Springer optical networks and technologies 2005

Perspectives on Optical Networks and Technologies 1 Developments in Optical Seamless Networks Andrea Spaccapietra and Giovanni Razzetta invited paper 3 Cinema-class Digital Content Distr

OPTICAL NETWORKS AND TECHNOLOGIES

IFIP was founded in 1960 under the auspices of UNESCO, following the First World Computer Congress held in Paris the previous year An umbrella organization for societies working in information processing, IFIP’s aim is two-fold: to support information processing within its member countries and to encourage technology transfer to developing nations As its mission statement clearly states,

IFIP’s mission is to be the leading, truly international, apolitical organization which encourages and assists in the development, exploitation and application

of information technology for the benefit of all people.

IFIP is a non-profitmaking organization, run almost solely by 2500 volunteers It operates through a number of technical committees, which organize events and publications IFIP’s events range from an international congress to local seminars, but the most important are: The IFIP World Computer Congress, held every second year;

to extensive group discussion.

Publications arising from IFIP events vary The papers presented at the IFIP World Computer Congress and at open conferences are published as conference proceedings, while the results of the working conferences are often published as collections of selected and edited papers Any national society whose primary activity is in information may apply to become a full member of IFIP, although full membership is restricted to one society per country Full members are entitled to vote at the annual General Assembly, National societies preferring a less committed involvement may apply for associate or corresponding membership Associate members enjoy the same benefits as full members, but without voting rights Corresponding members are not represented in IFIP bodies Affiliated membership is open to non-national societies, and individual and honorary membership schemes are also offered.

OPTICAL NETWORKS AND

Department of Electronics and Information Systems

Osaka University, Japan

Print © 2005 by International Federation for Information Processing.

All rights reserved

No part of this eBook may be reproduced or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without written consent from the Publisher

Created in the United States of America

Boston

©200 5 Springer Science + Business Media, Inc.

Visit Springer's eBookstore at: http://ebooks.kluweronline.com

and the Springer Global Website Online at: http://www.springeronline.com

Perspectives on Optical Networks and Technologies 1

Developments in Optical Seamless Networks

Andrea Spaccapietra and Giovanni Razzetta (invited paper)

3

Cinema-class Digital Content Distribution via Optical Networks

Tetsuro Fujii, Kazuhiro Shirakawa, Mitsuru Nomura, and Takahiro Yamaguchi (invited paper)

11

Next Generation Networks – a Vision of Network Evolution

Howard Green and Pierpaolo Ghiggino

19

An Overview of Key Technologies for the Next Generation Networks

Howard Green and Pierpaolo Ghiggino

31

PART A1: Optical Packet Switching / Optical Burst Switching 45

Guaranteeing Seamless End-to-End QoS in OBS Networks

Maurizio Casoni, Maria Luisa Merani, Alessio Giorgetti, Luca Valcarenghi and Piero Castoldi

47

A Framework for the Analysis of Delay Jitter in Optical Packet SwitchedNetworks

F Callegati, W Cerroni, G Muretto, C Raffaelli and P Zaffoni

Cost Efficient Upgrading of OPS Nodes

J Cheyns, C Develder, D Colle, E Van Breusegem and P Demeester

71

A Scheduling Algorithm for Reducing Unused Timeslots by ConsideringHead Gap and Tail Gap in Time Sliced Optical Burst Switched

79Networks

Takanori Ito, Daisuke Ishii, Kohei Okazaki, Naoaki Yamanaka and Iwao Sasase

WONDER: Overview of a Packet-Switched MAN Architecture

A Bianciotto and R Gaudino

Assessing the Benefits of Wavelength Selection vs Wavelength

Routing and Wavelength Assignment for Scheduled and Random

Lightpath Demands: Bifurcated Routing versus Non-Bifurcated

An Experimental GMPLS-based Wavelength Reservation Protocol forFlooding Global Wavelength Information in Uni-ring-based MAN

Raul Muñoz, Ricardo Martinez, Jordi Sorribes and Gabriel Junyent

163

GMPLS with Interlayer Control for Session-Uninterrupted Disaster

Recovery across Distributed Data Centers

Tetsuo Imai, Soichiro Araki, Tomoyoshi Sugawara, Norihito Fujita and Yoshihiko Suemura

Intelligent OTN in the TLC Operator Infrastructures

Ovidio Michelangeli and Alberto Mittoni

195

Novel Active Monitoring of Customer Premises using Bluetooth in

Optical Access Network

S B Lee, W Shin and K Oh

203

Shared Memory Access Method for a Computing Environment

Hirohisa Nakamoto, Ken-ichi Baba and Masayuki Murata 210

PART A4: Traffic engineering 219

A Multilayer-Routing-Strategy with Dynamic Link Resource

Multi-Layer Recovery Enabled with End-to-End Signaling

D Verchere, D Leclerc, A Noury, B Ronot, M Vigoureaux, O Audouin, A Jourdan, D Papadimitriou, B Rousseau, G Luyts, S Brockmann, W Koeber and G Eilenberger

