towards a Route Planning Tool for Open Optical Networks in the Telecom Infrastructure Project (Tu3E.4)
17:30 - 17:45     Contributed Submission
Presenter: Brian Taylor, Facebook, United States
We explore the validity of the Gaussian-noise (GN) model as the basis for an open-source optical network planning tool. Comparison of experimental results and TIP-PSE GN-model based predictions suggest the GN model a feasible choice.

Demonstrating Network-scale Gain Transient Impact of Multiple Series EDFAs in Link Failure Cases (Tu3E.5)
17:45 - 18:00     Contributed Submission
Presenter: Yusuke Hirota, NICT, Japan
This paper identifies that a link failure causes temporal QoT degradation propagation of 40.2% existing paths in a whole network by conventional EDFA and experimentally demonstrates this transient degradation can be suppressed by burst-mode EDFA.

Design and Deployment of Optical White Box (Tu3E.6)
18:00 - 18:30     Invited Submission
Presenter: Niall Robinson, ADVA Optical Networking AG, United States
White box solutions have been deployed inside data centers for many years. Like SDN this technological approach is spilling into the transmission world. Let’s review current industry status and explore the future of this exciting space.


Machine Learning and Network Availability (W1D)
Wednesday, March 14, 8:00 - 10:00

Guaranteed-Availability Network Function Virtualization in Inter-Datacenter Networks (W1D.1)
8:00 - 8:15     Contributed Submission
Presenter: Jian Kong, The University of Texas, Dallas, United States
Considering the availability of the datacenter’s network elements, we propose a coordinated protection mechanism that adopts both backup path protection and SFC replicas distributed among datacenters to support high availability while reducing total cost.

Data Analytics and Machine learning applied to transport layer (W1D.2)
8:15 - 9:15     Tutorial Submission
Presenter: Massimo Tornatore, Politecnico di Milano, University of California, Davis, Italy
After a generic introduction to machine-learning concepts and tools, some applications of machine learning in optical networks are introduced and discussed, with a focus on QoT estimation, failure detection and failure identification.

Q-availability based Virtual Optical Network Provisioning (W1D.3)
9:15 - 9:30     Contributed Submission
Presenter: Martin Bouda, Fujitsu Laboratories of America, United States
We show significant gain in Virtual Optical Network (VON) capacity by provisioning based on Q-availabilities of optical channels instead of fixed Q-margin. The Q-availability is calculated using stochastic simulations based on GN model.

Real-time Spectrum Surveillance in Filterless Optical Networks (W1D.4)
9:30 - 9:45     Contributed Submission
Presenter: Luis Velasco, Universitat Politecnica de Catalunya, Spain
A monitoring system exploiting data analytics and cost-effective optical spectrum analyzers with under 1.2GHz resolution is proposed. Its performance is demonstrated in a filterless network experimental test-bed. The system enables prompt action before lightpath disruption.

Applying Data Visualization for Failure Localization (W1D.5)
9:45 - 10:00     Contributed Submission
Presenter: Luis Velasco, Universitat Politecnica de Catalunya, Spain
Data visualization is applied to BER measures in a bigdata repository. Bubble charts are produced to identify lightpaths with increasing BER and spectrum color maps are then used to identify the most likely degraded link.


Performance Monitoring and Nonlinear Transmission (W1G)
Wednesday, March 14, 8:15 - 10:00

Learning from the Optical Spectrum: Soft-Failure Identification and Localization [Invited] (W1G.1)
8:15 - 8:45     Invited Submission
Presenter: Luis Velasco, Universitat Politecnica de Catalunya, Spain
The availability of coarse-resolution cost-effective Optical Spectrum Analyzers (OSA) allows its widespread deployment in operators’ networks. In this paper, several machine learning approaches for failure identification and localization that take advantage of OSAs are presented.