Inter-Domain Routing in Optical Networks

Américo Muchanga, Lena Wosinska, Fredrik Orava and Joanna Haralson

263

Optical Network Unit Based on a Bidirectional Reflective

Semiconductor Optical Amplifier

Josep Prat, Cristina Arellano, Victor Polo and Carlos Bock

273

Optical Label Recognition Based on Additional Pre-spread Coding

Hideaki Furukawa, Tsuyoshi Konishi, Kazuyoshi Itoh, Naoya Wada and Fumito Kubota

279

Optical Feedback Buffering Strategies

Ronelle Geldenhuys, Jesús Paúl Tomillo, Ton Koonen, and Idelfonso Tafur Monroy

Optimal Span Length Determination in Transmission Systems with

Hybrid Amplification

J D Ania-Castañón, I O Nasieva, S K Turitsyn, C Borsier and

E Pincemin

307

Separate Evaluation of Nonlinearity-Due Q Penalties in Long-Haul

Very Dense WDM Optical Systems

Livio Paradiso, Pierpaolo Boffi, Lucia Marazzi, Nicola Dalla Vecchia, Massimo Artiglia and Mario Martinelli

313

Suppression of Transient Gain Excursions in EDFA’s

Mladen Males, Antonio Cantoni, and John Tuthill

Techno-Economic Analysis of Dispersion-Tolerant Transmission

Techniques for 10Gb/s DWDM Systems

Cornelius Fürst, Helmut Griesser, Jörg-Peter Elbers and Christoph Glingener

346

2.5 Gbps 2-PSK Ultra-Dense WDM Homodyne Coherent Detection

using a Sub-Carrier based Optical Phase-Locked Loop

S Camatel, V Ferrero, R Gaudino and P Poggiolini

357

Influence of Optical Filters on the Performance of Fsk/Im Transmission

Combined (Symbol and Classical) DWDM Data Transmission

A O Nekuchaev and U Yusupaliev

384

Integrated Direct-Modulation based Quantum Cryptography System

Johann Cussey, Matthieu Bloch, Jean-Marc Merolla and Steven W Mc.Laughlin

A 40 GHz Polarization Maintaining Picosecond Modelocked Fiber

Laser Employing Photonic Crystal Fiber

Kazi S Abedin and Fumito Kubota

403

40 GHz Adiabatic Soliton Generation From a Dual Frequency Beat

Signal using Dispersion Decreasing Fiber Based Raman

Amplification

Ju Han Lee, Taichi Kogure, Young- Geun Han , Sang Hyuck Kim, Sang Bae Lee and David J Richardson

409

All-Optical Nonlinear Signal Processing at a RZ Receiver

Sonia Boscolo, Sergei K Turitsyn and Keith J Blow

416

Modification of Decoder for 2-D Wavelength/Time Optical CDMA

System by Optical Hard-Limiters

Jozef Chovan and František Uherek

422

BER Improvement using a 2-R Regenerator Based on an Asymmetric

Nonlinear Optical Loop Mirror

Markus Meissner, Klaus Sponsel, Kristian Cvecek, Andreas Benz, Stefan Weisser, Bernard Schmauss and Gerd Leuchs

438

Calculate BER Improvement due to Nonlinear Regenerators

F G Sun, Z G Lu, G Z Xiao and C P Grover

445

Bragg Gratings Photoimprinted in Integrated Optical Components:

Improving of Apodization Profiles

Lech Wosinski, Romano Setzu and Matteo Dainese

Adaptive Electronic Processing in Optical PMD-Impaired Systems

T Foggi, G Colavolpe, E Forestieri and G Prati

499

PSO Algorithm used for Searching the Optimum of Automatic PMDCompensation

Xiaoguang Zhang, Yuan Zheng, Yu Shen, Jianzhong Zhang and Bojun Yang

507

Dynamical Limitations of Single-Stage PMD Compensators

Ernesto Ciaramella

517

New Approach to Optical Polarisation Mode Dispersion Mitigation:

Experimental Analysis of the Dynamic Performances of a Cost-DrivenDevice

Raoul Fiorone, Aldo Perasso, Massimo Speciale, Marco Camera and Andrea Corti

All Optical 3R Regeneration and Wavelength Convertion

Davide Massimiliano Forin , Franco Curti, Giorgio Maria Tosi Beleffi, Francesco Matera , Andrea Reale, Silvello Betti, Simone Monterosso, Alessandro Fiorelli, Michele Guglielmucci and Sergio Cascelli