White Gaussian Noise Based Capacity Estimate and Characterization of Fiber-Optic Links (W1G.2)
8:45 - 9:00     Contributed Submission
Presenter: Roland Ryf, Nokia Bell Labs, United States
We use white Gaussian noise as a test signal for single-mode and multimode transmission links and estimate the link capacity based on a calculation of mutual information. We also extract the complex amplitude channel estimations and mode-dependent loss with high accuracy.

Non-Linearity Modeling at Ultra-High Symbol Rates (W1G.3)
9:00 - 9:15     Contributed Submission
Presenter: Pierluigi Poggiolini, Politecnico di Torino, Italy
We investigate the accuracy of several versions of the GN and EGN models for symbol rates up to 426 GBaud, over about 2THz of optical bandwidth, considering PM-32/64/128/256 QAM and addressing non-linear phase-noise mitigation.

Observing the Interaction of PMD with Generation of NLI in Uncompensated Amplified Optical Links (W1G.4)
9:15 - 9:30     Contributed Submission
Presenter: Mattia Cantono, Politecnico di Torino, Italy
We present simulative analyses displaying how PMD has negligible interaction with the generation of nonlinear interference, and show that the GN-model is accurate yet conservative in estimating the generalized SNR also in wide-bandwidth transmission.

A Novel Detection Strategy for Nonlinear Frequency-Division Multiplexing (W1G.5)
9:30 - 9:45     Contributed Submission
Presenter: Stella Civelli, Tecip, Scuola Superiore Sant'Anna, Consorzio Nazionale Interuniversitario per le Telecomunicazioni, Italy
A novel decision feedback detection strategy exploiting a causality property of the nonlinear Fourier transform is introduced. The novel strategy achieves a considerable performance improvement compared to previously adopted strategies in terms of Q-factor.

100 Gbps b-modulated Nonlinear Frequency Division Multiplexed Transmission (W1G.6)
9:45 - 10:00     Contributed Submission
Presenter: Son Le, Nokia Bell Labs, Germany
We demonstrate that the performance of 100 Gb/s NFDM system employing the continuous nonlinear spectrum is significantly enhanced by modulating only the b-coefficient, providing a gain up to 1.5 dB over the conventional OFDM system


Data Center Traffic Management (W4I)
Wednesday, March 14, 16:30 - 18:30

Network Traffic Characteristics of Data Centers (W4I.1)
16:30 - 17:00     Invited Submission
Presenter: Theophilus Benson, Duke University, United States
Although there is tremendous interest in designing improved networks for data centers, very little is known about the network-level traffic characteristics of current data centers. In this paper, we conduct an empirical study of the network traffic in 10 data centers belonging to three different types of organizations, including university, enterprise, and cloud data centers. Our definition of cloud data centers includes not only data centers employed by large online service providers offering Internet-facing applications, but also data centers used to host data-intensive (MapReduce style) applications. We collect and analyze SNMP statistics, topology, and packet-level traces. We examine the range of applications deployed in these data centers and their placement, the flow-level and packet-level transmission properties of these applications, and their impact on network utilization, link utilization, congestion, and packet drops. We describe the implications of the observed traffic patterns for data center internal traffic engineering as well as for recently proposed architectures for data center networks.

Leveraging Deep Learning to Achieve Efficient Resource Allocation with Traffic Evaluation in Datacenter Optical Networks (W4I.2)
17:00 - 17:15     Contributed Submission
Presenter: Ao Yu, BUPT, China
This paper first presents a deep learning-based resource allocation strategy supported by global evaluate factor in intra-datacenter optical networks. Numerical results show the proposed strategy improves traffic prediction accuracy and has superior performance.

A Novel Buffering Design and Performance Evaluation of Optical Flow Switch with Smart Scheduling Algorithms (W4I.3)
17:15 - 17:30     Contributed Submission
Presenter: Yhi Shiau, TL, Chunghwa Telecom Co., Ltd., United States
In this paper, buffering design and performance evaluation of OFS with smart scheduling algorithms are proposed. The buffer scheduling methods are used for a hybrid switching module or all-optical switching module on the data center.