537

Experimental Study of Reshaping Retiming Gates for 3R Regeneration

M Gay, L Bramerie, G Girault, V Roncin and J.-C Simon

545

Chirp-Free Transmission through a NOLM Based Optical Regenerator

K Sponsel, M Meissner, K Cvecek, B Schmauss and G Leuchs

Two Pump OPA for OTDM Pulses Amplification

Lucia Marazzi, Paola Parolari, Pierpaolo Boffi, Elisabetta Rognoni, Paolo Gaviraghi and Mario Martinelli

573

Applications of Free Space Optics for Broadband Access

E Leitgeb, M Gebhart, U Birnbacher, S Sheikh Muhammad and

Ch Chlestil

579

Polarization Conversion Induced in A Non-Conventionally Biased

Centrosymmetric Photorefractive Crystal

Claudio Crognale and Luigi Rosa

optical domain to provide full-connectivity to support a wide variety of

services

This book stems from the technical contributions presented at theOptical Networks and Technology Conference (OpNeTec), inaugurated thisyear 2004 in Pisa, Italy, and collects innovations of optical networktechnologies toward the century network High-quality recent researchresults on optical networks and related technologies are presented,including IP over WDM integration, burst and packet switchings, controland managements, operation, metro- and access networks, and componentsand devices in the perspective of network application An effort has beenmade throughout the conference, hopefully reflected at least partially in thisbook, to bring together researchers, scientists, and engineers working bothacademia and industries to discuss the relative impact of networks ontechnologies and vice versa, with a vision of the future Too often thephotonic communication field is approached as it were a mature field wheresystems and technologies have their own lives Photonics is still in itsinfancy, playing the correlation and reciprocal influence of technology and

system/network solutions a key role, deserving more attention andconsideration on both sides.

Ken-ichi Kitayama Francesco Masetti-Placci

Giancarlo Prati

The editors wish to express their sincere thanks to the members of theInternational Program Committee of the First Optical Networks &Technologies Conference (OpNeTec), which was held in October 18-20,

2004 in Pisa, Italy, whose cooperation was essential to the organization ofthe conference and to the publication of this book

The conference would not have been possible without the support of theItalian National Consortium for Telecommunications (CNIT) and the work

of the Organizing Committee

The editors would also like to gratefully acknowledged the followingorganizations and institution hereafter:

General Chairman

Giancarlo Prati, Scuola Superiore Sant’Anna & CNIT

Technical Program Co-Chairs

K Kitayama (Co-chair), Osaka University, Japan

F Masetti-Placci (Co-chair), Alcatel CIT, France

International Program Committee

S Araki, NEC Corporation, Japan

K Asatani, Kogakuin University, Japan

K Blow, Aston University, U.K

A Bonati, Alcatel Italy

A Bononi, University of Parma, Italy

A Cantoni, W Australian Telecomm Research Inst., Australia

R Castelli, Alcatel, Italy

D Chiaroni, Alcatel CIT, France

W Chujo, Communication Research Lab., Thailand

S Dixit, Nokia Research Center, USA

A Fumagalli, University of Texas at Dallas, USA

P Franco, Pirelli Labs, Italy

P Gambini, Agilent Technologies, Italy

R Gangopadhyay, IIT Kharagpur, India

P Ghiggino, Marconi, U.K

C Glingener, Marconi, Germany

E Guarene, Telecom Italia Lab, Italy

L Jereb, Budapest Univ of Tech & Econ., Hungary

L Kazovsky, Stanford University, USA

S Bae Lee, KIST, Korea

H T Muftah, University of Ottawa, Canada

G Morthier, Ghent University, Belgium

F Neri, Politecnico di Torino, Italy

H Onaka, Fujitsu Lab, Japan

H Perros, North Carolina State University, USA

F Russo, University of Pisa, Italy

R Sabella, Ericsson, Italy

S Saracino, Siemens CNX, Italy

K Stubkjaer, Tech University of Denmark, Denmark

L Wosinska, Royal Institute of Technology, Sweden

Organizing Committee

A Bogoni, CNIT, Italy

P Castoldi, Scuola Superiore Sant’Anna & CNIT, Italy

E Ciaramella, Scuola Superiore Sant’Anna & CNIT, Italy

S Cinquini, Telecom Italia, Italy

F Di Pasquale, Scuola Superiore Sant’Anna & CNIT, Italy

K Ennser, CNIT, Italy

S Giordano, University of Pisa, Italy

L Poti’, CNIT, Italy

CNIT Secretariat

M E Razzoli, CNIT, Italy

A Letta, Scuola Superiore Sant’Anna, Italy

Publications

K Ennser, CNIT, Italy

PERSPECTIVES ON OPTICAL NETWORKS AND

TECHNOLOGIES

DEVELOPMENTS IN OPTICAL SEAMLESS NETWORKS

Invited paper

Andrea Spaccapietra1,Giovanni Razzetta2

1

VicePresident Optical Core Networks, Marconi Corporation,

New Century Park, Coventry, West Midlands, CV3 1HJ, United Kingdom,

andrea.spaccapietra@marconi.com

2

Photonics System Design Manager, Marconi Communications,

Via A Negrone 1A, 16153 Genova, Italy

giovanni.razzetta@marconi.com

Abstract: This paper give a view on key technologies that are emerging as the enabler

for evolving Core Transport Network towards the delivery of the customer experience expected by the service users community Optical technologies will be dealt with first, explaining how they are fitting in the medium and long term evolution of commercial optical transmission systems Hardware and software technologies involved in the shift toward data centric services are addressed, identifying the path to a full integrated transport and switching core network, with the ultimate objective of maximizing the user benefits and reducing cost.

National incumbents, large fixed line second operators and some key Mobileoperators have spent significant time and Capital building large, highly reliable,resilient, carrier-class Networks to support Voice and leased lines services Thishad been a stable business model for well over a decade, with the largest challengebeing how to scale

As we know the existing model is being challenged It is anticipated that therevenue our customers will generate from traditional voice services will be flat at

services (i.e triple play: voice, data and video).

Residential broadband services typically consume 10 times the bandwidth ofnarrowband users, but are offered at no more than twice the existing narrowbandsubscription rates Equally the increasing requirement for Enterprise businesses tostore, protect and retrieve information and records is doubling data traffic in theWide area every two years

Broadband services dramatically increase capacity demands on the networkbut do not return a proportional increases in revenue Therefore network operatorsneed to substantially reduce network operational costs, provide capacity at muchlower cost per bit and deliver new revenue generating services

In the following the enabling technologies for achieving such objectives areillustrated and placed in the context of new generation core transport network,delivering seamless services to network operators and their customers

1.1 Optical seamless network

Optical Seamless network is a key component in building serviceinfrastructures that deliver a delighting customer experience, innovative products,rapid time to market for new services and transform the cost base for the network

An example can be found in Figure 1

The goal of the service infrastructure is to provide a “simple and complete”communications service to customers, regardless of time or place

The pillars of this network vision are:

an high performance, integrated, cost effective transport infrastructure,evolving from current transport networks;

new platforms for services based on the delivery of content and applications,supporting both multi-media and mobile services;

OSS increasingly becoming part of the service, and ultimately convergingwith the network intelligence components;

standards to define the architecture components

The optical seamless network is a rationalized optical transport enabling anultra broadband data network Access networks will be converging onto a multi-service platform

Figure 1- Service network architecture

SEAMLESS NETWORK

After positioning the Optical Seamless network in the bigger picture of overallservice network, we need to identify the requirements that service objectives set tothe transport network and how they can be best addressed using emergingtechnologies Requirements can be better understood if the transport network islogically divided into a data plane and a control plane From the technologicalpoint of view the data plane can be further divided into three technology layers: theoptical plane, the electrical plane and the data plane in strict sense (i.e packet/cellbased)

Key requirements and related technologies for the optical plane enabling theseamless operation can be identified in transparent bypass of wavelengths (OXC,ROADM, EDFA and Raman amplifiers) and compatibility with existing fibre plant(dispersion resistant modulation schemes for combined 40Gbit/s/10Gbit/stransport)

Emerging technologies for the electrical plane ready for a massive roll-out inthe electrical plane are integration of SDH and wavelength (ODU) switching,standards compliant mapping of data into transport frames (GFP/LCAS/VCAT).The data plane will be initially provided by separated plug-in cards andultimately a protocol agnostic transport switching can allow further integrationwith the electrical plane

Figure 1 – Network planes

The control of the network will progressively migrate from centralizedoperation support systems or network management system to a distributed controlplane (the same way large data networks are managed today), although functionslike Alarm Supervision, Performance Monitoring and Inventory will remain underthe TMN domain

A common control scheme for data and transport simplifies operationalprocesses and ease the convergence of data and transport

2.1 Optical plane

After the massive investment in the “Holy Grail” of optical technologies duringthe years of the “Telecom bubble”, we are now in a situation where componentssuppliers are bringing selected optical technologies to the marketplace for beingusefully utilized in the design of optical system The economical benefit of the socalled “optical bypass ” is now recognized in the transmission community and anumber of component and system technologies are available Some of them arewell established, such as dynamic power management, dispersion management,error correction, some of them are being developed and will become common inthe near future, such as electrical distortion compensation (EDC), innovativemodulation formats and sophisticated methods for optical performance monitoring

& fault localization All of these ingredients are preparing current transmission

infrastructure for scaling towards high channel bit-rates (i.e 40G), that has alreadybeen experimented in the field, both in Europe and US.