Optical Networks Throughput Enhancement via TCP Stop-and-Wait on Hybrid Switches (W4I.4)
17:30 - 17:45     Contributed Submission
Presenter: Artur Minakhmetov, Telecom ParisTech, France
We report on possible 50% throughput increase in Optical Packet Switching (OPS) data-center networks by replacing all-optical switches with optical switches with shared electronic buffers further enhanced with TCP Stop-and-Wait algorithms.

Optical Circuit Switching Enabled Reconfigurable HPC Network for Traffic Pattern (W4I.5)
17:45 - 18:00     Contributed Submission
Presenter: Yu Shang, BUPT, China
We propose a reconfigurable hybrid network architecture using optical circuit switches for HPC systems, and evaluate the performance of different topologies according to different traffic patterns through high throughput and low latency.

Role of Standards in Web-Scale Datacenters (W4I.6)
18:00 - 18:30     Invited Submission
Presenter: Mark Filer, Microsoft Corp., United States
Standards play an increasingly important role for cloud providers considering the dramatic growth that cloud services are experiencing. Distinctions are made between open consortia, multi-source agreements, and standards, and case studies with lessons learned are presented.


Deployable Transport Networks (W4A)
Wednesday, March 14, 16:30 - 18:30

Benefits of Performance Awareness in Coherent Dynamic Optical Networks (W4A.1)
16:30 - 17:00     Invited Submission
Presenter: Juraj Slovak, Coriant R&D,
Awareness of optical network performance and its accurate prediction enables squeezing of margins allocated for stable operation. We discuss potential of online performance assessment by leveraging functionalities of bandwidth variable transponders in flexible optical networks.

Extension of SDN networks to the Satellite Domain; Integration of a SDN enabled WAN network, with terrestrial and submarine elements, with command and control of multiple satellite constellations (W4A.2)
17:00 - 17:30     Invited Submission
Presenter: Robert Kimball, Ciena Corporation, United States
Dynamic software controlled terrestrial fiber networks combined with an agile satellite communications (SATCOM) service management system increases system resiliency; bringing resources into the operational environment. This includes failover of critical communications paths over different technologies.

Capacity Enhancement in Optical Networks using Margin Extraction (W4A.3)
17:30 - 17:45     Contributed Submission
Presenter: Mohammad Sheikh Zefreh, Ciena, Canada
In this paper, the value of margin extraction in optical networks is investigated. Simulation results for a sample network show that up to 64% of multi-rate transponders can run with higher rates using extracted margins.

Demonstration of Automatic Connection Pair Discovery and Path Setting in Filter-less Point-to-Point WDM Systems (W4A.4)
17:45 - 18:00     Contributed Submission
Presenter: Yutaka Takita, Fujitsu Labs. Ltd., Japan
We propose the filter-less point-to-point WDM system with IP-layer driven automatic optical connection pair discovery and path setting function. We confirmed its basic operation in a 100km span 32-lambdas WDM system with enhanced transponders.

How Much Is CD ROADM Contention Blocking? (W4A.5)
18:00 - 18:15     Contributed Submission
Presenter: Guangzhi Li, Futurewei Technology Inc., United States
This paper models CD ROADM network contention blocking and shows that the contention blocking impact is marginal when CD ROADM add/drop port usage is less than 75%. Numerical results verify the effectiveness.

Influence of the maturity of technology on the benefit of 75 GHz-spaced 64 GBaud channels in WDM elastic networks (W4A.6)
18:15 - 18:30     Contributed Submission
Presenter: Thierry Zami, Nokia Corporation, France
We quantify how the benefit of 75 GHz-spaced 64 GBaud channels in WDM core networks depends on both the associated elastic modulation set and the maturity of transponder implementation correlated to the WDM transmission reaches


Machine Learning for Network Control and Management (W4F)
Wednesday, March 14, 16:30 - 18:30

Realizing AI-Assisted Multi-Layer Restoration in a Software-Defined IP-over-EON with Deep Learning: An Experimental Study (W4F.1)
16:30 - 16:45     Contributed Submission
Presenter: Siqi Liu, Univ of Science and Technology of China, China
By using deep learning, we experimentally demonstrate AI-assisted multi-layer restoration in an IP-over-EON, which recovers affected traffic timely with congestion-avoidance rerouting.

Deep-RMSA: A Deep-Reinforcement-Learning Routing, Modulation and Spectrum Assignment Agent for Elastic Optical Networks (W4F.2)
16:45 - 17:00     Contributed Submission
Presenter: Xiaoliang Chen, University of California Davis, United States
This paper demonstrates Deep-RMSA, a deep reinforcement learning based self-learning RMSA agent that can learn successful policies from dynamic network operations while realizing cognitive and autonomous RMSA in EONs.

ANN-Based Transfer Learning for QoT Prediction in Real-Time Mixed Line-Rate Systems (W4F.3)
17:00 - 17:15     Contributed Submission
Presenter: Weiyang Mo, The University of Arizona, NEC Laboratories America, United States
Quality of transmission prediction for real-time mixed line-rate systems is realized using artificial neural network based transfer learning with SDN orchestrating. 0.42 dB accuracy is achieved with a 1000 to 20 reduction in training samples.

Machine learning-assisted management of a virtualized network (W4F.4)
17:15 - 17:45     Invited Submission
Presenter: Michiaki Hayashi, KDDI Research Inc., Japan
Toward proactive network management, machine learning is expected especially from enhancing fault management and sustaining automated system itself. This paper describes the motives for using machine learning, and provides some proof-of-concept demonstrations.

Field Trial of Gaussian Process Learning of Function-Agnostic Channel Performance Under Uncertainty (W4F.5)
17:45 - 18:00     Contributed Submission
Presenter: Fanchao Meng, University of Bristol, United Kingdom
We successfully demonstrated a novel performance learning method using monitoring and Gaussian process. After 436km dark fiber transmission the model captures most of the test data with 0.7dB mean error and enables robust QoT predictor.

Analytics-Driven Fault Discovery and Diagnosis for Cognitive Root Cause Analysis (W4F.6)
18:00 - 18:15     Contributed Submission
Presenter: Danish Rafique, ADVA Optical Networking, Germany
We propose an SDN-integrated framework for distributed cognitive fault discovery and diagnosis across end-to-end network infrastructure. The approach is evaluated on a real-world use-case and proactively identifies root cause of autonomously-detected optical power level anomalies.

Experimental Demonstration of Cognitive Provisioning and Alien Wavelength Monitoring in Multi-domain EON (W4F.7)
18:15 - 18:30     Contributed Submission
Presenter: Roberto Proietti, University of California Davis, United States
This paper proposes a cognitive multi-domain EON architecture with machine-learning aided RMSA and alien wavelength monitoring. Testbed experiments show modulation format recognition, QoT monitoring and cognitive routing for a 160 GBd alien multi-wavelength lightpath.


5G Transport (Th1B)
Thursday, March 15, 8:30 - 10:00

Access Network Economics: A Total-Cost-of-Ownership Perspective (Th1B.1)
8:30 - 8:45     Contributed Submission
Presenter: Christoph Lange, Deutsche Telekom, Germany
Several fiber-based access solutions are compared regarding TCO and NPV using an analytic framework and real service area data. Fiber-copper and fixed-wireless solutions show improved TCO and NPV efficiency short-term, FTTH in longer-term constellations.

Software Defined 5G Converged Access as a viable Techno-Economic Solution (Th1B.2)
8:45 - 9:00     Contributed Submission
Presenter: Andrea Marotta, University of L'Aquila, Italy
Software Defined Converged Access represents a feasible solution to effectively address 5G traffic demands. This paper proposes an integrated mobile-optical control for wavelength and bandwidth allocation. Evaluations of bandwidth utilization and techno-economic viability are provided.