All the above needs to be achieved through a “platform concept”, in order todeliver inexpensive unlimited bandwidth with the maximum flexibility Forexample a single common chassis type that has multi function slots that canaccommodate the relevant transponders, amplifiers or optical switching unit

7

2.2 Electrical plane

If we look at today SDH layer of a Core transmission network we can see that

we have a potential like for like element replacement of 16:1 or greater The highcapacity OCS nodes can consolidate multiple 1st generation ADM Elements intoone element This results in considerable infrastructure savings with the associatedbenefits this brings This allows significant improvements to be made at all levels

of a core implementation It also allows simplification of potential co-locatedelements which are interconnecting to other Access rings bringing operational andreal estate benefits

Moving to the layer above SDH we can see savings gained through integrationbetween layers of the core Network The recently standardized Optical TransportNetwork frame structure (OTN/ODU) is now integrated in optical cross-connects,that now can handle wavelength bandwidth granularity

Along with the evolution toward larger granularity and massive integration, thecapability of handling data streams by simply mapping them in transport structure

is now being deployed Bandwidth optimisation through aggregation and dynamicbandwidth allocation enhance the competitiveness of the solutions with respect topure data networks This identifies a definite trend for transport networks toencompass layer 2 and even layer 3 functionalities, but with the carrier classavailability performance that transport network only can deliver

This provides a base architecture that can adapt to accommodate any servicemix

2.3 Control Plane

The “Seamless’ attribute of new generation transport network is associatedwith an increased level of intelligence that is inherent within the network elementsand is realised via the implementation of an ASTN/GMPLS control plane, whichsimplifies network operation and optimises network resources

Via the neighbour discovery function newly added network elements and nodesare automatically recognised by the network The additional capacity and routediversity becomes automatically part of the network resource pool and canimmediately be used Long and cumbersome manual configuration processes

standardised user-network-interface (UNI) The network automatically finds thebest and least expensive route through the mesh and protects it according to theselected grade of service New services are provisioned at a fraction of the timeand cost compared to today’s manual processes Additional flexibility can beoffered to the end-user, optical virtual private networks (OVPN) become acompelling reality.

Mesh based restoration schemes increase the resilience of the network againstany type of failure Therefore new survivability schemes can be offered in addition

to the well-known SDH/OTN protection mechanisms and guaranteeing comparableswitching time

2.4 The convergence of data and transport

The reality of an optical layer capable of dynamic provisioning and restoration

of optical circuits offers the opportunity for an architecture where a reconfigurableSDH/OTN network delivers connectivity to the nodes of a packet backbone Thereconfigurable optical layer can be shared among other service networks such asATM, Frame Relay or leased lines

User to Network Interface between IP and transport systems is standardisedand interoperability between the two has been demonstrated and tested IP overtransport network approach is key to guarantee the scalability of the switchedoptical backbones and the cost-effectiveness of this approach has beendemonstrated by network modelling studies

The final convergence step will be represented by a unique Core Node wherethe Core Data functionality (IP/MPLS) is fully integrated inside the TransportNode As mentioned before the availability of protocol agnostic switching fabricswill surely enable this evolution

The flexibility of a single platform, Multi-application, next generation solutionalso simplifies the core network, significantly reducing the operational expendituredue to reduced Maintenance, powers, space, spares and training Modular &Scalable platforms enables the network to scale with traffic requirements whilstalso ensuring low first in cost

Through implementation of intelligent switching and software mechanisms, theimplementation of products with “change aware” hardware enables Dynamic TMNcapabilities to the network allowing on the fly provisioning of services andrestoration This dynamic network environment can be utilised for creatingcompetitive advantage by introducing new enhanced differentiated dynamicservices

Open interfaces conformant to the standards developed by the telecom industryensure inter-operability between different layer networks and between differentvendors equipment in the same layer

By the evolution path outline, significant and measurable benefits of the opticalseamless network are realised

Jan Späth: “Impact of traffic behaviours on the performance of dynamic WDM transport networks.” ECOC 2002

L Blair et al.: “Impact of switch node architecture upon capacity efficiency in Williams North American network.” ECOC 2002.

Sudipta Sengupta, Vijay Kumar, Debanjan Saha: “Switched Optical Backbone for Cost-Effective Scalable Core IP Networks.” IEEE Communications Magazine, June 2003

Agostino Damele, Andrea Spaccapietra: “Telecom Network Architecture: Multi-Layer Switching Solution.” FITCE 2002

Joerg-Peter Elbers: “High-capacity DWDM/ETDM transmission.” OFC 2002 L.M Gleeson, M.F Stephens, P Harper, A.R Pratt, W Forysiak, D.S Govan, B.K Nayar, I.D Phillips, B Charbonnier, M.D Baggott, H.S Sidhu,

managed all-optical cross connects and add-drop multiplexing.” ECOC2003 Stefan Herbst, Heinrich Lücken, Cornelius Fürst, Silvia Merialdo, Jörg-Peter Elbers, Christoph Glingener: “Routing criterion for XPM limited transmission

in transparent optical networks.” ECOE 2003.

[10]Silvia Merialdo, Jörg-Peter Elbers, Cornelius Fürst, Stefan Herbst, Christoph Glingener: “Path tolerant dispersion management for transport optical networks.” ECOC 2003

[11]Helmut Griesser, Joerg–Peter Elbers, Christoph Glingener: “A generalised concatenated error correcting code for optical fibre transmission.” ECOC 2003 [12]Cornelius Fürst, Roman Hartung, Jörg-Peter Elbers, Christoph Glingener:

“Impact of spectral hole burning and Raman effect in transparent optical networks.” ECOC 2003

[13]G L Jones, W Forysiak, J H B Nijhof : “Economic benefits of all-optical cross connects and multi-haul DWDM systems for European national networks.” OFC 2004

[14]A R Pratt, P Harper, S B Alleston, P Bontemps, B Charbonnier, W Forysiak, L Gleeson, D S Govan, G L Jones,

[15]D Nesset, J H B Nijhof, I D Phillips, M F C Stephens, A P Walsh, T Widdowson and N J Doran: “5,745 km DWDM transcontinental field trial using 10 Gbit/s dispersion managed solitons and dynamic gain equalization.” OFC 2003

[16]A R Pratt, B Charbonnier, P Harper, D Nesset, B K Nayar and N J Doran: “40 x 10.7 Gbit/s DWDM transmission over a meshed ULH network with dynamically re-configurable optical cross connects.” OFC 2003

CINEMA-CLASS DIGITAL CONTENT

DISTRIBUTION VIA OPTICAL NETWORKS

Invited paper

Tetsuro Fujii, Kazuhiro Shirakawa, Mitsuru Nomura, and Takahiro Yamaguchi

NTT Network Innovation Laboratories

1-1 Hikarinooka Yokosuka-shi,Knanagawa 239-0847 Japan

fujii.tetsuro@lab.ntt.co.jp

Abstract: To transmit and display high quality movies via optical networks, a

new Super High Definition (SHD) digital cinema distribution system with the resolution of 8-million pixel is developed Its image quality

is four times of HDTV in resolution, and enables us to replace conventional 35mm films This system is based on JPEG 2000 coding technology and transmits high quality digital cinema over high-speed IP networks All digital cinema data are continuously transmitted at up to 500 Mbps This system opens the door to the next generation of cinema-class digital content distribution over optical networks.

The growth of broadband networks has stimulated the development ofapplications that use high quality image communications To satisfy professionalusers in industry, i.e printing, medicine, and image archiving, a precision colorimaging system is required to achieve the digital images of excellent qualitybeyond HDTV An image category, called Super High Definition (SHD) images[1,2] is defined to have a resolution of at least 2000 pixels vertically with 24-bitcolor separation The SHD images surpass the quality of 35-mm films in terms ofspatial resolution In our first study on the SHD images and their applications, wedeveloped a high quality still image system with 28.3 inches LCD display of2560x2048 pixel resolution This image system features GbE as high speed

At the same time, we have developed a new platform for high quality digitalcinema with 8 million pixels, called SHD digital cinema This SHD digital cinemascales the heights now occupied by 35mm film It offers large venue support, largescreen projection to fully realize the promise of digital cinema, scalable integratedmedia production, and film-less cinema distribution via broadband networks Thenew SHD digital cinema format is defined as 2000 or more scan lines,progressively scanned, running at 24 frames per second To evaluate this newformat, we have developed real-time DECODER and a projector capable ofhandling SHD digital cinema with an effective resolution of 3840 x 2048 pixels(square sample) SHD digital cinema contains roughly four times the pictureinformation of HDTV 1080p/24 A comparison is made in Figure 1 SHD digitalcinema features RGB color encoding and 30 bits per pixel, for a much more film-like visual richness Motion picture people in Hollywood count up the cinemaresolution from the point of horizontal pixel Therefore, they call our system as

“4K Digital Cinema”

To transmit the movie contents using optical networks, an exceptionally highperformance decoder and an imaging system are required to process the movies inreal-time This is because the total bit rate of an SHD digital cinema can equal 5.6Gbps (3840 x 2048 pixels, 24 fps and 30-bit color), and the movie should becompressed to 10:1 - 20:1 in order to transmit them by wide-area IP networks.Eventually a special combination of a real-time DECODER and a projectiondevice is required to show SHD digital cinema In this paper, we introduce an SHDdigital cinema distribution system

Figure 1 Resolution and frame rate of SHD images

We have developed a prototype digital cinema system that can store, transmitand display SHD digital cinema of 8 million (3840x2048) pixel resolution usingJPEG2000[3] coding algorithm The SHD digital cinema distribution system isshown in figure 2 and 3 This is the third generation of our SHD digital cinemadistribution system[4,5,6] The transmission from the server to the real-timeDECODER is done over GbE (Gigabit Ethernet) It consists of three main devices,

a video server, a real-time DECODER, and a LCD projector We assume that themovie data have been compressed and stored in advance The real-timeDECODER decompresses the video streams transmitted from the server usingparallel JPEG2000 processors, and outputs the digital video data to an LCDprojector with 3840x2048 pixel resolution and RGB 30 bits with 24 fps

Figure 2 System conciguration of prototype SHD digital cinema distribution system

Figure 3 Photograph of SHD digital cinema distribution system

The SHD digital cinema system uses the JPEG2000 algorithm to construct areal-time DECODER From the point of communication traffic and storage cost,inter-frame coding such as MPEG-2 is better at compressing movie data However,

we selected JPEG2000 for the following five reasons (1) There is no internationalstandard to compress video for RGB 30 bits or more (2) Other sets of compresseddata with lower resolution can be generated easily by using the facility forembedded scheme of layered coding algorithm (3) Intra-frame coding schemesremain important because of their support of video editing (4) It is much easier toimplement a parallel processing decoder by using compact JPEG2000 CODECchips (5) The decoder is robust against data error Error recovery is achievedsimply by discarding the corrupted image frame The reasons (3) to (5) arecommon for JPEG and JPEG2000, but only JPEG2000 satisfies reasons (1) and (2)for excellent image reproduction which is requested by motion picture people

2.2 Real-time DECODER

The DECODER can perform the real-time decompression at a speed of 500Mpixels per second, using parallel JPEG2000 processing elements As a JPEG2000processor, Analog Devices Inc ADV202 is selected The decoder consists of 2circuit blocks, a PC/LINUX part with GbE interface, and newly developedJPEG2000 decoder boards shown in figure 4 4 chips of ADV202 are installed ineach board Total 4 boards are installed on the PCI-X-bus in order to process 24frames of 4K x 2K pixel up to 36-bit RGB color images (4 :4 :4) in a second Atthe same time, This board supports 10 bits YCbCr (4 :2 :2) mode and transferfunction from YCbCr to RGB is installed This kind of flexibility is obtained bythe FPGA based circuit design

Figure 4 Photograph of JPEG2000 decoder board

The PC part receives the coded streams of 200M to 500Mbps, then transfersthem to JPEG2000 decoder boards The PC part consists of dual CPUs runningwith LINUX A control program runs as an application that consists of two threadsthat share the PC’s main memory as a large data buffer One thread reads the datareceived from GbE-NIC, and the other reformats and forwards them to each of fourdecoder boards The buffer size is only limited by the main memory size of the PC,usually set to 200MB

2.3 LCD Projector

The prototype projector itself was developed by JVC using D-ILA technology.The high picture quality of D-ILA is derived from the advanced LCOS (LiquidCrystal on Silicon) technology and a high-precision optical system The majorfeatures of D-ILA are high brightness, high resolution, high contrast ratio, analoggradation and high-speed response The SHD LCD projector uses three pieces of3840x2048 pixel reflective D-ILA panels for each RGB 10-bit colors, and its size

is 1.7 inches The effective brightness exceeds 5000 ANSI lumens by using a1600W xenon lamp, which is bright enough to show images on as large as 300-inch diagonal screens The refresh rate of the projector is chosen to 96Hz Thishigh refresh rate thoroughly eliminate flicker, and is compatible to 24fps movie.Every frame of decoder output is simply displayed for times in the projectorwithout any interpolation between adjacent frames

2.4 PC Video Server

The PC server consists of LINUX running on dual CPUs (Pentium III, 1.44GHz),

an IDE-RAID (200GB x 6, RAID0, striping mode), and a GbE NIC The moviefilms are digitized to a large set of still images, and compressed in advance Theoriginal movie data are divided into 960x512 pixel image tiles, and they have 30bits RGB color components A data transfer command reads the data from theRAID and writes them periodically to the GbE NIC Like the DECODER controlprogram, a large size shared buffer is used in the server in order to enhance themaximum transmission rate by averaging the disk read speed Contrary to anordinal streaming system that runs by its own clock, the data rate of the server isprecisely controlled by the DECODER via acknowledge signal This is becausethe decoder generates the master-clock of the movie system of 24Hz to yieldsmooth replay without the lack or duplicate of frame

The quality of SHD movies and performance of the transmission system areevaluated by using long movie data that are digitized and compressed from avariety of actual movies.

3.1 Digital Cinema Data Acquitision

To evaluate the SHD movie system, a lot of movie sequences are required tocompare with the conventional film movies, So we provided various types of highquality movie data sequences with less blurs, scratches, nor grain noises Some ofthe sequences are digitized from original negative films, such as “Circle of Love”provided by ARRI and short test sequences provided by Hollywood studios.Others are from inter-positive (IP) films, such as “Tomorrow’s Memory” shot byNTT, and dupe-negative (DN) films, such as “Tomb Raider” of full Hollywoodmovie Usually, we use an IMAGICA’s “Imager XE” film scanner to digitize thesefilms This scanner can yield image files of 4096x3012 pixel as maximum sizewith RGB 10 bit log Cineon format and it takes 10 seconds per frame for scanning.This is a very time consuming process of data acquisition Recently, ThmosonGrass Valley announced the development of high speed film digitizer This kind ofhigh speed film digitizer is indispensable to make large variety of digitized cinemacontents to form a cinema archiving with reasonable costs in the near future.Master image file is obtained from this 10 bit log Cineon format image filesthrough the color correction process As master image file format, we use RGB30bit TIFF file for each frame In the case of “Tomb Raider” of 101 minutesmovie, the size of master file is 4.5 Tera Bytes from 144,000 frames

As far as sound material, we don’t apply any compression technologies and weuse uncompressed sequence Usually, master sound can be obtained as theTASCAM DA98 tapes The 6-channel sound tracks are extracted from this tape.The digital sound data are stored in the Linux/PC server as WAV files andtransmitted as IP stream with image data As the highest sound quality, 24 bit perchannel with 96 KHz sampling is available for the play back of orchestra andmusical

3.2 Network Transmission

We selected TCP for the connection protocol from server to real-timeDECODER TCP is adequate for the stable connected transmission and bestmethod to share the bandwidth of IP router based network Within the multiplexcinema, it is very easy to use But for large RTT (Round trip time) network, it isvery difficult to extract its full performance In order to verify the performance, we

had an experiment with a long distance high-speed IP network environment calledInternet 2, on 29th October, 2002 In this case, we used the second generation ofour SHD digital cinema distribution system[5,6]

The network configuration is shown in figure 5 We set up the server in theElectronic Visualization Laboratory (EVL) at the University of Illinois, Chicago(UIC) The real-time DECODER and the projector were installed at the RobertZemeckis Center of the School of Cinema-Television at the University of SouthernCalifornia (USC), Los Angeles The distance between the server and the decoderwas more than 3000 km There were six router hops between them The RTT(Round Trip Time) of the network was measured to be 59ms The targettransmission rate is 300 Mbps Many power users were sharing Internet2 while ourexperiments

Figure 5 Network configuration of long distance transmission with Internet2

To overcome the long distance, we applied (1) a large TCP window, 4 MByte,(2) multiple TCP connections, and (3) a shaping control function The TCPwindow size is the amount of data able to be sent without acknowledgement Thereare theoretical limitations to TCP-window-based flow control The configurationguideline is “window size > Required TCP throughput x RTT ” The TCP windowsize of the system was extended to 4 MByte from its initial 64 KByte value.Measured throughput was increased from 8 Mbps to 50 Mbps Theoretically,extension of window size is enough, but it is impossible to extract real high speed

To improve the performance, we increase the number of TCP connectionsbetween the server and the real-time CODEC The server application divided themovie data into equal segments and sequentially wrote them to multiple TCPsockets As the number of TCP connections increased, it was confirmed that

But the stream’s bit rate could not be raised even when the number of connectionswas increased more than 64.

We used an application traffic monitor to observe the traffic pattern to aresolution of 1 ms We found that the nature of data transmission was very bursty

To suppress the burstiness, a shaping control function for the data transmission wasbuilt into the socket writing process of the server application As a result,transmitted movie data traffic reach 300 Mbps We could succeed in transmittingSHD movie over 3000 Km at 300 Mbps with TCP/IP protocol To implementadequate functions, TCP can be applied to long distance transmission

We have developed an SHD digital cinema system that offers the resolution of3840x2048 pixels and the quality of RGB 30 bits We have exerted concertedefforts in realizing a complete digital cinema system that fully match the quality of35mm film Now the technology of optical network has matured and the new era

of broadband network is coming We believe that a lot of image service willappear using this kind of high quality digital content distribution platform

NEXT GENERATION NETWORKS –

A VISION OF NETWORK EVOLUTION

Howard Green 1 , Pierpaolo Ghiggino 1

1

Marconi Communications, Stoneleigh House, New Century Park, Coventry, CV3 1HJ, UK

Abstract: This article presents a view of the needs and developments for the “Next

Generation Networks” It starts from a market and service context following the burst of the Internet bubble and sketches the likely evolution of services

by end user type It is centered, however on a vision of network evolving to architectures necessary to support the needs of operators with special emphasis on the European environment.