DBA Capacity Auctions to Enhance Resource Sharing across Virtual Network Operators in Multi-Tenant PONs (Th1B.3)
9:00 - 9:15     Contributed Submission
Presenter: Marco Ruffini, CONNECT Center, Trinity College Dublin, Ireland
We propose an economic-robust auction mechanism for multi-tenant PON's capacity sharing that operates within the DBA process. We demonstrate that our mechanism improves PON utilization by providing economic sharing incentives across VNOs and infrastructure providers.

Deep Neural Network Based Dynamic Resource Reallocation of BBU Pools in 5G C-RAN ROADM Networks (Th1B.4)
9:15 - 9:30     Contributed Submission
Presenter: Weiyang Mo, The University of Arizona, United States
An LSTM network is developed to predict BBU pool traffic in 5G C-RAN ROADM networks. 5G throughput improvement and resource savings are observed with resource reallocation by reconfiguring the optical network 30 minutes in advance.

Joint Optimization of BBU Pool Allocation and Selection for C-RAN Networks (Th1B.5)
9:30 - 9:45     Contributed Submission
Presenter: Yao Li, The University of Arizona, United States
BBU pool allocation and selection are jointly optimized for maximizing wireless traffic capacity while minimizing wavelength resource occupation in optical networks. Numerical results show optimal BBU pool locations under different traffic patterns and network capacities.

MixCo: Optimal Cooperative Caching for Mobile Edge Computing in Fiber-Wireless Access Networks (Th1B.6)
9:45 - 10:00     Contributed Submission
Presenter: Ning Wang, Soochow University, China
We consider the optimal content caching problem among Mobile Edge Computing (MEC) servers in a Fiber-Wireless (FiWi) access network, to minimize the average content delivery latency subject to limited storage and computing capacity of each server. An MILP model and a Mix-Cooperative (MixCo) caching strategy are developed for efficient performance.


Application Awareness and Online Optimization (Th1D)
Thursday, March 15, 8:00 - 10:00

The Role of Open-source Network Optimization Software in the SDN/NFV World (Th1D.1)
8:00 - 9:00     Tutorial Submission
Presenter: Pablo Pavon-Marino, Universidad Politécnica de Cartagena, E-lighthouse Network Solutions, Spain
SDN/NFV means an unprecedented network control, for an unprecedented resource dynamicity. Manual optimization is unmanageable. This tutorial covers open source optimization software initiatives for offline planning and online provisioning and orchestration of SDN/NFV networks.

Hysteresis-based Margin Allocation for Adaptive Coding in SDN-enabled Optical Networks (Th1D.2)
9:00 - 9:15     Contributed Submission
Presenter: Yao Li, The University of Arizona, United States
Hysteresis-based margin-allocation for adaptive coding is experimentally investigated against PDL-induced OSNR fluctuations in a SDN-enabled multi-domain optical network. Up to 90% network outage reduction can be observed under different margins.

An Automated Service-downgrade Negotiation Scheme for Application-centric Networks (Th1D.3)
9:15 - 9:30     Contributed Submission
Presenter: Antonio Marsico, FBK CREATE-NET, Italy
We propose a novel negotiation scheme for an application-driven relaxation of different requirements in multi-layer networks. Simulative results show that it improves service acceptance while keeping requirements' degradation much lower than applications' worst-case acceptable values.

Application aware multilayer control and optimization of elastic WDM switched optical networks (Th1D.4)
9:30 - 10:00     Invited Submission
Presenter: Ioannis Tomkos, Athens Information Technology Center, Greece
In dynamic networks with diverse application requirements, Software Defined Networking (SDN) principles enable application-aware in-operation planning. EU project ACINO built a network orchestrator as the connecting component between network applications and the underlying network infrastructure.

Sponsored by: