leos - aristotle university of thessalonikiusers.auth.gr/npleros/publications/journal...

56
OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage you to attend the first LEOS Annual Meeting being held outside North America, November 10-14, in Glasgow. Special fea- tures of this year’s meeting include a 40 Gbit/s Symposium and five Short Courses on Sunday. Last year, we had the first Chapter Retreat to which we invited all Chapter Chairs to attend and share their experiences, thoughts, and best practices. This year, we continue with the Chapter Retreat and add a Technical Committee Retreat, to provide a similar forum where Committee Chairs can share their thoughts. The reception will be held in the new Glasgow Science Center. Knowing how fun-loving our Scottish friends are, you can look forward to a most entertaining evening. Haggis procession complete with bagpipe music, no doubt. Based on your feedback, I am gratified that some of you actually read my ram- bling. So, I will continue my stream of con- sciousness: Leadership is a fascinating subject. So much has been written about it with so many different interpretations. I wonder how one can instantaneously recall and fol- low the long lists of dos and don’ts. And the gurus keep changing on us. Charisma has long been listed as an important leader- ship characteristic, for example, and now people are saying that too much emphasis on charismatic leaders is the cause of our economic woes. The amazing thing is you see so many people in leadership positions doing the obviously wrong things that you wonder if they really understand what lead- ership is all about. Over the years, I have tried to internalize what I read, and distill the material to a few principles to guide my actions in business. I hope you will benefit from what I’ve learned. Almost everything that has been written about leadership more or less points to the word “protection.” Based on the principle of “give-to-get” in relationships, a leader must intuitively understand and conscious- ly take care of the needs of employees in exchange for their productivity or even loy- alty. Otherwise, what’s in it for the employ- ee to align his/her energy to the goals of the company? Employees are willing to follow leaders who they believe can provide them safety. For example, a lot of leaders in the limelight are charismatic so that’s an easy generalization for employees to have confi- dence that the company under a charismatic leader can succeed. The same can be said about having a vision, being strategic, or having the ability to communicate. But many great leaders are not charismatic at all and others do not have good verbal skills. The point is each of us has our own ways to convey a sense of safety and security to the people around us, but true “protection” we must provide at the end. Unfortunately, some people in leadership positions are good at faking it rather than providing what people really need. “Protection” sounds far out when com- panies are laying off employees, and if this concept is blindly applied can be detrimen- tal. Protection must be viewed in the con- text of what is good for the company and the individuals involved long-term. I emphasize the importance of “long-term” good because one of the easy pitfalls is to be soft on people. It is all too easy to come up IEEE A publication of the IEEE Lasers and Electro-Optics Society President’s Column MILTON CHANG VOLUME 16 NUMBER 4 OCTOBER 2002 ISSN: 1060-3301 www.i-LEOS.org In this Issue . . . President’s Column 1 Editor’s Comments 2 LEOS 2002 Annual Meeting, San Diego, CA 3 LEOS Profile Kristian Elmholdt Stubkjaer 6 IEEE LEOS Summer Topicals Meetings 7 Special Issue All-Optical Clock Recover Using Self-Pulsing Two-Section Gain-Coupled DFB Lasers 9 Monolithically Integrated Wavelength Converter 11 All-Optical Processing in Switching Networks 13 All-Optical Transmission and Wave-length Conversion of 40 Gb/s Signals Over One Million Kilometers of Fiber 15 All-Optical Clock Recovery for Signal Processing and Regeneration 17 System Perspective for All-Optical Switching 19 Fast Processes in Semiconductor Optical Amplifiers 21 Hierarchical Hybrid OTDM/WDM Network Based on Fast Optical Signal Processing 24 Optical Header Recognition Using Fiber Bragg Grating Correlators 29 OTDM Packet Networking Devices at 100 Gbit/s and Beyond 33 Applications of Fiber-Based Optical Parametric Amplifiers 36 Optical Packet Switching and Associated Optical Signal Processing 39 Ultrahigh-Speed Optical Signal Processing with Symmetric-Mach-Zehnder-Type All- Optical Switches 43 New Senior Members 45 Conference Calendar 46 continued on page 50 LEOS N E W S L E T T E R

Upload: phunglien

Post on 31-Jul-2019

216 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 1

Greetings! I encourage you to attendthe first LEOS Annual Meetingbeing held outside North America,

November 10-14, in Glasgow. Special fea-tures of this year’s meeting include a 40Gbit/s Symposium and five Short Courseson Sunday. Last year, we had the firstChapter Retreat to which we invited allChapter Chairs to attend and share theirexperiences, thoughts, and best practices.This year, we continue with the ChapterRetreat and add a Technical CommitteeRetreat, to provide a similar forum whereCommittee Chairs can share their thoughts.

The reception will be held in the newGlasgow Science Center. Knowing howfun-loving our Scottish friends are, you canlook forward to a most entertainingevening. Haggis procession complete withbagpipe music, no doubt.

Based on your feedback, I am gratifiedthat some of you actually read my ram-bling. So, I will continue my stream of con-sciousness:

Leadership is a fascinating subject. Somuch has been written about it with somany different interpretations. I wonderhow one can instantaneously recall and fol-low the long lists of dos and don’ts. Andthe gurus keep changing on us. Charismahas long been listed as an important leader-ship characteristic, for example, and nowpeople are saying that too much emphasison charismatic leaders is the cause of oureconomic woes. The amazing thing is yousee so many people in leadership positionsdoing the obviously wrong things that youwonder if they really understand what lead-ership is all about. Over the years, I havetried to internalize what I read, and distill

the material to a few principles to guide myactions in business. I hope you will benefitfrom what I’ve learned.

Almost everything that has been writtenabout leadership more or less points to theword “protection.” Based on the principleof “give-to-get” in relationships, a leadermust intuitively understand and conscious-ly take care of the needs of employees inexchange for their productivity or even loy-alty. Otherwise, what’s in it for the employ-ee to align his/her energy to the goals of thecompany? Employees are willing to followleaders who they believe can provide themsafety. For example, a lot of leaders in thelimelight are charismatic so that’s an easygeneralization for employees to have confi-dence that the company under a charismaticleader can succeed. The same can be saidabout having a vision, being strategic, orhaving the ability to communicate. Butmany great leaders are not charismatic at alland others do not have good verbal skills.The point is each of us has our own ways toconvey a sense of safety and security to thepeople around us, but true “protection” wemust provide at the end. Unfortunately,some people in leadership positions aregood at faking it rather than providingwhat people really need.

“Protection” sounds far out when com-panies are laying off employees, and if thisconcept is blindly applied can be detrimen-tal. Protection must be viewed in the con-text of what is good for the company andthe individuals involved long-term. Iemphasize the importance of “long-term”good because one of the easy pitfalls is to besoft on people. It is all too easy to come up

IEEE

A publication of the IEEE Lasers and Electro-Optics Society

President’s ColumnMILTON CHANG

VOLUME 16NUMBER 4OCTOBER 2002 ISSN: 1060-3301

www.i-LEOS.org

In this Issue . . .

President’s Column 1

Editor’s Comments 2

LEOS 2002 Annual Meeting, San Diego, CA 3

LEOS Profile

Kristian Elmholdt Stubkjaer 6

IEEE LEOS Summer Topicals Meetings 7

Special Issue

All-Optical Clock Recover Using Self-PulsingTwo-Section Gain-Coupled DFB Lasers 9

Monolithically Integrated WavelengthConverter 11

All-Optical Processing in Switching Networks 13

All-Optical Transmission and Wave-lengthConversion of 40 Gb/s Signals Over OneMillion Kilometers of Fiber 15

All-Optical Clock Recovery for SignalProcessing and Regeneration 17

System Perspective for All-Optical Switching 19

Fast Processes in Semiconductor OpticalAmplifiers 21

Hierarchical Hybrid OTDM/WDM NetworkBased on Fast Optical Signal Processing 24

Optical Header Recognition Using FiberBragg Grating Correlators 29

OTDM Packet Networking Devices at 100Gbit/s and Beyond 33

Applications of Fiber-Based OpticalParametric Amplifiers 36

Optical Packet Switching and AssociatedOptical Signal Processing 39

Ultrahigh-Speed Optical Signal Processingwith Symmetric-Mach-Zehnder-Type All-Optical Switches 43

New Senior Members 45

Conference Calendar 46

continued on page 50

LEOSN E W S L E T T E R

Page 2: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

IEEE Lasers and Electro-Optics Society

2 IEEE LEOS NEWSLETTER OCTOBER 2002

PresidentMilton Chang Incubic 855 Maude Avenue Mountain View, CA 94043Tel: +1 406 218 8658Fax: +1 650 960 0933Email: [email protected]

President-ElectG-D KhoeCOBRA InstituteEindhoven University of TechnologyFaculty E. Building EH-12; P.O. Box 5135600 MB Eindhoven, NetherlandsTel: +31 40 247 3880Fax: +31 40 245 [email protected]

Secretary-TreasurerAndrew M. Weiner Purdue University School of Elec. And Computer

Engineering1285 Electrical Engineering BuildingWest Lafayette, IN 47907-1285Tel: +1 765 494 5574Fax: +1 765 494 6951Email: [email protected]

Junior Past President Philip J. AnthonyJDS Uniphase 1865 Lundy Avenue San Jose, CA 95131Tel: +1 408 546 4562Fax: + 408 546 4547Email: [email protected]

Senior Past President Gordon DayOptoelectronics Division, 815.00National Institute of Standards &

Technology325 BroadwayBoulder, CO 80303-3328 Tel: +1 303 497 5204Fax: +1 303 497 7671Email: [email protected]

Executive DirectorPaul ShumateIEEE/LEOS445 Hoes LanePiscataway, NJ 08855-1331Tel: +1 732 562 3891Fax: +1 732 562 8434Email: [email protected]

Board of GovernorsJ.J.Coleman B.Y. KimP.J. Delfyett J.H. MarshM.M. Fejer S.A. NewtonC. Harder K. OkamotoP. Iannone J.B. SooleU. Keller K. Stubkjaer

Vice PresidentsConferences - H. S. Hinton Finance & Administration - G.D. Khoe Membership & Regional Activities

Americas - K. BergmanAsia & Pacific - S. Sudo Europe, Mid-East, Africa - S. Donati

Publications - R. J. Lang Technical Affairs - I.C. Khoo

Newsletter Staff

Executive EditorMary Y.L. WisniewskiIBM T J Watson Research Center Route 134 Yorktown Heights, NY 10598Tel: +1 914 945 2125Fax: +1 914 945 1358Email: [email protected]

Associate Editor Michael HaydukAir Force Research Lab - SNDR25 Electronic Parkway Rome, NY 13441-4515Tel: +1 315 330 7753Fax: + 315 330 7901Email: [email protected]

Associate Editor David V. PlantMcGill UniversityMcConnell Eng. Bldg., Room 7563480 University StreetMontreal, Quebec H3A-2A7, CanadaTel: +1 514 398-2989Fax: +1 514 398-3127Email: [email protected]

Contributing EditorDiana HuffakerUniversity of New Mexico, CHTM1313 Goddard SEAlbuquerque, NM 87106Tel: +1 505 272 7845Fax: +1 505 272 7801Email: [email protected]

Staff EditorGail WaltersIEEE/LEOS 445 Hoes LanePiscataway, NJ 08855Tel: +1 732 562 3892Fax: +1 732 562 8434Email: [email protected]

This October issue represents the fifthannual Special Issue of the IEEE-LEOSNewsletter. Our goal in publishing an

annual Special Issue is to provide LEOSmembers with an overview of research in aselected topic.

This year’s Special Issue features articleson the subject of “Fast Optical SignalProcessing in Optical Transmission” byresearchers who presented invited talks at the2002 LEOS Summer Topical on July 15-17.2002, at Mont Tremblant, Quebec, Canada.The Chair of this conference is Prof. KristianStubkjaer, who is profiled in this SpecialIssue. Prof. Stubkjaer is an elected member ofthe LEOS Board of Governors and is present-ly at the Technical University of Denmark.For more information on the 2002 LEOSSummer Topicals, see the website:http://www.ieee.org/organizations/society/leos/LEOSCONF/SUM2002/sum02.htm. Wehope that you enjoy reading the invited arti-cles that are contained in this issue.

The last meeting in 2002 of the LEOSBoard of Governors will be held on Nov. 11-14 in Glasgow, Scotland. For more details, seethe LEOS Conference web page at the URL:http://www.i-leos.org/info/calendar2002.html.

Please feel free to suggest comments andsuggestions to improve the Newsletter and tosubmit articles: [email protected]

Editor’sCommentsMARY WISNIEWSKI

IEEE LEOS Newsletter (ISSN 1060-3301) is pub-lished bimonthly by the Lasers and Electro-OpticsSociety of the Institute of Electrical andElectronics Engineers, Inc., Corporate Office: 3Park Avenue, 17th Floor, New York, NY 10017-2394. Printed in the USA. One dollar per memberper year is included in the Society fee for eachmember of the Lasers and Electro-OpticsSociety. Periodicals postage paid at New York,NY and at additional mailing offices. Ride-alongenclosed. Postmaster: Send address changes toLEOS Newsletter, IEEE, 445 Hoes Lane, P.O. Box1331, Piscataway, NJ 08855-1331.

Copyright © 2002 by IEEE: Permission to copywithout fee all or part of any material without acopyright notice is granted provided that thecopies are not made or distributed for directcommercial advantage, and the title of thepublication and its date appear on each copy.To copy material with a copyright notice requiresspecific permission. Please direct all inquiries orrequests to IEEE Copyrights Office.

Page 3: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 3

The fifteenth Annual Meeting of theIEEE Lasers and Electro-OpticsSociety (LEOS) will be held from

November 10-14, in Glasgow, Scotland’slargest city. This meeting will be the firstannual meeting held outside the UnitedStates of America, to reflect the Society’sgrowing overseas membership and activi-ties. To further promote global participa-tion, LEOS is now offering registration feesreduced by up to 75% based on WorldBank country classifications.

Submissions to the conference continue tobe strong, and the technical sessions promiseto be as vigorous as ever. Four plenary talks tokick-off the meeting are “New Age FibreCrystals” by Philip St. J. Russel of theUniversity of Bath, “Advances in Vertical-Cavity and Widely-Tunable Lasers using InP-Based PIC Technology” by Larry A. Coldrenof the University of California, Santa Barbara,“The Future of Optical Communications” byShigeyuki Akiba of KDDI Submarine CableSystems, and “Optical Sciences in Scotland”by John Marsh of the University of Glasgow.A record number of the special symposia forthis year’s meeting cover a variety of “hot”topic areas: 40 Gb/s Systems – the NextFrontier in Optical Communications, Lasersin Medicine & Biology, PhotonicsTime/Frequency Metrology, Nonlinear Opticsfor Coherent Sources from THz Radiation toX-Rays, Agile Optical Beams andApplications, and Photonic Integration.Moreover, the above six special symposia andthe sixteen regular technical sessions will fea-ture a large selection of invited talks, the hall-mark of the LEOS Annual Meeting, to covera wide range of disciplines under LEOS.

For new entrants to the field as well asthose wishing to expand their technicalscope, we offer on Sunday six short coursestaught by recognized technical experts.They are “Vertical Cavity Surface EmittingLasers: Technologies and Applications” byFumio Koyama of the Tokyo Institute ofTechnology, “40 GB/sec Technologies” byLeda Lunardi of JDS Uniphase and AlanWillner, “Design of AWGs forMultiplexing, Demultiplexing, Power-monitoring, Dispersion Compensation,

Gain-flattening and other Applications” byHenk Bulthuis and Martin Amersfoort ofKymata , “Wavelength-tunable LaserDiodes for Optical Communications” byMarkus C. Amman of the TechnischeUniversitaet Muenchen, “Optical Beam-forming for Phased-Array Antennas” byHenry Zmuda of the University of Florida,“Design of EDFAs, Raman Modules andHybrid Raman/EDFA Amplifiers” byDoug Butler of Corning.

The purpose of this annual meeting isto give LEOS members an opportunity togather together to present and discuss theirwork. The informal atmosphere encouragesintroduction and interaction. All confer-ence attendees are invited to attend theMonday evening conference reception andawards ceremony for an evening of foodand fun and a chance to connect withfriends and colleagues. We hope you willtake the time to join us in Glasgow,Scotland, this year, to enjoy one of theliveliest and most cosmopolitan destina-tions in Europe and, of course, to attendthe conference. You can find more informa-tion regarding the Annual Meeting at theLEOS website, http://www.i-leos-org/. Seeyou in Glasgow, Scotland.

Topic areas are as follows:

Displays“Organic Light Emitting Diode (OLED)Displays”. The workshop is going to be pri-marily focused on organic electrolumines-cent devices and organic light emittingdiode (OLED) displays. OLED based flatpanel displays are rapidly reaching commer-cialization. OLEDs are being activelyexplored for applications in solid state light-ing and imaging applications. In addition todisplays, organic electronics is rapidlyemerging as an important topic, so materialsand device research activities in OLEDs willenhance growth in organic electronics.

Electro-Optic Sensors & SystemsThe Electro-Optic Sensors and Systems pro-gram includes papers by world class scien-

tists and engineers on research, develop-ment and applications of optical, electro-optical, and optoelectronic systems and sen-sors for ultra fast systems, 3D display, 3Ddata processing, information systems, secu-rity systems, radar, and image and datasensing. There will be sessions whichinclude papers describing systems and tech-niques for high speed information process-ing and communications systems, opticalinterconnects, light modulators, smart pix-els for information processing, optical stor-age/memory for information systems, threedimensional imaging systems, three dimen-sional image recognition, adaptive optics,radar, fiber optics sensors, and imaging sys-tems for intelligent transportation systems.

Integrated Optics &OptoelectronicsThe integrated optics and optoelectronicssub-committee has put together an excel-lent program covering a broad range of rel-evant topics, including a special sympo-sium on Photonic Integration. There isincreasing interest in the field of photoniccrystals for possible applications in photon-ic routing and signal processing. Enablingdevices are highlighted through invitedand contributed papers. High speed modu-lation, optical amplification, and a varietyof filter technologies are also covered ininvited as well as contributed talks byexperts and researchers in these fields.

Microwave PhotonicsMicrowave and millimeter-wave photonictechnologies are critical for high-speed fiberand wireless communications, and remotesensing applications. Photonic techniquesare now used for the generation, transmis-sion, detection, and processing of millime-ter-wave signals with carrier frequenciesexceeding 100 GHz. This year’s 9 invitedand 7 contributed papers on microwavephotonics are divided into sessions coveringradio-over-fiber communications, advancedcomponents, antennas and receivers, andnovel measurement techniques. Specific

A Preview of the 15th Annual Meeting

LEOS 2002 ANNUAL MEETINGGLASGOW, SCOTLAND

Page 4: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

4 IEEE LEOS NEWSLETTER OCTOBER 2002

topics include millimeter-wave signal trans-mission, optical sampling and down-con-version, optoelectronic antennas, high-speedlasers and modulators, and microwave spec-trum analysis.

Nonlinear OpticsNonlinear Optics – LEOS 2002 has over30 papers evenly distributed between invit-ed and contributed papers. There will betwo sessions on nonlinear effects in micro-and nanostructured materials, which willinclude photonic crystal fibers and quasi-phase matched materials. Another sessionwill be on nonlinear effects in lasersimpacting the design and dynamical effectsinfluencing applications. There will be asession on spatial and temporal solitons. Asession on nonlinear optical materials willinclude talks on the steering of moleculesby multiphoton coherent control, and tai-loring of the complex nonlinear index oforganic, polymeric, and nanostructuredmaterials. The remaining sessions arefocused on nonlinear interactions, includ-ing ultrafast optical signal processing usingsemiconductor quantum dot amplifiers, all-optical switching using near-infrared inter-subband transitions in quantum wells, andanalytical solutions of the nonlinearSchrödinger equation with gain.

Optical CommunicationsThis year’s sessions in OpticalCommunications cover a broad range oftopics from optical components to highcapacity networks. Among the highlightswill be sessions on nonlinear-effects inoptical communications systems, opticalcomponents, polarization mode dispersion,fiber and semiconductor amplifiers, andwavelength conversion. New topicsinclude discussions on the progress in opti-cal signal processing, improving spectralefficiency and reducing nonlinear interac-tions in optical fiber transmission systems.Eleven technical sessions running over allfour days of the conference will featurethirteen invited talks by leading experts inthe field.

Optical Fiber & PlanarWaveguide TechnologyOptical communications networks increas-ingly require real time system optimiza-tion. Advances in optical fiber and planar

waveguide technology are meeting thischallenge. Four sessions in OFPW willcover these areas. One session will focus onadvances in fiber optic switch technology.In another session, measurement and char-acterization science will be covered.Another session will cover new and novelcomponents. These include, gratings andfilters, dynamic gain equalizers, and PMDand dispersion compensators.

Optical Interconnect &Processing SystemsThis year’s symposium on optical intercon-nects and processing systems highlights avariety of new and exciting technologiesand applications. The symposium consistsof 5 sessions featuring both invited andcontributed papers from around the world.One session will focus on optoelectronicswitching systems featuring both passiveall-optical and regenerative optoelectronicmethods. The optical interconnects sessioncontains two invited talks on the latest inVCSEL arrays and applications to PCBinterconnects. The system design andmodeling session will highlight analyticand CAD design methods for optimizingsystem design. Finally there will be twosessions covering novel devices and materi-als, featuring invited talks on novel tech-niques using nanophotonics, MEMs, andholographic-based technologies.

Optical Networks & SystemsTremendous interest in high capacitytransmission systems and optical networkshas been fueled by unprecedented growthin data and internet traffic. Advances innetwork architectures and device technolo-gy enable the transition from point-to-point systems to optical networking. TheOptical Networks and Systems committeeis addressing this trend with sessionsfocusing on optical crossconnects, model-ing of optical networks, and advanceddevices and network concepts. In addition,there is a special session onvisionary/future optical networking ser-vices and technologies. These sessions areanchored by more than a dozen invitedtalks by noted experts in the field.

Optoelectronic Materials &ProcessingOptical Materials and Processing papers

cover a wide range of applications and top-ics including crystal growth of semicon-ductor materials and nanostructures aswell as emerging device fabrication ofgratings, photonic crystals and quantumwell intermixing. Invited papers cover spinopto-electronics, long-wavelength materi-als, InP photonic circuits and photoniccrystal development. Contributed papersspan topics ranging from microcavityemitters for blue to infrared wavelengthsto advances in intermixing technologiesfor device integration.

Optoelectronic Packaging,Manufacturing & ReliabilityFour Optoelectronic Packaging,Manufacturing and Reliability sessionswill feature exciting selected presentationsfrom 16 speakers. A session dealing withdesign and reliability of MEMs and wave-length lockers will set the stage for ses-sions dealing with progress toward compo-nent and module reliability, low cost pack-aging and manufacturing. Presentations inthese sessions treat a variety of packagingissues including: platform technology,assembly and automation. Two sessionscovering the issues involved in manufac-turing and reliability of low-cost compo-nents, modules, and optoelectronic pack-ages conclude these sessions.

Photodetectors & ImagingThe five sessions of Photodetectors andImaging will include seven invited and fif-teen contributed papers with topics rang-ing from new device structures and materi-al systems to novel approaches to photore-ceiver design and integration. Presentationswill include advances in ultra-high speeddetectors, highly linear detectors, theoreti-cal and experimental work on avalanchephotodiodes, and photoreceivers employingboth hybrid and monolithic integration. Afull session will be dedicated to progress onsilicon-based photodetectors for silicon ICintegration, and progress in detectors andarrays for long-wavelength imaging willalso be presented.

Semiconductor LasersSemiconductor lasers continue to be heavi-ly studied with important new develop-ments being rapidly commercialized. Atthe 2002 LEOS Annual Meeting, this area

Page 5: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 5

comprises nine different sessions includingten invited presentations and approximate-ly thirty contributed presentations. Newresearch results will be presented in thealready commercially important technolo-gy areas of high power lasers, vertical cavi-ty surface emitting lasers (VCSELs), andgrating-based lasers for telecommunica-tions and other applications.

In addition, new diode laser researchareas are particularly emphasized. Severalinvited and contributed presentations arein the emerging areas of ultrafast and tera-hertz laser diodes. Recent results thataddress the utilization of novel III-Vnitride materials in long wavelength edgeemitting and vertical cavity lasers are pre-sented. Finally, leading researchers in theareas of quantum dots for lasers and pho-tonic crystal semiconductor lasers willintroduce their work in two sessions devot-ed to these topics.

Short Wavelength and Gas LasersThis session will highlight new develop-ment in the generation of coherent soft x-ray, VUV and UV radiation. Advances inthe generation of soft x-ray radiation byhigh order harmonics, soft x-ray lasers andfemtosecond x-rays from a synchrotronwill be reported. They will focus on thehigh brightness, energy scaling of the highorder harmonic generation, tunability, andsaturated amplification. In addition to thegeneration of soft x-ray radiation, theirnovel applications such as time-resolvedprobing of plasma created by femtosecondlaser pulse and the study of structuraldynamics in condensed matter will be dis-cussed. Significant advances in the UV andVUV radiation will be reported as well.Development of high-power cw 252 nmcoherent light sources for laser cooling ofsilicon atoms will be presented. High-power 157 nm discharge-pumped molecu-lar fluorine lasers for microlithography ofLSI will be reported as well.

Solid State LasersThis section will highlight emerging andnovel technologies in solid state lasers.Sessions will cover recent developments inhigh power lasers, novel laser architectures,fiber laser, diode-pumping and materialsfor non-linear conversion. These papersreflect the maturing states of the solid state

laser field, with most papers representingthe state-of-the-art in their technologyarea. Research results include developmentof UV and visible solid state lasers, non-linear optics, novel fiber lasers, unique cav-ity configurations, and short pulse lasers.

Special Symposium – Lasers inMedicine & BiologyThis symposium consists of four sessionsfocused on some of the hot topics in med-ical optics. These topics deal with the useof optical coherence tomography for imag-ing in a manner that allows penetration ofthe surface of a material; optical andRaman spectroscopy; fluorescent imaging;novel biophotonics and cell growth sen-sors; biophotonic film assembly and genemodulation; photonic microinstrumenta-tion and practical problems in dentistry.This series of talks will bring both expertsand novices up to speed on this field whichis rapidly growing.

Special Symposium –Photonics Time-FrequencyMetrologyThe Special Symposium on PhotonicsTime-Frequency Metrology highlightsexciting new developments in this emerg-ing field. There are seven invited papers,one tutorial, and seven contributed papers,organized in four sessions. The first sessionof the Symposium begins with a tutorialon the subject of precision measurementsthrough control of the phase and frequencyof mode-locked lasers. The tutorial is fol-lowed by an invited talk on the latestdevelopments in connecting the opticalfrequency domain with the RF domain.The second session is devoted to ultra-short pulse metrology, and characterizationof phase and amplitude of optical pulses.This session includes two invited presenta-tions and two contributed ones. The thirdsession of the symposium is organizedaround the subject of optical frequencystandards, with an invited talk on recentresults with optical clocks, and four con-tributed papers describing optical frequen-cy references in various regions of the spec-trum. The last session is on the subject ofultra-low timing jitter. Three invitedpapers describe generation, characteriza-tion, and applications of low jitter signals,with a contributed paper on characteriza-tion of noise. The four sessions represent a

forum where critical aspects of frequency-time metrology are addressed, and whereemerging new directions are emphasized.

Special Symposium –Nonlinear Optics for CoherentSources from THz Radiation toX-RaysStarting with convenient visible or near-infrared laser sources, researchers have tra-ditionally relied on nonlinear optics,through processes such as harmonic andparametric generation, to extend the avail-able spectral range of coherent sources. It isthe purpose of this symposium to presentrecent progress in the coherent generationand detection of radiation from the soft x-rays to the THz region and their applica-tions. In three sessions, the role of nonlin-ear optics in broadening the spectrum ofcoherent radiation is highlighted. The firstsession includes a tutorial about the use ofx-ray radiation for time resolved studies ofexcited matter. The second session gives anoverview of the exciting new area ofattosecond metrology including high-har-monic generation extending into the soft x-ray region. The third session is devoted tothe generation, detection and application offar-infrared and THz radiation.

Special Symposium – AgileOptical Beams andApplicationsThe focus of this special symposium is toreport recent research and technologydevelopments related to the area of agilelaser beams. Papers are solicited in worksfor basic materials, optical beam controldevices, and systems using agile opticalbeams. Example materials include liquidcrystals, and optical micro-electromechani-cal systems (MEMS) fabrication materials.Devices include chip-scale, integratedoptic, and bulk-type optical beamformersusing for example the mentioned materi-als. System applications of interest foragile beams include, for example, lasercommunications, laser radar, and infraredcountermeasures and tracking.

Special Symposium – 40 Gb/sThis year, there will be a special sympo-sium on 40 Gb/s transmission systems.The sessions are highlighted by one tutori-al and 8 invited talks intended to provide

Page 6: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

6 IEEE LEOS NEWSLETTER OCTOBER 2002

insight into current research and commer-cial deployment of 40 Gb/s systems.

Special Symposium – PhotonicIntegrationPhotonic Integration can offer the benefitsof increased functionality at a lower cost.Highlighting this fact, we have organizeda “ Special Symposium on PhotonicIntegration “, that opens with a tutorialtitled “High Density Integrated Optics”by Prof. Herman Haus of MIT, to set thestage. Seven invited talks offer an excellentoverview of this very important area ofresearch and application. The presentationsoffer an even balance between the futurepromise of Photonic crystals and the more

mature integration methods in use today.The four contributed papers selected com-plement the invited talks to round out thisSymposium on Integration.

Ultrafast Optics & ElectronicsThe ultrafast optics and electronics com-munity continues to be very active, with aprogram consisting of five sessions witheight invited and thirteen contributedpapers focusing on the latest advances inthe field of ultrafast phenomena and tech-nology. Some of the ways in which ultrafastoptoelectronics plays an increasinglyimportant role are highlighted in a sessionon broadband femtosecond techniques,which includes an invited talk on the ori-

gin of supercontinuum generation inmicrostructured fibers. These have recentlyproduced a unique coupling between theultrafast and metrology communities (seethe Special Symposium on PhotonicsTime/Frequency Metrology). Terahertz sci-ence and technology is well represented inseveral sessions covering new terahertz toolsand techniques applied to imaging, biologyand semiconductor spectroscopy (see alsothe Special Symposium on NonlinearOptics for Coherent Sources from THzRadiation to X-rays).\ New techniques inthe generation and characterization ofultrashort electromagnetic pulses are cov-ered in several sessions which include invit-ed papers on arbitrary waveform synthesisand high repetition rate sources.

LEOS Profile

Kristian Elmholdt StubkjaerKristian Elmholdt Stubkjaer is a newlyelected member of the LEOS board ofGovernors. He is director for ResearchCenter COM at the TechnicalUniversity of Denmark, where he is alsoa professor.

After college Kristian started studiesof Electrical Engineering at theTechnical University of Denmark in1972. During his study time ProfessorPalle Jeppesen and Professor MagnusDanielsen initiated the field of opticalcommunication in Denmark. It was aneasy decision to join this exciting andvery dynamic field and Kristian endedup specializing in system properties ofsemiconductor lasers both in his Masterthesis project that was completed in1977 and in the Ph.D. study that hecompleted in 1981. Kristian’s Ph.D.study in Denmark was interrupted byan 18-month stay in Japan, where hejoined the well-known group ofProfessor Suematsu at the TokyoInstitute of Technology. Around 1979the Tokyo group was one of the onlyones in the world that could producelong wavelength InGaAsP semiconduc-tor lasers. Being part of the Japaneseresearch team was a unique opportunityand an experience that has brought last-

ing relations with manygood colleagues in Japan.

After the Ph.D. degreeKristian was drafted for mili-tary service and spend timein the army studying under-water acoustics – so slightlylonger wavelengths comparedto optics. This was followedby a one-year employment atthe IBM T.J. WatsonResearch Center in USAwhere he had another chanceto carry out research at one ofthe top laboratories of the world.

In 1983 he became a faculty mem-ber at the Technical University ofDenmark where his research concentrat-ed on active components for optical sys-tems and networks. Much effort wasdevoted to optical amplifiers and lateron to more advanced semiconductorstructures for wavelength conversionand simple signal processing. A fieldin which his team of young, brilliantPh.D. students did many pioneeringexperiments.

He and his research team have alsobeen active on many European researchprojects where collaboration withEurope’s leading companies and univer-sity groups has been extremely stimu-

lating for the research envi-ronment offered to staff andstudents.

Over the years Kristianhas been active on many con-ference committees includingthose for OFC, ECOC,IOOC, ECIO, Optical ampli-fiers and their Applications,IEEE LEOS summer topicalmeetings. This year heserves as the technical pro-gram chair for the EuropeanConference on Optical

Communication that is taking place inCopenhagen in September. He is also aDanish representative in the manage-ment committee of the Europeanresearch program IST that is a 3.5 bil-lon $ program covering InformationTechnologies in a broad sense.

Since 1998 he has devoted most ofhis time to management as director forResearch Center COM that is the centerfor telecommunication at the TechnicalUniversity of Denmark. With 135 staffmembers COM covers central aspects oftelecommunications, including opticalcommunication. Research includescomponent technologies, system andnetwork technologies as well as applica-tions and services.

Kristian ElmholdtStubkjaer

Page 7: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

IEEE LEOS Summer Topical MeetingsJuly 15–17, 2002

Fast Optical Signal Processing in Optical Transmission wasone of the highlights at this years IEEE LEOS SummerTopical Meetings. It is an exciting field that becomes more

and more interesting as the progress in component technologyenables higher efficiency and switching speed. To avoid any misun-derstandings, focus is not on massive or complex signal processingas would be required in optical computing. On the contrary focus ison simple optical processing elements that can play a role in futurehigh capacity transmission systems in overcoming speed limitationsof electronics and at the same time avoid tedious opto-electronicconversions. Simple optical signal processing will also play a role inoptical switching nodes that must handle huge amounts of traffic.In the nodes we need all-optical functions for packet routing, 3Rregeneration, buffering, etc. Again, it must be emphasised that it isa main task to understand the role sharing between electronic andoptical signal processing technologies.

Placed in the wonderful settings of Mont Tremblant in Canadathe meeting with its mixture of invited and contributed papers fromall over the world gave a very comprehensive coverage of the state ofoptical signal processing. Both the components and systems aspectswere covered in the densely packed two and a half day program.

A number of papers covered the systems perspectives of signalprocessing in transparent optical networks, packet networks,OTDM packet networks as well as in OTDM systems generally.Besides a joint session with two of the other collocated topicalmeetings addressed all optical networking issues. The presentations

served the very important task of giving a better understanding ofwhere all-optical signal processing has a role to play.

On the components and sub-unit level a number of contributionsfocused on signal processing using semiconductor optical amplifiers(SOA’s) in interferometric structures. They operate at data rates exceed-ing the capabilities of electronics and a number of presentations con-centrated on achieving even higher switching speeds by optimising theoperation of the structures and also by tailoring their material proper-ties. Experiments at data rates higher than 300 Gbit/s were reported. Anumber of presentations showed how the components can be used forlabel swapping, header correlation and recognition, OTDM channelselection, bypass switching, all optical XOR, bit level synchronizationand other important functionalities. All-optical clock recovery was alsopresented in a number of papers demonstrating the steady progress inthe realization of all-optical clock recovery in TDM as well as Packetnetworks. Also units for duobinary transmission of optically waveformshaped pulses were reported. Moreover, one presentation gave state ofthe art of parametric optical amplifiers that is a field of increasinginterest because of the emergence of new non-linear materials .

This LEOS Newsletter features extended versions of the invitedconference papers. As you will see optical signal processing hascome a long way. It is, however, also a field where there is stillmuch exciting research to be done in developing new componentsand also new systems concepts. I am convinced that we will seemore and more examples where it makes sense to implement opti-cal processing units in future optical systems.

We help entrepreneurs build great companies.

855 MAUDE AVENUE MOUNTAIN V IEW, CA 94043 T 650 960 2383 F 650 960 0933

i N C U B i C

Visit our portfoliocompanies online:

www.lightconnect.com

www.oepic.com

www.opvista.com

www.precisionphotonics.com

www.incubic.com

Featuring: LIGHTCONNECTThis dynamic components company was founded by top technologists Asif Godil and David Bloom. Their fundamental diffractive MEMS concept was licensed from Stanford University, then expanded with significant telecom- specific inventions at LIGHTCONNECT. Today, the company ships multiple leading products to top system and subsystem builders. Its VOA (variable optical attenuator) is the world's smallest, fastest, and first ever MEMS VOA to pass Telcordia reliability testing.

The DGE (dynamic gain equalizer) dynamically flattens EDFA’s for DWDM, providing unparalleled performance on all fronts: resolution, speed, size, power dissipation.

Copyright ©2002 Incubic, LLC All rights reserved.

ww

w.in

cubi

c.co

m

OCTOBER 2002 IEEE LEOS NEWSLETTER 7

Page 8: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage
Page 9: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 9

All-Optical Clock Recovery Using Self-Pulsing Two-SectionGain-Coupled DFB Lasers

GUIFANG LI, YUHUA LI and WEIMING MAO

School of Optics/CREOL, University of Central Florida, Orlando, FL 32816-2700

ABSTRACTUltrafast all-optical clock recovery up to 100 Gb/s has beendemonstrated using two-section gain-coupled DFB lasers. Thissimple and compact clock recovery circuitry is wavelength andpolarization insensitive, and has lower jitter, high sensitivity, largedynamic range, fast lockup time and long hold time.

All-optical clock recovery (CR), in general, will potentially beused when data rates of optical communication systems exceed elec-tronic capabilities. One of the applications of all-optical clockrecovery is OTDM demultiplexing, where the OTDM line rate ishigher than the electronic bottleneck. Both optoelectronic phase-locked loops (PLL) [1, 2] and passively mode-locked semiconductorlasers MLL [3] have been demonstrated for this purpose. Theseapproaches recover a subharmonic clock signal, and hence are notsuitable for another application of all-optical clock recovery, namely,optical 3R (retiming, reshaping and reamplification) regeneration,which requires a recovered clock at the optical line rate.

High-speed optical CR at the optical line rate has been realizedusing actively mode locked fiber lasers [4, 5] and passively modelocked semiconductor lasers [6]. Fiber lasers have been shown to beable to operate above 100 Gb/s but suffer from high power require-ment and are very complicated. Passively mode-locked semiconductorlasers have shown to operate up to 40 Gb/s. An alternative approachis to use self-pulsing gain-coupled DFB lasers [7]. The two-sectiongain-coupled DFB laser has two sections that share a common sub-strate. The contacts for the two sections are electrically isolated.Unlike index-coupled DFB lasers, the gain-coupled DFB grating isetched into the gain region [8]. Gain coupling leads to a large (> 40dB) single-mode suppression ratio (SMSR) in single-section DFBlasers. For two-section gain-coupled DFB lasers, gain coupling results

in relatively independent lasing in the two sections. As a result prop-er design and operation of the device lead to self pulsing: periodicintensity output with DC bias currents for both sections [9].

Experiments have shown that self pulsing can be injectionlocked to optical clock signals embedded in data streams thus real-izing the clock recovery function. There are three modes of opera-tion for all-optical clock recovery using two-section gain-coupledDFB lasers [10]. The first mode is the so-called the incoherentclock recovery as shown in Figure 1. The CR circuitry is character-ized by the following: 1) Injection data is carried on a wavelengththat is at least 1 nm from that of the free running TS DFB, and 2)Injection power is >1 mW. The mechanism of CR is also shown inFigure 1. Injection data creates carrier modulation at the clock fre-quency in the front section of the two-section DFB laser. This carri-er modulation creates sidebands from the mode in the front section.One of the resulting sidebands all-optically injection locks themode in the back section. Beating of the two modes results in a

SPECIAL ISSUE

AMP

I1 I2

TSDFB

BPFλ0

λ1RZ Data

λ

Second modeFirst mode

Side bands

Figure 1. Schematic and mechanism of incoherent clock recovery.

Figure 2. Schematic and mechanism of coherent clock recovery. Figure 3. Schematic of wavelength- and polarization-insensitive coherent clock recovery.

Page 10: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

10 IEEE LEOS NEWSLETTER OCTOBER 2002

recovered clock. The advantages ofthis mode of CR are 1) it is wave-length insensitive as the CR mecha-nism depends on intensity-inducedcarrier modulations, and 2) it can bepotentially polarization insensitive ifthe quantum-well gain region isproperly strained. The disadvantagesof this mode of CR are 1) it requiresrelatively high injection power, i.e.,the sensitivity of CR is low, and 2)jitter of the recovered clock is rela-tively high.

The second mode is the so-calledcoherent clock recovery as shown inFigure 2. The CR circuitry is charac-terized by the following: 1) Injectiondata carried on almost the same wave-length as that of the free-running TSDFB, and 2) Injection power is≈ 50µW . The mechanism of CR isalso shown in Figure 2. The injectiondata contains sidebands due to exis-tence of the clock component. One ofthe sidebands is aligned with thefront section and the optical carrier isaligned with the back section. Opticalinjection locking takes place in bothsections. Beating of the two opticallyinjection locked modes results in a recovered clock. The advantagesof this mode of CR are 1) it requires extreme low injection power,i.e., the sensitivity of CR is high, independent of the bit rate and 2)the jitter of recovered clock is very small. The disadvantages of thismode of CR are 1) it is wavelength insensitive, and 2) it is polariza-tion insensitive even if the quantum-well gain region is properlystrained.

The third mode is the so-called wavelength- and polarization-insensitive coherent clock recovery as shown in Figure 3. The injec-tion data is carried on arbitrary wavelength. Probe wavelength isnominally the same as free-running TS DFB Laser. The semiconduc-tor optical amplifier (SOA) is both wavelength- and polarization-insensitive. The injection power into the SOA is < 100 mW. Themechanism for this wavelength- and polarization-insensitive coher-ent clock recovery scheme is straightforward. Clock component car-ried on the arbitrary wavelength of the injection data is converted toprobe wavelength, which is the same as the free-running DFB laser.The wavelength conversion is insensitive to both wavelength andpolarization. After the bandpass filter, CR operates in the coherentmode, identical to the coherent mode as shown in Figure 2.

The wavelength and polarization-insensitive coherent clockrecovery is the preferred mode of operation since it combines theadvantages of the previous two modes. It is both wavelength andpolarization insensitive and it offers low jitter. Compared to theincoherent clock recovery scheme, the only additional component isthe probe laser. This wavelength- and polarization-insensitivecoherent clock recovery circuitry has demonstrated the followingcharacteristics simultaneously:

• Recovered clock signal should have low jitter (<1 ps at 40Gb/s and < 250 fs at 100 Gb/s)

• high sensitivity and large dynamicrange (-10 dBm and 7 dB, respectively)• high-speed operation above electron-ic speed limit (at least 100 Gb/s)• wavelength insensitive• polarization insensitive• a fast lock up time for reducedlatency (<1ns)• a long hold time to eliminate pat-tern-dependent effects (5 ns)• simple, compact, low-power con-sumption and reliable

Figure 4 contains time-domainmeasurements of the input data andthe recovered clock at 96 Gb/s madepossible by the recent availability ofthe Terascope by Agilent. The eye dia-gram of the input data is shown inFig. 4(a). Some non-uniformity can beobserved in the 8 optical time-divisionmultiplexed (OTDM) channels at 12Gb/s, each of which is a pseudo-ran-dom bit stream (PRBS) with a patternlength of 231 − 1. The recovered clockat 96 GHz is shown in Fig. 4 (b). Theclock recovery circuitry successfullyremoved the pattern effect in theOTDM channels. The RMS timing jit-

ters of the recovered clock was 499 fs compared with the 436 fsRMS jitter of the original OTDM signal. The jitter of the originalOTDM signal is mainly due to OTDM multiplexing. Assumingthat the jitter of the original OTDM signal and the intrinsic jitterinduced by the clock recovery circuitry (wavelength converter andthe TS-DFB laser) are the only jitter sources of the recovered clockand they are statistically independent, the intrinsic jitter of the clockrecovery circuitry is

√4992 − 4362 ≈ 243 fs, which is very close to

the instrument timing jitter (200fs) of the Terascope.This work has been supported by the National Science

Foundation.

REFERENCES1. D. T. K. Tong, et. al., IEEE Photon. Technol. Lett., vol. 12, pp.

1064-1066, Aug. 2000.2. O. Kamatani, et. al., IEEE Photon. Technol. Lett., vol. 8, pp. 1094 -

1096, Aug. 1996.3. I. Ogura, et. al., ECOC’97 Technical Dig., vol. 2, pp. 77-80.4. S. Bigo, et. al., IEEE J. Select. Topics Quantum Electron., vol. 3, pp.

1208-1223, Oct. 1997.5. H. Yokoyama, et. al., Proc. CLEO/pacific Rim 2001, vol. 2, pp.

498-499.6. X. Wang, et. al., J. Opt. Soc. Am. B, vol.16, pp. 2030-2039, Nov. 1999.7. W. Mao, et. al., Electron. Lett., vol. 37, pp. 1302 –1303, Oct. 2001.8. B. Sartorius, et. al., IEEE J. Select. Topics Quantum Electron., vol. 1,

pp.535-538, June 1995.9. C. Bornholdt, et. al., OFC’2002 Technical Dig., Paper TuN6, pp. 87-89.10.W. Mao, et. al., to appear in IEEE J. Lightwave Technol., Sept.

2002.

(a)

(b)

Figure 4 Eye diagram of the input data signal (a) and waveform of therecovered clock (b) at 96 Gb/s.

Page 11: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

Monolithically Integrated Wavelength Converter:Sagnac Interferometer Integrated with Parallel-AmplifierStructure (SIPAS) and Its Application

YASUHIRO SUZUKI, TOSHIO ITO, and YASUO SHIBATA

NTT Photonics Laboratories, NTT Corporation3-1, Morinosato Wakamiya, Atsugi-shi, Kanagawa, 243-0198 JapanTel: +81 46 240 2847, Fax: +81 46 240 4383, E-mail :[email protected]

AbstractA Sagnac interferometer monolithically integrated with parallel-amplifier structure (SIPAS) is described. Filter-free wavelengthconversion and full bit-rate conversion from 10-Gb/s randomWDM channels to a 40-Gb/s channel were demonstrated.

IntroductionThe explosive growth of the internet has pushed an increase in thebit rate of WDM networks and has accelerated the construction ofphotonic networks, which have the potential to overcome the speedlimit of electrical devices through all-optical signal processing.

All-optical wavelength conversion will play an important role infuture photonic networks. A wavelength converter based on thecross phase modulation (XPM) in a semiconductor optical amplifier(SOA) is one of most promising devices. Differential phase modula-tion (DPM), which uses XPM in a differential scheme, can over-come the speed limitation of carrier lifetime in an SOA[1].Wavelength conversion at over 100 Gbit/s has been reported usingDPM[2]. In this scheme, an optical filter is necessary in order toreject the input signal. In the case of wavelength-tunable conver-sion, the response time of this filter may limit system performance,so simple filter-free operation is desirable [3,4].

With the increase in the bit rate of networks, bit-rate conversionfrom low-speed (<10-Gb/s) WDM-LANs to high-speed (40-Gb/s)networks is also desirable. All optical bit-rate conversion providesbit-rate-free operation.

In this article, we first describe filter-free wavelength conversionusing a newly developed Sagnac interferometer integrated with paral-lel-amplifier structure (SIPAS)[5,6]. SIPAS has low wavelength depen-dence, single-signal input without using a delay line and can dividethe input and converted signal. We then discuss full bit-rate conver-sion from 10-Gb/s random WDM channels to a 40-Gb/s channel,including non-return-zero (NRZ) to return-to-zero (RZ) format con-version and reconversion, as an application of SIPAS. Clear eye open-ings and low power penalties were observed in experimental testing.

Configuration of SIPAS SIPAS is a Sagnac interferometer with parallel-ampli-fier structure (PAS), which is a Mach-Zehehnderinterferometer (MZI) having polarization insensitiveSOAs in each arm. It was fabricated using monolithicintegration techniques (Fig. 1), and is 4 mm long.

The operating principle is similar to that of theSLALOM [7]. An input CW light is divided into

clockwise (CLW) and counterclockwise (CCW) traveling lights.Since the PAS is placed asymmetrically in the loop, two lights reachthe SOAs at different times, which leads to a different phase modu-lation between the lights when a signal light is input into theSOAs. After traveling the loop, the lights are superimposed andtransmitted to the output port due to differential phase modulation(DPM). We placed the PAS asymmetrically by 0.5mm so that theswitching window due to DPM is about 10 ps, which enables high-speed operation over 40 Gbit/s. As the PAS is set in the cross state,

the signal light cannot enter the loop, resulting infilter-free wavelength conversion.

Filter-free wavelength conversionWe first examined the filtering effect of the PAS.A CW light and signal light were injected fromthe CW and signal-input ports respectively.

Figure 2 shows the spectrum from the output

Fig.1 Structure of SIPAS

signalCW

1540 1545 1550 1555 1560 1565 1570-70

-65

-60

-55

-50

-45

-40

-35

-30

-25

-20O

ptic

alPo

wer(

dBm

)

Wavelength (nm)Wavelength (nm)

Opt

ical

Pow

er(d

Bm

)

1540 1545 1550 1555 1560 1565 1570-70

-65

-60

-55

-50

-45

-40

-35

-30

-25

-20

Opt

ical

Powe

r(dB

m)

Wavelength (nm)

signal

CW

27 dB

Wavelength (nm)

Opt

ical

Pow

er(d

Bm

)(a) (b)

Fig.2 Spectrum of output optical signal:signal: 1552.6 nm CW:1550.0 nm

(a) PAS off, SOA1: 212 mA, SOA2: 0 mA(b) PAS on, SOA1: 212 mA, SOA2: 217 mA

(25 ps/div)

Fig.3 Eye pattern in filter-free operation

OCTOBER 2002 IEEE LEOS NEWSLETTER 11

Page 12: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

12 IEEE LEOS NEWSLETTER OCTOBER 2002

port. Figure 2(a) shows the case when a drive current was injected intoonly SOA1, which corresponds to the conventional SLALOM[7]. Bothconverted and signal lights were observed. When drive currents wereinjected into both SOAs, on the other hand, the signal light was sup-pressed and converted light was increased as shown in Fig.2(b). Thesuppression ratio was as large as 27 dB, as shown in the figure, whichis large enough for filter-free operation.

We then performed filter-free wavelength conversion experi-ments under the conditions shown in Fig. 2(b). The signal lightwas modulated with a 10-Gb/s RZ signal. Figure 3 shows the eyepattern of the converted signal in filter-free operation. Clear eyeopenings were observed. A power penalty of 0.9 dB was obtainedcompared to back-to-back in filter-free operation.

Application: Full bit-rate conversion from 10-Gb/srandom WDM channels to a 40-Gb/s channelUsing SIPAS, we demonstrated bit-rate conversion. Figure 4 showsthe configuration of the demonstration system. The key featuresare as follows.

<NRZ/RZ and MUX>Four 100-GHz spacing WDM channels of 10-Gb/s NRZ format

(point a) are launched into an EA modulator and converted simulta-neously to RZ formats (point b). A fiber loop arranges the piled upRZ pulse into a serial bit stream. This simple technique for streamforming can also be used in large-scale WDM systems with 8 or 16channels, for example.

<Bit-rate conversion using SIPAS>For 40-Gb/s bit-rate conversion, the four different wavelengths

of the multiplexed bit stream are converted into a single wave-length by using SIPAS (point c). SIPAS has a low wavelengthdependency.

<DEMUX and RZ/NRZ>An EA modulator is used to demultiplex the converted 40-Gb/s

stream into 10-Gb/s RZ format (point d). Another DPM devicewith a gating window of 100 ps enlarges the pulse width, thuscompleting the 10-Gb/s RZ to NRZ reconversion (point e).

<Experimental result>Figure 5 shows the eye pattern measured at points (a) to (e). Clear

eye openings were observed at each point. Figure 6 shows the bit errorrate from points (a) to (e). The receiver sensitivity at an error rate of10-9 was less than -32 dBm (Fig. 6). Only a small power penalty ofless than 0.8 dB was observed.

ConclusionWe have developed a monolithically integrated Sagnac interferom-eter integrated with parallel-amplifier structure (SIPAS). Filter-freewavelength conversion at a bit rate of 10 Gb/s and full bit-rateconversion from 10-Gb/s random WDM channels to a 40-Gb/schannel were demonstrated. These results show that the SIPAS willbe useful in future high-bit-rate photonic networks.

References1. B. Mikkelesen et al., Electron. Lett., vol. 33, pp. 2137 (1997)2. Y. Ueno et al., Proc. of ECOC 2000, Munich, Germany, pp. 13

(2000)3. J. Leuthold et al., IEEE J. of Lightwave Technol. vol.17, pp.1056

(1998) 4. D. Wolfson et al., Proc. of OFC 2000, Baltimore, USA, paper TuF35. Y. Shibata et al., Proc. of OECC/IOOC 2001, Sydney, Australia, pp.

212 (2001)6. T. Ito et al., Proc. of OAA 2001, Stresa, Italy, paper OWA3-17. M. Eiselt et al., J. of Lightwave Technol. vol.13, pp. 2099 (1995)

SIPAS

DEMUXMUX WDM/TDM

Fiber Loop

DPM

EAEA(a) (b) (c) (d) (e)

λ1

λ2

λ3

λ4

λ1- λ4

λ5 λ6

Fig. 4. Full bit-rate conversion

Fig. 5. Eye patterns in bit-rate conversion and RZ/NRZ conversion

-36 -34 -32 -30 -28 -26received optical power, dBm

10 Gb/s 27-110-4

10-5

10-12

10-10

10-8

10-6

bite

rror

rate

backtobackfull bit-rate conversion

Fig. 6. Bit error rate in bit-rate conversion

Page 13: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 13

All-Optical Processing in Switching Networks

PAUL R. PRUCNAL, VARGHESE BABY, DARREN RAND, BING C. WANG, LEI XU, and IVAN GLESK

Department of Electrical Engineering, Princeton University, Princeton, NJ 08544

In the backbone of today’s high per-formance networks, optical fibersprovide enormous point-to-point

communications capacity. With thedeployment of DWDM equipment,aggregate throughputs on the order of afew Tbps per fiber are being achieved [1].However, despite the recent success offiber optics, it has so far been used pri-marily as a low loss, high bandwidthreplacement to electrical cable in point-to-point transmission links. In these sys-tems, optical signals are usually convertedto the electrical domain at intermediatenodes in order to perform switching andsignal processing. For example, in the Internet, electronicswitches are used to route packets to their destinations.However, in this approach, the maximum serial line rate islimited by the bandwidth of electronics, which is consider-ably less than the bandwidth available in optical fiber. Ineffect, an “electronic bottleneck” is created in the system.This article summarizes the research efforts at PrincetonUniversity towards the development of network nodes capa-ble of all-optical signal processing and routing.

Optical time division multiplexing (OTDM) based tech-niques have demonstrated the capability to provide routing,switching and processing at very high speeds. In OTDM,the electronic baseband signal is modulated on narrow puls-es that occupy a small fraction of the bit period. The opticalpulses from different baseband channels are multiplexedinto different timeslots to create an aggregate line rateequal to the number of channels multiplied by the rate ofthe baseband electronic channel. Since OTDM assigns dif-ferent temporal positions within the bit period to differentchannels, the processing latency of OTDM based tech-niques is determined by the timeslot access time. Highlyscalable architectures of timeslot channel selectors havebeen demonstrated with timeslot channel access latency of afew bit periods [2,3].

Future optical networks may exploit the benefits of both OTDMand WDM technology. Fig. 1 shows the architecture of an OTDM-WDM hybrid network node capable of signal processing. The nodeincludes several all-optical functions: routing, 3R regeneration (re-amplifying, re-shaping and re-timing), and wavelength and formatconversions. Regeneration is done to compensate for the degrada-tion of the signal quality during transmission. Wavelength conver-sion is useful in avoiding wavelength blocking in transmission sys-tems, thus greatly increasing the flexibility and the capacity of thesystem. Format conversion is needed as an interface technologybetween network layers that use different formats.

All-optical switches in an OTDM-WDM framework present aversatile approach to all-optical processing since a variety of appli-

cations can be performed without significantly changing the devicearchitecture. Various approaches have been proposed for the devel-opment of such switches. In general, all these approaches involvethe utilization of a nonlinearity in various media like optical fiber,passive waveguides and active semiconductor optical amplifiers(SOA). Although passive all-optical switches have demonstratedthe fastest switching capabilities to date, they typically requirehigher switching energies than the active switches based on thenonlinearity of the SOA. Also, ultrafast all-optical SOA-basedswitches are very compact and offer the possibility for monolithicintegration together with other photonic devices. These SOA-based all-optical switches were initially limited by the slow recov-ery rate associated with the active nonlinearity in the SOA.

Fig. 1: The architecture of a generic node in a transmission network with the various signal processing capabilities

SOA

InputCW Input

CCW

Control

50:50

Filter

Input

Control

Output

Fig. 2: The structure of the Terahertz Optical Asymmetric Demultiplexer (TOAD)

Page 14: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

14 IEEE LEOS NEWSLETTER OCTOBER 2002

Research at Princeton University has led to the development of theTerahertz Optical Asymmetric Demultiplexer (TOAD) [4], basedon a Sagnac interferometric structure (see Fig.2), that allows signalprocessing functions at rates faster than the SOA recovery time byutilizing a differential onset of the nonlinearity in the SOA. It canbe viewed as a generic 3-port device, with an input, control, andoutput port.

The TOAD is essentially a fiber loop joined at the base by anoptical coupler, which splits an input signal into two equal partsthat counter-propagate around the loop and recombine at the cou-pler. In addition, there is a coupler on the loop for the injection ofa control signal, and an SOA in the loop to produce a large phaseshift in the data signal.

In the absence of a control signal, the device is off—the twoparts of the input pulse see the same medium as they go aroundthe loop, arrive at the coupler in phase, and return to the inputfiber. To turn the switch on, a control pulse is inserted thatdepletes the gain of the SOA and, according to theKramers–Kronig relations, changes the index of refraction. Bycarefully timing the control pulse, we can induce a phase differ-ence between the two counter-propagating input pulses, whichpass through the SOA at different times. If the phase shift isproperly adjusted, the two parts recombine at the input coupler insuch a way that the whole signal passes to the output fiber. Thecontrol signal is eliminated at the output by a polarization orwavelength filter.

Because the SOA recovers slowly from the blast of the controlpulse, input pulses that enter the switch immediately after thecontrol pulse see the SOA in approximately the same recoverystate, and are therefore affected similarly. The temporal duration ofthe switching window τwin can be adjusted by moving the locationof the SOA, according to the relation τwin = 2∆xSOA/cfiber, where∆xSOA is the distance of the SOA from the center position of theloop and cfiber is the speed of light in the fiber. Switching windowswith a temporal width approaching 1 ps have been achieved withlow control pulse energies (less than 500 fJ).

The same principle of the differential onset of the fast nonlin-earity in an SOA has been used to build other all-optical devices inthe Mach-Zender [5] and Michelson interferometric configura-tions. An integrated version of a Mach-Zehnder all-optical switchusing an asymmetric twin-waveguide structure was demonstratedat Princeton University [6]. This structure couples the active andpassive optical waveguides by growing the active layer on top ofthe passive guide, thus reducing the optical coupling loss betweenthe active and passive waveguides. This also results in improvedyield since device growth can be performed in a few steps and doesnot require re-growth.

As mentioned before, the temporal length of the switchingwindow is determined by the offset of the SOA in the loop. Theposition of the SOA in the loop and the configuration of theTOAD can be varied to obtain the desired optical processing func-tionality. For example, in high-speed demultiplexing applications,a short switching window is used. The aggregate data stream isinjected into the input port and the clock pulses into the controlport [7]. However, for 3R regeneration, a large switching windowis used with the data stream injected into the control port and thelocal clock pulses into the input port [8]. Short switching windows

have also been used in high-speed analog sampling [9], whereaslonger switching windows have been used for wavelength and for-mat conversion [10] and all-optical packet routing [11]. TheTOAD has been employed in network demonstrations for applica-tions in photonic packet switching [12] and high speed intercon-nects [13].

In summary, the TOAD has demonstrated the versatility toperform many processing functions all-optically. Due to the lowcontrol energy requirements and potential for integration of thisdevice, it is expected to present a viable approach to all-opticalsignal processing in both transmission systems and local areanetworks.

References1. A. K. Srivastava, S. Radic, C. Wolf, J. C. Centanni, J.W. Sulhoff,

K. Kantor and Y. Sun, “Ultra-dense terabit capacity WDM trans-mission in L-band”, IEEE Photonics Technology Letters 12, 1570(2000).

2. K.-L Deng, K.I. Kang, I. Glesk, P.R. Prucnal, “A 1024-channelfast tunable delay line for ultrafast all-optical TDM networks,”IEEE Photonics Technology Letters 9, 1496 (1997).

3. B. C. Wang, I. Glesk, R. J. Runser, and P. R. Prucnal, “FastTunable Parallel Optical Delay Line,” Optics Express 8, 559 (2001).

4. J. P. Sokoloff, P. R. Prucnal, I. Glesk, M. Kane, “A Terahertz opti-cal asymmetric demultiplexer,” IEEE Photonics Technology Letters 5,#7, pp 787-790 (1993)

5. K.I. Kang, I. Glesk, T.G. Chang, P. R. Prucnal, and R. K. Boncek,“Demonstration of all optical Mach-Zehnder demultiplexer,”Electronics Letters 31, 749 (1995).

6. P. V. Studenkov, M. R. Gokhale, J. Wei, W. Lin, I. Glesk, P. R.Prucnal and S. R. Forrest, “Monolithic, all-optical Mach-Zehnderdemultiplexer using an asymmetric twin-waveguide structure,”IEEE Photonics Technology Letters 13, 600 (2001).

7. I. Glesk, J. P. Sokoloff, and P. R. Prucnal, “Demonstration of all-optical demultiplexing of TDM data at 250 Gbps,” ElectronicsLetters 30, 339 (1994).

8. Bing C. Wang, Lei Xu, Varghese Baby, Deyu Zhou, Robert J.Runser, Ivan Glesk, Paul R. Prucnal “Experimental Study on theRegeneration Capability of the Terahertz Optical AsymmetricDemutliplexer” Optics Communications 199, 83 (2001).

9. K. -L. Deng, R. Runser, I. Glesk, and P. R. Prucnal, “Single-shotoptical sampling oscilloscope for ultrafast optical waveforms”,IEEE Photonics Technology Letters 10, 397 (1998).

10.R. J. Runser, D. Zhou, C. Coldwell, B. C. Wang, P. Toliver, K. -LDeng, I. Glesk, and P.R. Prucnal, “Interferometric ultrafast SOA-based optical switches: From devices to applications.” Optical andQuantum Electronics 33, 841 (2001).

11.I. Glesk, K. I. Kang, and P. R. Prucnal, “Demonstration of ultra-fast all-optical packet routing,” Electronics Letters 33, 794 (1997).

12.P. Toliver, I. Glesk, R. J. Runser, K. -L. Deng, B. Y. Yu, and P. R.Prucnal, “Routing of 100 Gb/s words in a packet-switched opticalnetworking demonstration (POND) Node,” IEEE Journal ofLightwave Technology 16, 2169 (1998).

13.K. -L. Deng, R. J. Runser, P. Toliver, I. Glesk, and P. R. Prucnal,“A highly scalable, rapidly-reconfigurable, multicasting-capable,100-Gbit/s photonic switched interconnect based upon OTDMtechnology,” IEEE Journal of Lightwave Technology 18, 1892 (2000).

Page 15: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

All-Optical Transmission and Wavelength Conversion of 40Gb/s Signals over One Million Kilometers of Fiber

J. LEUTHOLD, G. RAYBON, Y. SU, and R.J. ESSIAMBRE

Lucent Technologies, Bell Labs, Holmdel, NJ 07733, USA

AbstractAll-optical 3R regenerators and regenerative all-optical wavelengthconverters are key components in future transparent networks. In atransparent network, transmission distances are not a priori knownand therefore regeneration is needed to reduce impairments thataccumulate. In addition, wavelength converters are needed to over-come blocking issues at network nodes. Here we review recentprogress towards fully regenerative schemes that allow transmis-sion of 40 Gb/s signals over virtually unlimited distances, i.e. wedemonstrated all-optical transmission over one million kilometers.

Introduction3R all-optical regenerative devices that can perform re-amplification(1R), reshaping (2R), and re-timing (3R) hold promise to replaceoptical-to-electrical-to-optical (OEO) components in future transpar-ent networks. In particular, 3R regenerators that do not rely onpower consuming broadband electronics [1]-[4] may replace OEOregenerator units that are needed to overcome transmission impair-ments. Additional functionality is needed in networks with opticalcrossconnects, where optical translator units are used to overcomewavelength blocking. Regenerative all-optical wavelength converters(AOWC) might replace OEO translator units in such crossconnects[5]. As a likely scenario “islands of transparency” [6] will firstemerge. In these islands of transparency, the aforementioned deviceswill enable transmission from source to desti-nation without intermediate conversion intothe electronic domain (Fig. 1).

This article reviews techniques that havedemonstrated full optical signal regenerationand thus enable transmission over virtuallyunlimited distances.

3R Regeneration in SolitonSystemsLong-haul experiments based on 40 Gb/sdispersion managed soliton transmissionhave demonstrated transmission distances ofup to 70,000 km at regenerator spacings of250 km [7]. To achieve these results bothreshaping based on soliton transmission andretiming exploiting the so-called synchro-nous modulation technique [8] has beenused. The soliton transmission techniqueprovides a very natural form of pulse shapeconservation since a soliton pulse preservesit’s form during the propagation along thefiber. The synchronous modulation tech-nique on the other hand provides retiming

by remodulating a signal after a certain transmission distance withthe clock derived of it’s own signal. An example of a synchronousintensity modulator is shown in Fig. 2(b). 10 Gb/s soliton trans-mission based on such techniques have already demonstrated trans-mission distances of up to one million kilometers in the past [8][9].

XCXC

XCXC

XCOEO

MetropolitanNetworkLong-Haul

Network

OEOCustomer

Island of TransparencyOEO

OEO

Fig. 1 Future networks are likely to consist of “islands of transparency” . In these islandsof transparency, the signals are transmitted without intermediate conversion into the elec-tronic domain from source to destination (Fig. 1). The usual optical-to-electrical-to-opti-cal (OEO) converter at network interfaces will be replaced by all-optical devices.

~~~

HNLFSMF

HNLF

1 nm

2R Regenerator (2 stages)

2.5 km0.5 km

2 km~~~

HNLF

1 nm

2R Regenerator

2 km

EA

Q

(c) Retiming and Wavelength Converter (Clock Recovery & Wavelength Converter)

Pclk

CR

EAQ

SOADI

SOA-DI

SA~~~

1 nm

Retiming &WavelengthConversion

(b) Retiming (Synchronous Modulator)

Pin

Pcv,3R

P2R

Preg

P2R

P2R P3R

P3R

(a) 3R Regenerator - Scheme

cw

Fig. 2 (a) Scheme of the 3R regenerator comprising fiber 2R regenerators and retiming elements. (b) SynchronousModulator scheme. (c) Clock recovery and semiconductor optical all-optical wavelength converter for performing retim-ing and wavelength conversion.

OCTOBER 2002 IEEE LEOS NEWSLETTER 15

Page 16: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

16 IEEE LEOS NEWSLETTER OCTOBER 2002

3R Regenerators for Pseudo-Linear TransmissionSystemsThe soliton transmission technique is difficult to apply over commer-cially deployed fibers. Instead a pseudo-linear transmission techiqueis used [10]. A 3R regenerator that works in the pseudo-linear trans-mission regime is shown in Fig. 2. The regenerator consists of threeelements [11]. It comprises two fiber based regenerators for signalreshaping and reamplification (2R) and a retiming section (3R).

The fiber based 2R regenerator is shown in more detail in Fig.2(a). The 2-stage fiber regenerator (left side of Fig. 2(a)) contains acompression stage and a regenerator stage [11]. The compressionstage is for suppressing Brillouin scattering. In the regeneratorstage self-phase modulation (SPM) of the nonlinear fiber is exploit-ed to broaden the optical spectrum of the input signal. A 1-nmoptical filter is placed at the output to select components of thesame intensity in the SPM modulated spectrum. Components withless intensity have less broadening and are therefore suppressed bythe filter. Components with higher intensity give larger broadeningand are brought back to identical intensities determined by the fre-quency offset of the filter. The 2R stage also provides some smallwavelength shift of 1.5 nm. The second single-stage regenerator istherefore used to reset the signal wavelength to its input value. Forall-optical regeneration where no wavelength conversion is needed,a simple 40 Gb/s synchronous intensity modulation scheme is usedto perform retiming, Fig. 2(b). The 40 Gbit/s signal is intensitymodulated synchronously by a sinusoidally driven electroabsorption(EA) modulator. The clock signal is derived from a very simpleclock recovery scheme as shown. The signal is detected using a 40GHz pin photodiode, then filtered using a very high Q (~900)microwave filter and then re-amplified before applying to thereverse biased modulator. The synchronous modulator can be placedeither in between the two 2R regenerators or before the first one.

Regenerative Wavelength ConversionWhen wavelength conversion is needed, the synchronous modula-tor is replaced with a 3R semiconductor optical amplifier (SOA)based delay interferometer (DI) wavelength converter [3]. In thisdevice, the input signal P 2R is used for gating the SOA-DI, suchthat the clock signal P clk can pass through the device dependingon whether the gate is opened or closed. The signal that passes thegate is the clock signal carrying the new wavelength of the clocksignal. Wavelength conversion can be performed to any wave-length within approximately 30 nm around the SOA gain maxi-mum. To further improve the signal quality after the wavelength

converter we added a saturable absorber (SA) behind the filter atthe output of the SOA-DI. The additional 2R regenerator at theoutput of the scheme improved the signal quality only little. Toperform wavelength converter over just a few 400 km loops we didnot need the additional fiber regenerator at the output. On theother hand it was needed to provide a perfect 3R regeneration suchas needed to demonstrate transmission over thousands of 400 kmloops. The clock signal is derived of the input signal. In our casewe use the 40-GHz pin photodiode and the high-Q filter of aboveto detect the signal and to drive an EA modulator which generatesthe clock signal. However, all-optical clock recovery schemes suchas suggested by Sartoris et al. could be used [13] and thus com-pletely eliminate the need for high-speed electronics.

Loop SetupA simplified loop set up to demonstrate long distance all-opticalregeneration is shown in Fig. 3(b). The transmitter is a 40 Gb/s,33% duty cycle, RZ signal that is obtained by multiplexing four10 Gbit/s signals with a PRBS of 231 -1 [11]. The receiver consistsof a high-Q filter based clock recovery and an OTDM demultiplex-er. An electroabsorption modulator demultiplexes the 40 Gbit/ssignals down to 10 Gbit/s for error detection. The transmissionspan consists of four 100 km spans of TrueWaveTM Reduced Slope(TWRS) nonzero dispersion fiber and dispersion compensated fiber.Erbium-doped fiber amplifiers (EDFAs) and backward pumpedRaman amplification was used to compensate the 21 dB span loss-es. A tunable dispersion compensator adjusts the dispersion beforethe regenerator. The launch power into the spans was between -1dBm and 3 dBm for the regenerator and the wavelength converterexperiments respectively. Input signal wavelengths were 1552.5nm. The fiber based regenerator and wavelength converter were setsuch that this wavelength was maintained in the loop. Input-signaland clock signal power into the SOA-DI were 6 and 8 dBm.

ResultsFig. 4(a) shows the measured Q value as a function of distance ofthe fiber-based 3R regenerator and the 3R all-optical wavelengthconverter [11][12]. Transmission over one million kilometers of

40 Gb/s Tx 3R Regenerator

EDFARamanPump

100 km TWRS

40 Gb/s Rx

x 4

EHS-DCM

Fig. 3 Loop Setup

Fig. 4 Quality of the signal versus transmission distances and (b) Bit-error rate ver-sus received power at one million km.

Page 17: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 17

fiber is shown for both schemes. The fiber based 3R regeneratorshows no sign of a degradation even after one million kilometer oftransmission. The wavelength converter shows a first sign of anerror floor that we attribute to polarization sensitivity issues of theSOA and the SA that occurred at the end of the experiment.

ConclusionsFull 3R regeneration and wavelength conversion have beenachieved exploiting both fiber nonlinear effects and SOA basednonlinearities. This way 40 Gb/s transmission over as much as onemillion kilometer of fiber has been demonstrated with no or onlylittle signal degradation.

References1. S. Fischer, M. Dulk, E. Gamper, W. Vogt, E. Gini, H. Melchior, W.

Hunziker, D. Nesset, A.D. Ellis ;El. Lett, vol. 35, no. 23, pp.2047-2049, Nov. 1999

2. Y. Ueno, S. Nakamura, K. Tajima; Photon. Technol. Lett., vol. 13,no. 5, pp. 469-471, May 2001

3. J. Leuthold, B. Mikkelsen, R.E. Behringer, G. Raybon, C.H.

Joyner, P.A. Besse; Photon.Technol. Lett., vol. 13, no. 8, pp. 975-877, Aug. 2001

4. T. Otani, T. Miyazaki, S. Yamamoto; J. of Lightwave Technol., vol.20., no. 2, pp. 195-200, Feb. 2002

5. J. Leuthold, R. Ryf, S. Chandrasekhar, D.T. Neilson, C.H. Joyner,C.R. Giles; Proc. Opt. Fiber Comm. Conf. OFC’2001, Anaheim,USA, March 2001, pd. 16

6 A.A.M. Saleh, IEEE/LEOS Summer Topical Meeting, p.36, July 19987. K. Suzuki, H. Kubota, A. Sahara, and M. Nakazawa; Electron.

Lett., vol. 34, no. 1, pp. 99-100, 1998.8. M. Nakazawa, E. Yamada, H. Kubota, K. Suzuki; El. Lett., vol. 17,

no. 14, pp. 1270, July 19919. G. Aubin, T. Montalant, J. Moulu, B. Nortier, F. Pirio, J.B.

Thomine; El. Lett., pp. 1163, July 99410. R.-J. Essiambre, b. Mikkelsen, G. Raybon; El. Lett., vol. 35, no. 18,

pp. 1576-157711. G. Raybon, Y. Su, J. Leuthold, R. Essiambre, T. Her, P. Steinvurzel,

K. Dreyer, K. Feder, C. Joergensen; Proc. Opt. Fiber Comm. Conf.OFC’2002, Anaheim, USA, March 2002, pd. FD10

12. J. Leuthold, Y. Su, G. Raybon, R. Essiambre; El. Lett., vol. 38, no. 16,pages XX, Aug. 2002

13. B. Sartorius; Proc. OFC’2001, p. MG7

All-Optical Clock Recovery for Signal Processing andRegeneration

B. SARTORIUS, S. BAUER, C. BORNHOLDT, O. BROX, M. MÖHRLE, H.-P. NOLTING and J. SLOVAK

Heinrich-Hertz-Institut für Nachrichtentechnik Berlin GmbH, Einsteinufer 37, D-10587 Berlin, GermanyTel +49 30 31002 508,FAX +49 30 31002 558 ([email protected])

Abstract Self-pulsating DFB lasers are developed for all-optical clock recov-ery. The system performance of these devices is demonstrated andtheir ultra high speed potential is pointed out.

1. IntroductionOptical signal processing offers the potential for ultra-high speedoperation (faster than electronics), for improved functionality(needed e.g. in packet switched networks), and for cost reduction(by avoiding the expensive opto-electronic conversions). An opticalclock is a key device needed for triggering the signal processingfunctions in digital optical communication systems /1/. In thispaper an overview on self-pulsating DFB lasers (section 2), theirapplication for all-optical clock recovery (section 3) and their speedpotential (section 4) is given.

2. Self-Pulsating DFB Lasers for Clock RecoverySelf-pulsating DFB lasers are compact devices – about 800 µmlong – comprising two DFB sections with detuned gratings and an

integrated phase tuning section (Fig. 1). The devices are driven bydc currents. However, they emit optical pulses at a high repetitionrate /2,3/. The self-pulsation frequency is electrically tunable. Fig.2 shows for example rf spectra of one device at various currents. Acontinuous tuning range from 6 to 46 GHz can be noticed /4/.This electrical tuning function is quite different to the characteris-tics of mode-locked lasers where the repetition frequency is definedby the length of the laser cavity.

©2002 Optical Society of AmericaOCIS codes: (070.4340) Non-linear optical signal processing, (230.1150) All-opticaldevices, (320.7090) Ultrafast lasers,

Fig. 1. 40 Gb/s input data signal and 40 GHz output clock signal of a self-pulsat-ing DFB laser.

Page 18: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

The application for clock recovery requires the synchronizationof the self-pulsation relative to the data signal. This synchronizationcan be achieved easily by injecting the optical data signal into theself-pulsating laser whose frequency is tuned electrically close to thedata rate (locking range about 100 MHz). Fig. 1 shows a 40 Gb/sPRBS data signal (upper trace) injected via a circulator into thedevice /4/. The lower trace shows the 40 GHz self-pulsation signalemitted from the device. The oscilloscope is triggered in parallel tothe data stream and thus only synchronized signals are displayed.The display of the self-pulsation trace thus verifies the locking ofthe optical clock. Analyzing the optical clock signal more in detailone finds sinus shaped pulses, stable in amplitude and with lowtiming jitter. The jitter derived from phase noise measurements isin the 300 fs range and can be – to some amount - lower than thejitter of a degraded input signal (regenerative function).

3. System Applications of Self-Pulsating Lasersfor Optical Clock Recocvery The most critical test of a device is its application in a regenerator.There the device has to operate with degraded input data signals thathave to be transformed to improved output signals. The 40 GHz all-optical clock has been applied in an all-optical 3R regenerator (3R:re-amplification, re-timing, re-shaping) based on the “synchronousmodulation” scheme /5/. This 3R regenerator has been evaluated inloop experiments. Error free transmission of 40 Gb/s PRBS signalsover more than 10.000 km was achieved /5/. Fig.3 shows the accord-ing eye diagrams which are clearly open. This result demonstrates thegood system performance and the regenerative function of the self-pulsating laser as 40 GHz all-optical clock recovery.

An important function needed in future packet switched net-works is the ultra-fast synchronization of clock recoveries to asyn-chronous data packets. Optical solution here can offer strong advan-tages compared to the electronics. In Fig. 4 the synchronizationtime of a 10 GHz self-pulsating laser (lower trace) to a packet of“1” bits (upper trace) is investigated. The non-synchronized noisysignal changes rapidly within 1 ns after injecting the data signal toa clear clock pulse trace. Only ten “1” bits are needed for the syn-chronization /6/. Analysis of the behavior at the end of the packetresulted in a stable clock signal for more than hundred “0” bits.The ultra fast locking and the long hold time make the self-pulsat-ing laser ideally suited for future packet switching applications.

4. Speed Potential of Self-Pulsating DFB Lasers The frequency for different configurations of self-pulsating DFBlasers has been analyzed by modeling. Fig. 5 shows a calculated

(VPI component maker) rf spectrum indicat-ing 160 GHz self-pulsation. Even higher fre-quencies seem to be achievable by designingthe DFB gratings properly and by usingsuited operating conditions. First devicestowards higher speed have already been fab-ricated. Fig. 6 shows a measured rf spectrumindicating 80 GHz self-pulsation /7/. Thelocking function to 80 Gb/s PRBS signalshas been verified, Fig. 7 shows the accordingstable 80 GHz pulse trace. The jitter perfor-mance of the clock has been evaluated byapplying the device in a transmission experi-ment. The phase noise jitter of the clockrecovered after 160 km transmission was less

frequency /GHz

rel.

el.i

nten

sity

/dB

m

frequency /GHz

rel.

el.i

nten

sity

/dB

m

Fig. 2. Electrical frequency tuning of a self-pulsating DFB laser

Fig. 3. Eye Diagrams of 40 Gb/s signal and demulti-pexed 10 Gb/s channel after 10.000 km transmission

Fig. 4. Synchronization time for a 10 GHz clock (lowertrace) to injected data packet with “1” bits (upper trace)

Fig.5. Modeling of 160 GHz SP. Fig.7. Pulse trace of 80 GHz clock.Fig.6. Measured 80 GHz self-pulsation.

18 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 19: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

then 300 fs and close to the jitter of the 80 Gb/s transmitter /8/.This indicates an excellent system performance of the high speedall-optical clock recovery based on the self-pulsating laser.

5. SummaryOptical clock recovery is a key function for signal processing infuture high speed and flexible all-optical networks. Self-pulsatingDFB lasers are developed for these applications. They are compactsemiconductor devices, easy to operate, and tuneable in frequencyvia the driving dc currents. Their good performance as opticalclock has been demonstrated in several system experiments.Important advantages are the ultra high speed potential exceeding

that of electronics and the fast locking function, needed for opera-tion in asynchronous packet switched networks.

6. References 1. B. Sartorius, OFC 2001, invited paper MG7, Anaheim, Cal, USA 2. M. Radziunas et al, IEEE J. QE 36, pp.1026, 20003. M. Möhrle et al, IEEE J. Sel. Topics in QE, 7, pp. 217, 20014. C. Bornholdt, et al, Electron. Lett. 36, pp.327, 2000.5. B. Sartorius et al, OFC 2000, paper PD 11, Baltimore, USA6. S. Bauer, et al., OFC 2000, paper TuF5, Baltimore, USA7. C. Bornholdt et al, ECOC 2001, paper Th.F.1.2, Amsterdam, NL8. C. Bornholdt et al, OFC 2002, paper TuN, Anaheim, Cal, USA

OCTOBER 2002 IEEE LEOS NEWSLETTER 19

System Perspective for All-Optical Switching

C. BINTJAS, N. PLEROS and H. AVRAMOPOULOS

Department of Electrical and Computer EngineeringNational Technical University of Athens, 15773 Zographou, Athens, Greece

It is true to say that the past year has been one of significantchanges in the telecommunications industry as a result of therecent market downturn. It is also true to say that before this,

the telecommunications industry had experienced an explosivegrowth for a good number of years. For wired networks this growthwas due to the unprecedented improvement in performance of thephotonic technologies as well as proof of their maturity that hasresulted in a spectacular decrease in transmission cost, which inturn has lead to the widespread adoption and construction of newfiber networks. At this point of the business cycle and in prepara-tion for the next crest of the wave, it is appropriate to reviewresearch lines that have been maturing over the years and whoseresults may be incorporated in next generation products. The pur-pose of this article is to argue that all-optical logic and all-opticalswitching techniques are such research themes.

All-optical switching, signal processing and more generally opti-cal computing has been the holy grail of researchers ever since theinvention of the laser. General purpose all-optical processing is stilla long way off and may never materialize as electronics possess andseem capable to maintain huge technical and cost advantages.However there are specialized applications in high data ratetelecommunications, where low complexity all-optical processingcircuits are ideally suitable to provide functionalities that electronicsolutions cannot, so that they may be of commercial value.

In order for all-optical switching techniques to become a seriousconsideration, they must simultaneously possess four properties: (a) aconsiderable speed advantage and otherwise ability to simplify thecircuit design, (b) switching energies that are similar to those of elec-tronics (c) capability for integration and of course (d) an applicationdomain where these advantages may be of use. It was relatively earlyrecognized that for the implementation of all-optical gates, interfero-metric arrangements offered speed advantages and Boolean logiccapability [1,2]. Figure 1 shows a generic interferometric arrange-ment drawn as a Mach-Zehnder interferometer, but which could infact be any type of interferometer as the Sagnac or the Michelson.

The interferometer consists of two separate optical paths where thephase of the optical field may be independently controlled in a non-linear optical medium by optical means. In the example of figure 1,assuming that on each of the two interferometer paths the phase ofthe optical field CLK may be changed by π depending on whetherone or both of the external controlling signals A and B are present,interference on the output coupler causes the result of the Booleanaddition XOR between A and B to be written on CLK. The rest ofthe Boolean operations may be implemented similarly.

Earlier research efforts used optical fiber as the nonlinear materi-al partly due to its ultrafast Kerr nonlinearity and mainly because itwas so much more easily available than for example optical semi-conductor devices. Very spectacular results in terms of speed havebeen achieved using the nonlinear optical fiber, Sagnac interferome-ter, including the demultiplexing of a 10 Gbps channel from a 640Gbps data stream in a 1.28 Tbps transmission experiment [3].Unfortunately the Kerr nonlinearity of optical fiber is weak, requir-ing the use of long pieces of fiber, making the devices hard to inte-grate and their switching energies high. In the example given abovethe interferometer used 450 m of fiber and required about 2 pJ ofswitching energy for 1 ps pulses [4], so that operation of the switch

Figure 1. Generic interferometric gate arrangement configured to perform BooleanXOR between signals A and B. The output is written on signal CLK.

Page 20: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

20 IEEE LEOS NEWSLETTER OCTOBER 2002

at the full data rate becomes energy-wise expensive. It was the realization that fast phase variations can also be

obtained in semiconductor optical amplifiers (SOAs) as a result of therelation between the refractive index and the carrier density, that hasmore recently lead to a whole new class of compact and very lowswitching energy devices. These can be broadly separated to eitherdiscrete SOA interferometric gates or integrated devices. The discretedevices are single optical path designs that are either of the dual armSagnac type [5-7], or single arm type [8,9] and include a single SOA.Integrated devices have been demonstrated by both monolithic inte-gration or hybrid integration of SOAs on Si-based planar lightwavecircuits and follow the Michelson or the Mach-Zehnder designs. Thetechnology status of semiconductor optical gates including integra-tion activities have been very thoroughly reviewed in reference [10].The basic nonlinearity that these devices exploit is resonant and forthis reason switching has been shown repeatedly to be possible withenergies of few fJ for pulses of few ps duration [11-13], in very com-pact devices. In this context demultiplexing from 336 Gbps to 10.5Gbps and wavelength conversion experiments have been shown [14],as well as more taxing experiments for the switching element at thefull data rate of 100 Gbps [15].

From the previous discussion it must be obvious that the recentadvances of all-optical switching techniques have addressed the issuesof speed, switching energy and capability for integration. It remainsto clearly identify application regimes for these devices where theydisplay superior suitability compared to alternative solutions.Applications related to ultra-high data rate single channel TDMtransmission such as demultiplexing were identified early. All-opticalregeneration was also identified as a key functionality that can beprovided by optical gates, by taking advantage of their amplitudeand timing jitter suppression properties [16-18]. It was however theoverwhelming penetration of WDM systems and the need for fastwavelength converters that has provided one of the main applicationareas [12, 19-20], because of the very steep transfer function and highextinction ratio that semiconductor interferometric devices can give.

Looking further ahead and in view of achieving better resource

use at the network level, some form ofpacket or burst switching is likely to beneeded. Currently only circuit opticalswitches are available in the market. All-optical packet switches exist only in labo-ratories and the demonstrations availabledo not offer the distinguishing features ofpacket switching, that is, on-demand useof bandwidth and the ability to switchbusty traffic without excessive packet loss.Proposals of how to build ‘true’, opticalpacket switches have been made in the past[21] and these rely on algorithms thatguarantee lossless routing through theswitch and optimal cost scaling of theswitch in terms of the burstiness of thedata. Such a switch consists of three mainunits: (a) the electronic control unit that isresponsible for running the routing algo-rithm from information obtained from thepacket headers, (b) a packet slot inter-changer unit at the input of the switchthat plays the role of buffer to avoid inter-

nal collisions and (c) a space switch for output routing. The routingalgorithm is still performed in the electronic domain, but the con-trol of the optical switching elements can be done all-optically, soas to avoid the complex electronic driver circuits for the opticalswitches and to improve the overall switch speed. To implementsuch design based on the all-optical control of the switching ele-ments, several functionalities are needed including packet synchro-nization [22] and parity checking [23]. Key optical circuits are alsoneeded as, for example, 2x2 exchange bypass switches [13] for thepacket slot interchanger or clock recovery units that may operate ona packet basis. Figure 2 shows a recently demonstrated all-opticalcircuit to acquire clock from packets [24]. The packet clock recov-ery unit uses a Fabry-Perot (FP) filter with a free spectral rangeequal to the data rate followed by a nonlinear, high speed, opticalgate. The role of the FP filter is to partially fill the ‘0s’ of theincoming packet and the optical gate employs its nonlinear transferfunction to equalize the amplitudes of the partially filled ‘1s’.Important features of this circuit are that it acquires clock withinvery few bits, the clock signal persists approximately for the dura-tion of the packet, it does not contain any high speed electronicsand does not require any external synchronization, so that it couldbe suitable to power an all-optically controlled packet switch.

In summary all-optical switching techniques have matured tech-nically, very significantly over the past few years and across allfronts. Advances have spanned from the material and device level,to the system and application level as well as performance and prac-ticality. Improvements were the result of effort in a number of labo-ratories across the world and these have brought all-optical switch-ing technology to the point of commercial exploitation. The nextfew years will call the commercial verdict on these efforts.

References1. A. Lattes, et. al., “An ultrafast all-optical gate,” IEEE J. Quantum

Electron., Vol. 19, pp. 1718-1723, 1983.2. G. Eichmann, Y. Li and R. R. Alfano, “Digital optical logic using a

Fab

ry-P

erot

Filt

er

localcw laser

Opt

ical

Gat

e

packetspackets packet clockpacket clock

Figure 2. All-optical clock recovery from short data packets at 10 Gbps. Note the short rise time of the packet clock andthat its duration is approximately equal to the original packet length.

Page 21: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 21

pulsed Sagnac interferometer switch,” Opt. Eng., vol. 25, p. 91, 1986.3. M. Nakazawa, T. Yamamoto and K. R. Tamura, “1.28 Tbit/s-70

km OTDM transmission using third- and fourth-order simultane-ous dispersion compensation with a phase modulator,” Electron.Lett. , Vol. 36, pp. 2027-2029, 2000.

4. T. Yamamoto, E. Yoshida and M. Nakazawa, “Ultrafast nonlinearoptical loop mirror for demultiplexing 640 Gbit/s TDM signals,”Electron. Lett., Vol. 34, 1013-1014, 1998.

5. J. P. Sokoloff, et. al., “A terahertz optical asymmetric demultiplexer(TOAD),” IEEE Photon. Technol. Lett., Vol. 5, pp. 787-790, 1993.

6. A. D. Ellis, et. al., “Ultra-high-speed OTDM networks using semi-conductor amplifier-based processing nodes,” J. of LightwaveTechnol., Vol. 13, pp. 761-770, 1995.

7. M. Eiselt, W. Pieper and H. G. Weber, “SLALOM: semiconductorlaser amplifier in a loop mirror,” J. of Lightwave Technol, Vol. 13,pp. 2099-2112, 1995.

8. K. Tajima, S. Nakamura, Y. Sugimoto, “Ultrafast polarization-dis-criminating Mach-Zehnder all-optical switch,” Appl. Phys. Lett.,Vol. 67, pp. 3709-3711, 1995.

9. N. S. Patel, K. A. Rauschenbach and K. L. Hall, “40-Gb/s demulti-plexing using an ultrafast nonlinear interferometer (UNI),” IEEEPhoton. Technol. Lett., Vol. 8, pp. 1695-1697, 1996.

10.K. E. Stubkjaer, “Semiconductor optical amplifier-based all-opticalgates for high-speed optical processing,” IEEE J. Select. TopicsQuant. Electron., 6, 1428-1435, 2000.

11.C. Bintjas, et. al, “20 Gb/s all-optical XOR with UNI gate, IEEEPhoton. Technol. Lett., Vol. 12, pp. 834-836, 2000.

12. Y. Ueno, et. al., “168-Gb/s OTDM wavelength conversion using anSMZ-type all-optical switch,” in Proc. ECOC 2000, pp. 13-14, 2000.

13.G. Theophilopoulos, et. al., “Optically addressable 2x2 exchange

bypass packet switch,” IEEE Photon. Technol. Lett., Vol. 14, pp.998-1000, 2002.

14. S. Nakamura, Y. Ueno and K. Tajima, “Ultrahigh-speed opticalsignal processing with Symmetric-Mach-Zehnder-type all-opticalswitches,” TuK4 in Technical Digest 2002 LEOS Summer TopicalMeetings, Mont Tremblant, Quebec, Canada, 2002.

15.K. L. Hall and K. A. Rauschenbach, “100-Gbit/s bitwise logic,”Opt. Lett., Vol. 23, pp. 1271-1273, 1998.

16.H. Avramopoulos and N. A Whitaker Jr., “Optical regenerationcircuit,” US patent No. 5369520, 1994.

17.B. Lavigne, et. al., “Full validation of an optical 3R regenerator at20 Gbit/s”, in Proc. OFC 2000, Vol. 3, pp. 93-95, 2000.

18.B. Sartorius, “3R All-optical signal regeneration”, in Proc. ECOC2001, Vol. 5, pp. 98-125, 2001.

19.C. Joergensen, et. al., “Wavelength conversion by optimized mono-lithic integrated Mach-Zehnder interferometer,” IEEE Photon.Technol. Lett., Vol. 8, pp. 521-523, 1996.

20.J. Leuthold, et. al., “All-optical wavelength conversion up to 100Gb/s with SOA delayed-interference configuration,” in Proc. ECOC2000, pp. 119-120, 2000.

21. E. A. Varvarigos, “The ‘Packing’ and ‘Scheduling packet’ switcharchitectures for almost all-optical lossless networks,” J. ofLightwave Technol. Vol. 16, pp. 1757-1767, 1998.

22.S. A Hamilton and B. S. Robinson, “40-Gb/s all-optical packet syn-chronization and address comparison for OTDM networks,” IEEEPhoton. Technol. Lett., Vol. 14, pp. 209-211, 2002.

23.J. Poustie, et. al., “All-optical parity checker with bit-differentialdelay,” Opt. Commun., Vol. 162, pp. 37-43, 1999.

24.C. Bintjas, et. al., “Clock Recovery Circuit for Optical Packets,” tobe published in IEEE Photon. Technol. Lett., September 2002.

Fast Processes in Semiconductor Optical Amplifiers: Theory and Experiment

JESPER MØRK, TOMMY W. BERG and SVEND BISCHOFF

COM, Technical University of Denmark, Building 345V, DK-2800 Kgs. Lyngby, DenmarkE-mail: [email protected], Tel: +45 4525 5765, Fax: +45 4593 6581

AbstractWe review the physical processes that are responsible for ultrafastgain and index dynamics in semiconductor optical amplifiers andwhich impact high-speed optical switching applications.

All-optical signal processing is expected to play an important rolein future high-capacity optical communication systems. Primaryincentives for this evolution are the larger data rates that can be han-dled by all-optical devices, as well as the cost-reduction and increasedflexibility that may be achieved by avoiding conversions between theoptical and the electronic domain. In order to be practical and com-petitive, all-optical switching devices should fulfil criteria similar tothose of electronics, i.e., be small and allow integration of differentfunctionalities, and have the potential for cheap mass-production.Presently, semiconductor optical amplifier (SOA) based devices areamong the primary contenders for integrated all-optical devices. The

large gain and large differential gain of SOAs allow switching withpower levels in the range of milliwats, and various functionalitieshave been demonstrated in a number of different schemes at speeds inexcess of 100 Gb/s, e.g. wavelength conversion at 168 Gb/s [1]. Inthis paper we will review and discuss the physical processes thatimpact the operation of SOAs at such high data-rates.

Common to the various schemes utilizing SOAs as the mainswitching element is the exploitation of saturation effects in theactive region of the waveguide. When an optical beam is injected inthe amplifier, the gain of the amplifier is saturated and, in conse-quence of the induced change of the carrier density in the activeregion, the refractive index of the waveguide is changed, cf. thequalitative illustration in Fig. 1. Both effects can be utilized for all-optical switching of a data signal [2]. In the simplest scheme ofcross-gain modulation (XGM), the gain change simply controls theamplitude of another (probe) beam transmitted through the wave-

Page 22: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

guide. The refractive index change is exploited in interferometricstructures, where the cross-phase modulation (XPM) imparted onthe probe beam in one arm of an interferometer can be converted toan amplitude change.

Following depletion by an injected optical pulse, the gain of theSOA will recover to its initial value by processes that restore thecarrier distribution in the active region. The gain and the index ofthe device in general depend on the detailed distribution of carriersin energy, but can be parameterized by the carrier density, if aquasi-equilibrium situation can be assumed. The recovery of thecarrier density in the active region is governed by the so-calledstimulated carrier lifetime, i.e., the lifetime decreases when theoptical power level is high. This can be utilized to increase thebandwidth of the SOA by ensuring a high optical power level,either through the data/control beams themselves and an amplifierwith high gain or by the application of a separate “holding” beam[3]. In this way gain recovery times as low as a few tens of picosec-onds have been realized and this is a main reason for the successfulhigh-speed demonstrations of XGM-based schemes. A related phe-nomenon is the “self-filtering” of the optical signal that takes placeupon propagation in the amplifier [4]. Thus, a modulation compo-nent, once generated, has a high-pass transmission characteristicthrough the amplifier, since higher frequencies saturate the amplifi-er less. This tends to equalize the lower efficiency with which the

high-frequency components can be generated by XGM, thusincreasing the bandwidth of the device [4].

Switching windows not limited in width by the carrier lifetimecan be realized in interferometers by utilizing that only the phasedifference between the two arms matters (differential scheme), thusenabling cancellation of a slow index recovery. Sub-picosecondswitching windows have been achieved for a periodic control signal[5], but since the magnitude of the index change depends on thesaturation of the amplifier, patterning effects anyway limit the bit-rate for data-controlled functionalities, such as wavelength conver-sion and regeneration.

If the characteristic time scale of the pump signal becomes com-parable to or shorter than the sub-picosecond scattering times onwhich a quasi-equilibrium carrier distribution is established in thedevice, the detailed evolution of the carrier distribution becomesimportant. The qualitative evolution of the carrier distribution fol-lowing excitation by a short optical pulse is illustrated in Fig. 2.Spectral holeburning (indicating the regime prior to establishmentof a Fermi-Dirac distribtion function for the carriers) and carrierheating effects thus significantly contribute to the gain and indexdynamics for short optical pulses and for high data rates.

– – –– – – –– – – ––

++ + +++ + + ++ + +

–– –– –––– –– –– –––– –– –– ––––

++++ ++ ++++++ ++ ++ ++++ ++ ++

I

P

NPin

I

P

NPin

Pin

Gain G

0

G0

Pin

Index n

n0

Pin

Gain G

0

G0

Pin

Index n

n0

Figure 1. Qualitative illustration of amplification and saturation in a semiconductoroptical amplifier.

Figure 2. Qualitative illustration of the evolution of the carrier distribution (electrondensity ρ vs. electron energy E ) in the active region of an SOA. Following stimulat-ed emission induced by a short optical pulse, the distribution recovers to equilibrium bycarrier-carrier scattering, establishing a Fermi-Dirac distribution function, carriertemperature relaxation and carrier injection.

Figure 3. (Left) Pump-probe measurements on a bulk SOA for different wavelengths; lower curve is for excitation in the gain region and upper curve is in the absorption region.(Right) Breakdown of various contributions to the response for excitation in the absorption region close to the the transparency point. TOT: total response. N: carrier density con-tribution. SHB: Spectral holeburning. CH: Carrier heating / temperature relaxation. TPA: Two-photon absorption.

22 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 23: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 23

The effects of carrier heating and spectral holeburning are wellknown to cause gain suppression and bandwidth limitations insemiconductor lasers. The temporal characteristics of the processesas well as the dependence on operation parameters, such as wave-length and injection current, can be characterized through hetero-dyne pump-probe measurements employing femtosecond opticalpulses [6]. A generalized rate equation model has been establishedthat well accounts for such measured gain and index dynamics inbulk and quantum well SOAs [7]. On the 100 fsec time scale,instantaneous (coherent) processes such as two-photon absorptionand optical Kerr effects are also found to significantly influence theresponse. Fig. 3 shows examples of pump-probe measurements fordifferent excitation wavelengths in a bulk InGaAsP optical amplifi-er. The figure also depicts an example of the breakdown of differentcontributions to the response for excitation on the short wavelengthside of the transparency point, where the device is absorbing.

The ultrafast contributions to the gain response strongly modifythe short pulse saturation properties of amplifiers. It has thus beenshown, experimentally [8] as well as theoretically [9], that for rela-tively low repetition rates and pulses shorter than about 10 ps, ultra-fast gain dynamics provide the dominant contribution to the satura-tion of the pulse gain. For optical switches, this has the implicationthat the carrier density change induced by a pump pulse with fixedenergy is reduced for shorter pulses [5,10]. On the other hand, thecontributions to the gain and index change from the intrabanddynamics increase for shorter pulses [5,6]. These opposing effectslead to switching windows with characteristics that depend on thepulsewidth and the operation point of the device, and careful opti-mization is necessary to exploit possible benefits of the fast processes.

Lasers with quantum dot (QD) active material have shownrecord-low threshold current densities and it is of interest to explorethe potential of QD SOAs for ultrafast processing [11].Experimental results presented in [12] indicate the possibility ofachieving very fast relaxation times in QD amplifiers. Fig. 4 depictsthe results of a comprehensive model for QD amplifiers, showinggood agreement with the experimental results [13]. Fast capturefrom a reservoir of carriers in the wetting layer is believed to beresponsible for the fast relaxation, pointing to the need of tailoringthe device design and operation to avoid long term depletion and

patterning effects. Recent results indicate the possibility of operationin a regime of such strong holeburning, to reduce patterning effectsdue to slow carrier recovery below the limits of bulk and QW SOAs.

In conclusion, the basic processes responsible for ultrafast gaindynamics in bulk and QW SOAs on a time scale of 100 fsec or longerseem to be rather well established. The influence of such fast processeshas been investigated for different specific switching applications, buta general picture valid for switching speeds in excess of 100 Gb/s hasnot yet been established. Results for quantum dot SOAs are beginningto appear. It seems that such devices may offer new regimes of opera-tion that enable high-speed all-optical signal processing.

References1. S. Nakamura, Y. Ueno and K. Tajima, “168-Gb/s all-optical wave-

length conversion with a symmetric-Mach-Zehnder-type switch”,IEEE Photon. Technol. Lett., vol. 13, pp. 1091-1093, Oct. 2001.

2. K. Stubkjaer, “Semiconductor optical amplifier-based all-opticalgates for high-speed optical processing”, IEEE J. Select. TopicsQuentum Electron., vol. 6, pp. 1428-1435, Nov. 2000.

3. R. J. Manning, A. D. Ellis, A. J. Poustie, and K. J. Blow,”Semiconductor laser amplifiers for ultrafast all-optical signal pro-cessing”, J. Opt. Soc. Am. B, vol. 14, pp. 3204-3216, Nov. 1997.

4. C. Joergensen et al., ”All-optical wavelength conversion at bit ratesabove 10 Gb/s using semiconductor optical amplifiers”,IEEE J. Select.Topics Quantum Electron., vol. 3, pp. 1168-1180, Oct. 1997.

5. S. Nakamura, Y. Ueno, and K. Tajima, “Femtosecond switchingwith semiconductor-optical-amplifier-based symmetric Mach-Zehnder-type all-optical switch”, App. Phys. Lett., vol 78, pp.3929-3931, June 2001.

6. K. L. Hall et al., “Nonlinearities in active media”, in Nonlinear Opticsin Semiconductors II. Semiconductors and Semimetals, vol. 59, Ed. byE. Garmire and A. Kost, Academic Press (San Diego), 1999.

7. J. Mørk and A. Mecozzi, “Theory of the ultrafast optical response ofactive semiconductor waveguides”, J. Opt. Soc. Am. B., vol. 13, pp.1803-1816, Aug. 1996.

8. P. Borri et al., ”Measurement and calculation of the criticalpulsewidth for gain saturation in semiconductor optical amplifiers”,Opt. Commun., vol. 164, pp. 51-55, June 1999.

0 2 4 6 8 10

0.6

0.7

0.8

0.9

1.0

(b)Occ

upa

tion

prob

abil

ity

Tim e delay [ps ]

GSESWL

0.0 0.5 1.0 1.5 2.0

0.6

0.7

0.8

0.9

1.0

Delay [n s]

-0.5 0.0 0.5 1.0 1.5 2.0-1.50

-1.25

-1.00

-0.75

-0.50

-0.25

0.00

(a)

ExperimentCalculated

Ga

inch

an

ge,

∆G[d

B]

Tim e delay [ps ]

Figure 4. (Left) Simulated and measured (from ref [12]) pump-probe response of a QD SOA. (Right) Variation of occupation probability for carriers in the ground state (GS),excited state (ES) and wetting layer (WL). The figure illustrates the sequential change of the carrier distributions following optical depletion of the ground state.

Page 24: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

9. A. Mecozzi and J. Mørk, “Saturation effects in non-degenerate four-wave mixing between short optical pulses in semiconductor laseramplifiers”, IEEE J. Sel.. Topics Quantum Electron., vol. 3, pp.1190-1207, 1997.

10.R. Schreieck et al., “Ultrafast switching dynamics of Mach-Zehnderinterferometer switches”, IEEE Photon. Technol. Lett., vol. 13, pp.603-605, June 2001.

11.M. Sugawara et al., “Quantum-dot semiconductor optical amplifiersfor high bit-rate signal processing over 40 Gbit/s”, Jpn. J. Appl.Phys., vol. 40, pp. L488-L491, May 2001.

12.P. Borri, W. Langbein, J. M. Hvam, F. Heinrichsdorff, M.-H. Mao,and D. Bimberg, “Ultrafast gain dynamics in InAs–InGaAs quan-tum-dot amplifiers,” IEEE Photon. Technol. Lett., vol. 12, pp.594–596, June 2000.

13. T. W. Berg, S. Bischoff, I. Magnusdottir, and J. Mørk, “Ultrafast GainRecovery and Modulation Limitations in Self-Assembled QuantumDot Devices”, Photon. Technol. Lett., vol. 13, pp. 541-543, June 2001.

14/T.W. Berg, A.V. Uskov, and J. Mørk, ”Ultrafast signal processing inquantum dot amplifiers through effective spectral holeburning”,Techn. Digest CLEO’02, paper CFH7, May 2002.

Hierarchical Hybrid OTDM/WDM Network Based on Fast Optical Signal Processing

HIDEYUKI SOTOBAYASHI(1)(2), WATARU CHUJO(1), and TAKESHI OZEKI(3)

(1) Communications Research Laboratory, IndependentAdministrative Institution, 4-2-1, Nukui-Kita, Koganei, Tokyo184-8795, Japan, Phone: +81-42-327-5320, Fax: +81-42-327-7035, E-mail: [email protected](2) Research Laboratory of Electronics, MassachusettsInstitute of Technology, Room 36-323, 77 MassachusettsAvenue, Cambridge, MA 02139, U.S.A., Phone: +1-617-253-8949, Fax: +1-617-253-9611, E-mail: [email protected](3) Dept. of Electrical and Electrical Engineering, SophiaUniversity, 7-1, Kioicho, Chiyodaku, Tokyo 102-8554, Japan,Phone: +81-3-3238-3330, Fax: +81-3-3238-3321, E-mail: [email protected]

I. IntroductionAs society is undergoing a fundamental change from an industrialsociety to an information society, the rapidly increasing demandfor bandwidth, driven by the Internet, has led to a paradigm shiftin the telecommunications industry to IP centric networks.Photonic networks are thus becoming to play important roles asglobal information infrastructures. Up to now, the primary genera-tion of photonic networks has thrust into the bandwidth expansionissue. However, the next generation of photonic networking is nowemerging, and the main issue is dynamic bandwidth provisioningon demand.

In this paper, we propose a hierarchical hybrid OTDM/WDMnetwork as the future core network. In order to achieve hierarchicalhybrid OTDM/WDM networks, ultrafast optical signal processingwould be key technologies. As enabling technologies for the hierar-chical hybrid OTDM/WDM network, we experimentally demon-strated three kinds of key functions such as wavelength-band gener-ation, all-optical OTDM/WDM mutual format conversion; and theinter-wavelength-band conversion.

II. Hierarchical hybrid OTDM/WDM networkThe hierarchical optical path architecture consisting of WDMbands and channels has been proposed to be suitable for large-scaleWDM backbone networks as shown in Fig.1(a) [1]. The groupingof wavelength optical paths, that is, forming a wavelength-band

WavelengthWDM channel

#2#1 #3 #m

Wavelength

Wavelength

WDM-to-OTDMOTDM-to-WDM

Conversion

Wavelength band #1

Group of WDM channels

Wavelength band #2 Wavelength band #3

Wavelength-band path

Wavelength band #1(OTDM)

WDMchannel path

Wavelength band #3(WDM)

Wavelength band #2(OTDM)

WavelengthWDM channel

#2#1 #3 #m

Wavelength

Wavelength

WDM-to-OTDMOTDM-to-WDM

Conversion

WDM-to-OTDMOTDM-to-WDM

Conversion

Wavelength band #1

Group of WDM channels

Wavelength band #2 Wavelength band #3

Wavelength-band path

Wavelength band #1(OTDM)

WDMchannel path

Wavelength band #3(WDM)

Wavelength band #2(OTDM)

High-levelnode

High-levelnode

High-levelnode

High-levelnode

L

L

L

LHigh-levelnode

L

L

LL High-

levelnode

L Low-level nodeWDM channel

Normal-priorityTraffic

Node Cut-through

Tera-bps OTDM

High-priorityTraffic

Band WDM

High-levelnode

Wavelength band

High-levelnode

High-levelnode

High-levelnode

High-levelnode

L

L

L

LHigh-levelnode

L

L

LL High-

levelnode

L Low-level nodeWDM channel L Low-level nodeWDM channel

Normal-priorityTraffic

Node Cut-through

Tera-bps OTDM

High-priorityTraffic

Band WDM

High-levelnode

Wavelength bandHigh-levelnode

Wavelength band

Error-Detection/All-opticalEqualization

BandMUX

BANDDEMUX

SPACESWITCH

Band-WavelengthConversion

Frequency Standardized

Wavelength-bandSignal

(> Tera-bit/s)

Error-Detection/All-opticalEqualization

Band signal-monitoring/All-opticalequalization

BandMUXBandMUX

BANDDEMUX

ATM-SW

Channel-to-

BandConvert

Channel-to-

BandConvert

Channel-to-

BandConvert

All-opticalFormatConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

All-opticalFormatConvert

IP Router ATM-Switch

Channel-to-

BandConvert

Channel

Convert

Channel-to-

BandConvert

Channel

Convert

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Band-to-

ChannelConvert

Band-to-

ChannelBand

Band-to-

ChannelConvert

Band-to-

ChannelBand

IP Router

SPACESWITCH

Band-WavelengthConversion

Frequency Standardized

Wavelength-bandSignal

(> Tera-bit/s)

Conversion

opticalFormat

All

Conversion

opticalFormat

All

Error-Detection/All-opticalEqualization

Error-Detection/All-opticalEqualization

BandMUX

BANDDEMUX

SPACESWITCH

Band-WavelengthConversion

Frequency Standardized

Wavelength-bandSignal

(> Tera-bit/s)

Error-Detection/All-opticalEqualization

Band signal-monitoring/All-opticalequalization

BandMUXBandMUX

BANDDEMUX

ATM-SW

Channel-to-

BandConvert

Channel-to-

BandConvert

Channel-to-

BandConvert

All-opticalFormatConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

All-opticalFormatConvert

IP Router ATM-Switch

Channel-to-

BandConvert

Channel

Convert

Channel-to-

BandConvert

Channel

Convert

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Band-to-

ChannelConvert

Band-to-

ChannelBand

Band-to-

ChannelConvert

Band-to-

ChannelBand

IP Router ATM-SW

Channel-to-

BandConvert

Channel-to-

BandConvert

Channel-to-

BandConvert

All-opticalFormatConvert

Channel-to-

BandConvert

Channel-to-

BandConvert

Channel-to-

BandConvert

All-opticalFormatConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

All-opticalFormatConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

Band-to-

ChannelConvert

All-opticalFormatConvert

IP Router ATM-Switch

Channel-to-

BandConvert

Channel

Convert

Channel-to-

BandConvert

Channel

Convert

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Channel-based OXCChannel-based OXC

Band-to-

ChannelConvert

Band-to-

ChannelBand

Band-to-

ChannelConvert

Band-to-

ChannelBand

IP Router

SPACESWITCH

Band-WavelengthConversion

Frequency Standardized

Wavelength-bandSignal

(> Tera-bit/s)

Conversion

opticalFormat

All

Conversion

opticalFormat

All

Conversion

opticalFormat

All

Conversion

opticalFormat

All

(a)

(b)

(c)

Fig. 1: (a) Layered structure of a hierarchical hybrid OTDM/WDM network. And(b) node cut-through and (c) node architecture in hierarchical hybrid OTDM/WDMnetwork.

24 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 25: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

path, is one way to reduce thecomplexity and size of theoptical path cross connect(OPXC). Hierarchical OPXCshave the following merits: (1)economical network by multi-protocol label switching node-cut-through at wavelength-band path level, (2) the modu-larity that means that relative-ly small-scale modular OPXCsmake up construct large-scalesystems, (3) the superior cross-talk specification of demulti-plexing, (4) the economicalinstallation according to thedemands, and (5) the reducednetwork management com-plexity through the use of thetwo-layered hierarchical net-work management [2,3]. Thenode-cut-through of the wave-length-band path is the keyconcept in the hierarchicalOTDM/WDM network archi-tecture as shown in Fig. 1(b).The wavelength-band is usedto transmit ultrahigh bit-rateOTDM signals instead of agroup of WDM channels. Oneof the aims of using OTDM inthe wavelength-band path ispossibility to monitor error-free transmission in the physi-cal layer, which is necessary tocut through the high-levelnodes. Figure 1(c) illustrates the node structure of the hierarchicalhybrid OTDM/WDM networks. The input wavelength-band sig-nals are demultiplexed by a band demultiplexing (DEMUX). Eachband should be first checked to guarantee error-free conditions. Thebands are then switched according to the cross-connect program.Node cut-through bands are switched to the output band-multi-plexing (MUX) after passing through the all-optical signal process-ing unit, which may transform them to the new wavelength of thebands for virtual wavelength-band path scheme. The other bandsare switched to drop-ports. The size of OPXC space switch isremarkably small compared to that of WDM channel OPXC. Thedropped bands are then assembled to WDM channels to feed chan-nel-based OPXCs. IP-routers and ATM-switches process the pack-et-based routing and switching and forward the packets to individ-ual channel-based OPXC. An important issue regarding the hierar-chical hybrid OTDM/WDM network nodes is how to realize fastoptical signal processing to construct efficient and economical nodestructures. The functions that must be provided by all-opticalmeans are (1) wideband signal generation, (2) multiplexing formatconversion, (3) wavelength-band conversion, (4) equalization ofwavelength and polarization-mode dispersions, (5) improvement ofspectral efficiency, and (6) error monitoring for node-cut-through.The first three key functions are described in the following sections.

III. Wavelength-band signal generationFigure 2(a) shows the operational principle of frequency standardizedsimultaneous wavelength-band generation [4]. 40 Gbit/s carrier-sup-pressed return-to-zero (CS-RZ) multiplications are performed bysupercontinuum (SC) generation [5], which is directly pumped by a40 Gbit/s CS-RZ signal and followed by spectrum slicing using anarrayed waveguide grating (AWG). The advantage of this method isthat the channel spacing is strictly fixed by the microwave mode-locking frequency of the source laser [6]. Figure 2(b) shows theexperimental setup. 40 Gbit/s CS-RZ format signals were generatedby using time-delayed optical multiplexer with phase shifter. TheSC spectrum was sliced and recombined by AWGs with 100 GHzchannel spacing. Tellurite-based erbium-doped fiber amplifiers (T-EDFAs) were used for amplifying the continuous signal band in theC- and L-bands. Figure 3(a) shows the optical spectra of SC at theoutput of SC fiber, signals before transmission, and after transmis-sion. The power difference in all WDM channels after 80 km trans-mission was about 7 dB. Figures 3(b) and 3(c) respectively show themeasured optical spectra after WDM demultiplexing of ch. 1(1535.04 nm) and ch. 81 (1600.60 nm). These figures clearly showthat optical carriers were suppressed even after 80 km transmission.Figures 3(d) and 3(e) respectively show the measured eye diagramsfor WDM ch. 1 and ch. 81. The eye diagrams for each of the WDM

OCTOBER 2002 IEEE LEOS NEWSLETTER 25

MLLD1530.33 nm

SCFLN-

MOD

OTDM MUXwith

PHSE-SHIFTER

10 GHz

PPG

SMF

10 Gbit/s 40 Gbit/s CS-MUX

Chirpcompensation

SMF29.4 kmRDF

10.6 km

SMF28.4 kmRDF

11.6 kmAWGWDM

DEMUX

SMZDEMUX

Opt.Pre-amp.Receiver

BERT

40 Gbit/s-to-10 Gbit/s DEMUX

EDFA

Tellurite basedEDFA

AWGWDMMUX

81 channel:100 GHzspacing

Tellurite basedEDFA

40 GHz

t

[ 0 π 0 π ]

Optical Carrier Phase

25 ps100 ps

f

CarrierFrequency

f

f

Spectrum slicing

SC generation

t

[ 0 π 0 π ]

Optical Carrier Phase

25 ps100 ps

(b)

(a)

Fig. 2: (a) Principle of simultaneous wavelength-band generation of frequency standardized WDM signal in CS-RZ format using SCgeneration and spectrum slicing. And (b) experimental setup for wavelength-band signal generation and transmission.

Page 26: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

26 IEEE LEOS NEWSLETTER OCTOBER 2002

1540 1560 1580 1600Wavelength [nm]

Pow

er [a

.u.]

(5 d

B/d

iv) 40 Gbit/s CS–RZ induced SC

before transmission

before WDM DEMUX

WDM ch.1 WDM ch. 81

1535 1536

Pow

er [a

.u.]

(5 d

B/d

iv)

Wavelength [nm]

40 GHz

Opticalcarrier

frequency

1600 1601

Pow

er [a

.u.]

(5 d

B/d

iv)

Wavelength [nm]

40 GHz

Opticalcarrier

frequency

(b)ch. 1

(c)ch. 81

(d)ch. 1

20 ps /div

(e)ch. 81

20 ps /div

(a)

1540 1560 1580 1600Wavelength [nm]

Pow

er [a

.u.]

(5 d

B/d

iv) 40 Gbit/s CS–RZ induced SC

before transmission

before WDM DEMUX

WDM ch.1 WDM ch. 81

1535 1536

Pow

er [a

.u.]

(5 d

B/d

iv)

Wavelength [nm]

40 GHz

Opticalcarrier

frequency

1600 1601

Pow

er [a

.u.]

(5 d

B/d

iv)

Wavelength [nm]

40 GHz

Opticalcarrier

frequency

(b)ch. 1

(c)ch. 81

(d)ch. 1

20 ps /div

(e)ch. 81

20 ps /div

(a)

Fig. 3: (a) Measured optical spectra of SC at the output of SCF (upper trace), signals before transmission (middle trace), and after transmission (lower trace). Measured opti-cal spectra after 80 km transmission and WDM DEMUX of (b) ch. 1 and (c) ch. 81. And measured eye diagrams of (d) ch.1and (e) ch. 81.

SC pulseP

t

SA Time-window

P

t

Time-gated SC pulse

WDM

OTDM

P

t

SA pump pulse

0

4 3 2 1

SC pulseP

t

SC pulseP

t

λ

SA Time-window

P

t

λ

Time-gated SC pulse

WDM

OTDM

P

t

SA pump pulse

λ0

λ4 λ3 λ2 λ1

SA Time-window

P

t

Time-gated pulse trains

WDM

OTDM

P

t

SA pump pulse

t

P High speed pulse train

4 3 2 1

0

0

SA Time-window

P

t

Time-gated pulse trains

WDM

OTDM

P

t

λ

SA pump pulse

t

P High speed pulse train

λ4 λ3 λ2 λ1

λ0

λ0

(a) (b)

SCF

BERT

LN-MOD

AWGWDM

DEMUX

1 2 3 4

TDMMUX

10-to-40Gbit/s

EDFA

MLLD0 SA

Fiber forpulse

stretching

10 Gbit/s

AWGWDMMUX

EDFA 3 nm

3 nm

SAEDFA

TDMMUX

EDFA3 nm

TDMDEMUX

40-to-10 Gbit/s

10-to-40 GHz

Opt.Pre-amp.Receiver

PPG

WDM-to-OTDMconversion

4 x 10 Gbit/s WDM

40 Gbit/s OTDM

40 Gbit/s OTDM

1.5 ps10 GHz,

1548.3 nm

Time gate Window: 10 ps

24 dBm

OTDM-to-WDMconversion

SCF

BERT

LN-MOD

AWGWDM

DEMUX

1 2 3 4λ 1 λ 2 λ 3 λ 4

TDMMUX

10-to-40Gbit/s

EDFA

MLLDλ 0 SA

Fiber forpulse

stretching

10 Gbit/s

AWGWDMMUX

EDFA 3 nm

3 nm

SAEDFA

TDMMUX

EDFA3 nm

TDMDEMUX

40-to-10 Gbit/s

10-to-40 GHz

Opt.Pre-amp.Receiver

PPG

WDM-to-OTDMconversion

4 x 10 Gbit/s WDM

40 Gbit/s OTDM

40 Gbit/s OTDM

1.5 ps10 GHz,

1548.3 nm

Time gate Window: 10 ps

24 dBm

OTDM-to-WDMconversion

(c)

Fig. 4: Operational principle of the photonic conversion of (a) OTDM-to-WDM, and (b) WDM-to-OTDM by using optical time-gating. And (c) experimental setup for for-mat conversions.

Page 27: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

channels have a good eye opening with BERs less than 10-9. It isthus concluded that a frequency standardized 3.24 Tbit/s (81 WDMx 40 Gbit/s) wavelength-band signal in CS-RZ format can be simul-taneously generated by using a single SC source [4].

VI. Multiplexing format conversionConversion of 40 Gbit/s OTDM-to-4 x 10 Gbit/s WDM is done byoptical time-gating of a highly linear chirped SC pulse as shown inFig. 4(a) [7]. By shifting the time-position of the optical time-gat-ing, the center wavelength of the time-gated SC pulse can be tuned.When 40 Gbit/s OTDM signals are used to control the time-gatingON/OFF window, the 10 GHz repetition rate SC pulses are convert-ed to 4 x 10 Gbit/s WDM signals. Fig. 4(b) shows the principleoperation of WDM-to-OTDM conversion [7]. Time-aligned 4 x 10Gbit/s WDM signals are used for controlling the ON/OFF gate win-dow of the 40 GHz repetition rate pulse trains, resulting in a con-version of WDM signals to 40 Gbit/s OTDM signals.

Figure 4(c) shows the experimental setup. A 10 GHz SC pulsewith a highly chirped, wide pulse width, and rectangular shapewas optically time-gated in a semiconductor saturable absorber(SA) [8], which was pumped by 40 Gbit/s OTDM signal.Conversion to WDM was done by WDM demultiplexing using a4 ch. AWG (λ11544.1 nm - λ4:1552.5 nm) because the centerwavelengths of time-gated SC pulses depend on the time positionof time-gating. For WDM-to-OTDM conversion, 40 GHz pulsetrains of λ0 are optical time gated in the SA pumped by 4 x 10Gbit/s WDM data. The experimentally measured mutual formatconversions are shown in Figs. 5(a)-5(e). These figures show that40 Gbit/s OTDM of λ0 is converted to 10 Gbit/s WDM channelsof λ1-λ4 and reconverted into 40 Gbit/s OTDM of λ0 with cleareye opening and BERs less than 10-9 [7].

V. Wavelength-band conversionFor wavelength-band path routing, inter-wavelength-band conver-sions of 640 Gbit/s OTDM signals both from C-to-L-band and L-to-

C-band followed by 640-to-10 Gbit/s OTDM DEMUX have beenexperimentally demonstrated [9]. Highly nonlinear dispersion-shiftedfibers (HNL-DSFs), which have high nonlinear coefficients and lowdispersion slopes [10], were used as wavelength converters and ultra-fast nonlinear optical loop mirror (NOLM) demultiplexings [11].

Fig. 6(a) shows the experimental setup for C-to-L-wavelength-band conversion of a 640 Gbit/s OTDM signal. A 10 GHz, 700 fspulse train at 1550 nm was split into two, one was used for a 640Gbit/s C-band OTDM signal and the other was used for the controlpulse for demultiplexing of the wavelength-band converted L-band640 Gbit/s OTDM signal. After polarization optimization, the sig-nal and pump were coupled in HNL-DSF #1 in order to generatefour wave mixing (FWM). The pump wavelength λp was set to1565.0 nm. A wavelength-band converted 640 Gbit/s signal at1580 nm was extracted by using an L-band EDFA, and then opticaldemultiplexed to 10 Gbit/s in a NOLM controlled by a 10 GHz,700 fs C-band pulse train. Fig. 6(b) shows the experimental setupof L-to-C wavelength-band conversion. A 10 GHz, 780 fs pulsetrain at 1580 nm was split into two, one was used for a 640 Gbit/sL-band OTDM signal and the other was for the control pulse indemultiplexing of the wavelength-band converted C-band 640Gbit/s OTDM signal. A wavelength-band converted 640 Gbit/ssignal at 1550 nm by FWM was extracted using a C-band EDFAand demultiplexed. Figures 7(a)-7(d) show the experimental results.In both wavelength-band conversions, 640 Gbit/s OTDM signalswere wavelength-band converted without significant pulse widthbroadening with BERs less than 10-9 [9].

IV. ConclusionKey technologies for hierarchical hybrid OTDM/WDM network,that is, 3.24 Tbit/s frequency standardized continuous C- and L-wavelength-band signal generation, 40 Gbit/s OTDM-WDMmutual multiplexing format conversions, and 640 Gbit/s OTDMwavelength-band conversions have been developed. The proposedschemes based upon fast optical signal processing will become cru-cial in the future hierarchical hybrid OTDM/WDM network.

OCTOBER 2002 IEEE LEOS NEWSLETTER 27

Fig. 5: Experimental results of (a) optical spectrum and (b) eye diagram of input 40 Gbit/s OTDM, (c) optical spectrum and (d) eye diagram of converted 4x 10 Gbit/s WDM,and (e) optical spectrum and (f) eye diagram of reconverted 40 Gbit/s OTDM.

Page 28: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

References1. K. Harada, K. Shimizu, T. Kudou, and T. Ozeki, “Hierarchical

optical path cross-connect systems for large scale WDM networks,”in Proc. Optical Fiber Communication Conference (OFC’99),WM55, pp. 356-358 (1999).

2. I. Nishioka, R. Izumailov, Y. Suemura, Y. Maeno, and S. Araki,“Aggregation of Dynamically Varying Demands in HierarchicalOptical Networks,” Technical Report of IEICE, PNI2001-14 (2001).

3. J. M. Simmos, “Hierarchical Restoration of Backbone Networks”,in Proc. Optical Fiber Communication Conference (OFC ’99), TuL2(1999).

4. H. Sotobayashi, A. Konishi, W. Chujo, and T. Ozeki, “Wavelength-band generation and transmission of 3.24 Tbit/s (81 WDM x 40Gbit/s) carrier suppressed RZ format using a single supercontinu-um source for frequency standardization,” to be published in OSAJ. Opt. Soc. Am. B, vol. 19, no. 11, 2002.

5. H. Sotobayashi and K. Kitayama, “325 nm bandwidth supercontin-uum generation at 10 Gbit/s using dispersion-flattened and non-decreasing normal dispersion fibre with pulse compression tech-nique,” IEE Electron. Lett., vol. 34, no. 13, pp.1336-1337 (1998).

6. H. Takara, T. Ohara, K. Mori, K. Sato, E. Yamada, Y. Inoue, T.Shibata, M. Abe, T. Morioka, and K.-I. Sato, “More than 1000channel optical frequency chain generation from single supercontin-

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

(a) (b)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

(a) (b)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

(a) (b)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

0 10 20 30Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalC–to–L–wavelength–band converted

640–to–10 Gbit/s DEMUX signalL–band (1580 nm)

(a) (b)

0 10 2 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

(c) (d)

0 10 2 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

(c) (d)

0 10 2 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

0 10 20 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

(c) (d)

0 10 20 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

0 10 2 0Time [ps]

Po

wer

(a.

u.)

640 Gbit/s OTDM signalL–to–C–wavelength–band converted

640–to–10 Gbit/s DEMUX signalC–band (1550 nm)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

1540 1560 1580 1600Wavelength (nm)

Po

wer

(10

dB

/div

: a.u

.)

(c) (d)

0

Fig. 7: Experimentally measured 640 Gbit/s OTDM signal after (a),(b) C-to-Lwavelength-band conversion, and (c),(d) L-to-C wavelength-band conversion.

HNL-DSF#2

MLLD1550.0 nm

10 GHz 10 Gbit/s

PLCTDM MUX

640 Gbit/sC-bandEDFA

P.C.

BERT

Signal

Opt.Pre-amp.Receiver

DDF

CW-LD1565.0 nm

Tellurite-basedEDFA

P.C.

LN-Mod. HNL-

DSF#1

L-bandEDFA

L-bandEDFA

Pump

1 nmFilter

1580nm

C-bandEDFA

C-bandEDFA

Pumprejection

filter

C-band, 700 fs pulse

C-to-L-wavelength-band conversion

640-to-10 Gbit/s OTDM DEMUX

1.5 ps 24 dBm

HNL-DSF#2

MLLD1550.0 nm

10 GHz 10 Gbit/s

PLCTDM MUX

640 Gbit/sC-bandEDFA

P.C.

BERT

Signal

Opt.Pre-amp.Receiver

DDF

CW-LD1565.0 nm

Tellurite-basedEDFA

P.C.

LN-Mod. HNL-

DSF#1

L-bandEDFA

L-bandEDFA

Pump

1 nmFilter

1580nm

C-bandEDFA

C-bandEDFA

Pumprejection

filter

C-band, 700 fs pulse

C-to-L-wavelength-band conversion

640-to-10 Gbit/s OTDM DEMUX

1.5 ps 24 dBm

MLLD1560.0 nm

10 GHz

ATT.

10 Gbit/s

PLCTDM MUX

640 Gbit/sL-bandEDFA

P.C.

SignalDDF

CW-LD1565.0 nm

Tellurite-basedEDFA

P.C.

LN-Mod. HNL-

DSF#1

C-bandEDFA

Pump

C-bandEDFA

SCF

3 nmFilter

1580 nm

HNL-DSF#2

BERTOpt.

Pre-amp.Receiver

C-bandEDFA

1 nmFilter

1550 nm

L-bandEDFA

Pumprejection

filterL-band, 780 fs pulse

L-to-C-wavelength-band conversion

640-to-10 Gbit/s OTDM DEMUX

MLLD1560.0 nm

10 GHz

ATT.

10 Gbit/s

PLCTDM MUX

640 Gbit/sL-bandEDFA

P.C.

SignalDDF

CW-LD1565.0 nm

Tellurite-basedEDFA

P.C.

LN-Mod. HNL-

DSF#1

C-bandEDFA

Pump

C-bandEDFA

SCFSCF

3 nmFilter

1580 nm

HNL-DSF#2

BERTOpt.

Pre-amp.Receiver

C-bandEDFA

1 nmFilter

1550 nm

L-bandEDFA

Pumprejection

filterL-band, 780 fs pulse

L-to-C-wavelength-band conversion

640-to-10 Gbit/s OTDM DEMUX

(a)

(b)

Fig. 6: Experimental setup for (a) C-to-L wavelength-band conversion and (b) L-to-C wavelength-band conversion.

28 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 29: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 29

uum source with 12.5 GHz channel spacing,” IEE Electron. Lett.,vol. 36, no. 25, pp. 2089-2090 (2000).

7. H. Sotobayashi, W. Chujo, and T. Ozeki, “Bi-directional photonicconversion between 4x10 Gbit/s OTDM and WDM by opticaltime-gating wavelength interchange,” Optical FiberCommunication Conference (OFC 2001), WM5, pp. WM5-1-WM5-3, Anaheim, March 2001.

8. H. Kurita, I. Ogura and H. Yokoyama, “Ultrafast all-optical signalprocessing with mode-locked semiconductor lasers,” IEICE Trans.on Electron. vol. E81-C, no. 2, pp. 129-139 (1998).

9. H. Sotobayashi, W. Chujo, and T. Ozeki, “Inter-wavelength-band

conversions and demultiplexings of 640 Gbit/s OTDM signals,”Optical Fiber Communication Conference (OFC 2002), WM2,Anaheim, March 2002.

10. O. Aso, S. Arai, T. Yagi, M. Tadakuma, Y. Suzuki and S. Namiki,“Broadband wavelength conversion using a short high-nonlinearitynon-polarization-maintaining fiber,” in Proc. European Conference onOptical Communication (ECOC’99), Th B1-5 (1999).

11. H. Sotobayashi, C. Sawaguchi, Y. Koyamada, and W. Chujo,“Ultrafast walk-off free nonlinear optical loop mirror by a simplifiedconfiguration for 320 Gbit/s TDM signal demultiplexing,” to be pub-lished in OSA Opt. Lett., vol. 27, no. 15 (2002).

Optical Header Recognition Using Fiber Bragg GratingCorrelators

JOHN E. MCGEEHAN, MICHELLE C. HAUER, and ALAN E. WILLNER

AbstractThere is increasing interest in performing many key networkingfunctions in the optical domain. Optical header recognition is onesuch key function that may enable rapid reading of optical packetsin the future all-optically-switched network core. Many of theseoptical header recognition functions are enabled through the use offiber Bragg grating-based optical correlators. A brief backgroundon optical header recognition and FBG correlators is presented,and two fiber Bragg grating-based optical header recognitionmethods are explained in detail: an optical bypass for an Internetrouter, and a multi-wavelength header recognition system usingsampled fiber Bragg grating optical correlators.

I. IntroductionFuture high-performance optical networks may require optical datapackets to be rapidly routed by all-optical switches. For efficient andhigh-throughput switching, the header bits within each packet mustbe recognized and acted upon quickly. Although electronic tech-niques are available for reading individual bits at lower bit rates (i.e.,<10 Gbit/s), there is a requirement for all-optical methods that canreadily “read” a series of header bits on-the-fly and either assist orreplace the electronics in making ultra-fast routing decisions.

In conventional Internet routers, packets are steered toward theirdestinations by interrogating their 32-bit destination addresses andmatching them to entries in a large routing table — a time consum-ing process as modern look-up tables can contain upwards of 100,000entries. Current optical signal processing technologies, however, donot scale well, making it impractical to build a 100,000-entry all-optical lookup table. However, in the network core, where routershave only four to eight output ports and only a few hops may be nec-essary to reach a core access point, it may be possible to determine apacket’s destination port by looking at only a subset of the bits in thedestination address. Recent research has shown that looking at a sub-set of a core routing table with as few as 100 entries can successfullyroute up to 90% of network traffic [1], and by looking at which sub-sets of those headers are required to identify an output port, a man-ageable optical lookup table can be constructed.

One common way to perform this optical bit-pattern recognitionis through the use of a time-domain optical correlator to match aseries of bits to an optical lookup table [2]. The purpose of a correla-tor is to compare an incoming signal with one that is “stored” in thecorrelator. At an appropriate sample time, a maximum autocorrela-tion peak will be produced if the input signal is an exact match tothe stored one. Numerous optical correlator designs have been experi-mentally demonstrated, including those using fiber Bragg gratings(FBGs) [3], FBGs in complement with novel periodically-poled lithi-um niobate (PPLN) waveguide wavelength shifters [4], erbium-doped fiber [5], free-space holography [6], and deposited metallicfiber mirrors with hardwired optical delays [7]. FBG-based correla-tors have some specific advantages over other designs, includingincreased tunability (the reflectivity of FBGs can be tuned via heating[8] or stretching [9] the gratings), ease of manufacture, and relativelylow cost. Using FBG-based correlators, a number of innovative opti-cal header recognition systems have been proposed and demonstrated:i) an optical bypass for an internet router using a single-gratingtopology that can scale to high bit rates, and ii) a reconfigurableWDM header recognition (WDM-HR) system using sampled FBGs.

II. Fiber Bragg Grating-based Optical CorrelatorsA set of fiber Bragg gratings can be used to construct an effectivefiber-based optical correlator [3]. An FBG is fabricated by creating aperiodic variation in the fiber’s index of refraction for a few millime-ters to a centimeter of length along the fiber core [10]. A conven-tional FBG acts as a reflective, wavelength-selective filter. A nicefeature of FBG filters is that the reflection spectrum can be adjustedby a few nanometers via heating or stretching of the grating. Thereflectivity of the grating is nearly 100% at the center of the reflec-tion spectrum and falls off quickly outside the grating bandwidth.FBG correlators can be constructed by writing many gratings in suc-cession with center-to-center spacings equal to 1

2 of a bit time sothat the round-trip time between gratings corresponds to a 1-bitdelay. To program a “1” bit in a correlation sequence, a grating istuned to be partially or wholly reflective, and to program a “0” bit, agrating is tuned away so that it does not reflect. The reflectivity of

Page 30: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

each grating must be chosen such that the pulses reflecting off eachgrating have equal power. In practice this can be determined byrepeatedly sending a single pulse into the FBG array and adjustingthe grating to equalize the powers of the time-delayed output pulseson an oscilloscope (each “1” bit grating will reflect the input pulseat different times, creating multiple pulses at the output). Theresulting grating array will produce a correlation output correspond-ing to the input sequence correlated with the programmed correla-tion sequence, and a threshold detector can sample the central lobeof this sequence to determine the presence of a pattern match, asshown in Fig. 1 for a correlation sequence of “1101.”

In general, optical correlators cannot uniquely identify a givenbit sequence. As gratings representing a “0” bit are tuned away anddo not reflect, those bits are actually “don’t-care” bits and a thresh-old detector will determine a match at the sample time regardlessof whether the “don’t-care” bits are a “0” or a “1” and may result ina false positive when a “1” bit is present where a “0” bit is desired.To uniquely recognize any of the 2N possible N-bit sequences, asecond correlator is used, configured in complement to the first onethat produces a “match” signal when zero power is present at thesample time. In this “zeros” correlator, a “0” bit is programmed bytuning a grating to be reflective, and a grating corresponding to a“1” bit is tuned away and does not reflect. As only an identical pat-tern will simultaneously provide a high threshold in the “ones” cor-relator and zero power in the “zeros” correlator at the sample time,by combining the outputs of the standard “ones” correlator withthis additional “zeros” correlator using an AND gate a matchingsignal is produced that can drive an optical switch.

III. Optical Bypass for an Internet RouterA true all-optical router would need to be capable of 32-bitlookups into 100,000-entry tables at ≥40 Gbit/s. Such capabilitiesare beyond current optical technologies, but some recent develop-ments hint at the feasibility of a partial solution. It may be feasibleto build an “optical bypass” to vastly accelerate a conventionalrouter [11]. As much as 90% of the incoming traffic may be rout-

ed by this bypass, and remaining traffic, requiring more compli-cated processing, can be handled by a conventional electronicrouter. This bypass is made possible through the use of FBG opti-cal correlators as header subset recognition devices. A diagram ofthis optically-boosted router is shown in Fig. 2. A new correlatordesign in which an FBG array is constructed from a single uniformFBG using separate, electrically-tunable thin-film micro-heaters isused to construct this optical bypass, and is shown in Fig. 3(a).The grating is fabricated out-of-band and the heaters are used totune individual parts of the grating in-band. Thus, a multi-bitcorrelator is constructed from a single long grating (Fig. 3(b)). Asthese thin-film micro-heaters can be made to lithographic preci-sion, millimeter grating spacings are easily achieved using thistechnology, making it scalable to ≥10 Gbit/s.

To configure the “ones” and “zeros” correlators to recognize a bitpattern of “xx1x01x0” (a header where a subset of only 4 out of the8 bits is required to determine the output port) in a system run-ning at 10 Gbit/s, the third grating in the “ones” correlator istuned to partially reflect, and the sixth grating is tuned for fullreflection. Likewise, the fifth grating in the “zeros” correlator istuned for partial reflection, and the eighth grating in the “zeros”correlator is tuned for full reflection. Thermal tuning is accom-plished by applying a voltage across each micro-heater. A packet-rate timing signal is used to trigger a set of low-speed decision cir-cuits to sample the correlation outputs at the proper time. The twooutputs are sent through an AND gate that provides a high signalwhen there is a match, and a low signal otherwise. The match/no-match signal that results can be used to control an optical switchand route matched packets to an appropriate output port, as shownin Fig. 4.

IV. Reconfigurable Multi-wavelength OpticalCorrelators Using Sampled Fiber Bragg GratingsThe above method for optical header subset recognition acts on asingle wavelength-division-multiplexed (WDM) channel, thusrequiring N complete modules in order to recognize the headers

30 IEEE LEOS NEWSLETTER OCTOBER 2002

Fig. 1. An FBG optical correlator is an array of FBG mirrors with tunable reflectivity R. The correlation sequence is programmed as “1101”, with the first, second, and fourthgratings tuned to be reflective, corresponding to “1” bits, and the third grating is tuned away so it does not reflect, corresponding to a “0” bit. The input sequence, also “1101”,reflects off each grating, and the result can be sampled electronically and a pattern match signal generated.

Page 31: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 31

on N different WDM channels. A correlation module that canallow for reconfigurable optical correlation of multiple WDMchannels simultaneously may use significantly fewer components[12]. Using a set of discrete sampled fiber Bragg gratings, a multi-wavelength FBG correlator can be constructed. A sampled FBG isan FBG that has a superstructure written on top of the grating forwhich the Fourier transform produces a reflective time delay thatis replicated at equal wavelength spacings [13]. When this type ofFBG is stretched or heated, the entire reflection spectrum shifts,causing the reflectivity at each wavelength to experience the samevariation, as shown in Fig. 4. Thus, the correlation sequence can bereconfigured for all incoming channels simultaneously.

While a sampled FBG correlator can be constructed in a mannersimilar to a standard FBG correlator, the spacing requirements canbe problematic. As it takes approximately ~1 cm of fiber to provide12 a bit time delay at 10 Gbit/s, the center-to-center spacingbetween gratings must equal this length. Due to their complexity,it can be difficult to manufacture sampled FBGs shorter than 1 cmthat have the high reflectivity required to produce good correlationresults. This problem can be solved by interleaving the gratingsbetween multiple fiber branches and using a splitter prior to thecorrelator. This decreases the spacing requirement by a factor equalto the number of branches. While these limitations on sampledFBG systems reduce the scalability of the architecture, a recent

Data B

Router

Electronic Processing

Data A

Data B1234

O/E

E/O

Auxilliary Port for optical “misses”

5

Data ASwitch

Tap

EDFA

Correlator 1

Correlator 2

Correlator 3

Correlator M

Port 1 Match?

# Correlators > # Ports, (M > N)

Port 2 Match?

Port 2 Match?

Port 4 Match?

Optical Correlator Module

Correlator Control & Timing

1xM

Correlator 4 Port 3 Match?

Data BData B

Router

Electronic Processing

Data AData A

Data BData B1234

O/E

E/O

Auxilliary Port for optical “misses”

5

Data AData ASwitch

Tap

EDFA

Correlator 1

Correlator 2

Correlator 3

Correlator M

Port 1 Match?

# Correlators > # Ports, (M > N)

Port 2 Match?

Port 2 Match?

Port 4 Match?

Optical Correlator Module

Correlator Control & Timing

1xM

Correlator 4 Port 3 Match?

Cache

Algorithm:Routing Table Optimization

& Correlator Control

~100,000 Entries 100

Entries 10 Correlators

Cache

Algorithm:Routing Table Optimization

& Correlator Control

~100,000 Entries 100

Entries 10 Correlators

Fig. 2. Conceptual diagram of a core router with N (in this case, 4) output ports assisted by a bank of M optical FBG-based correlators that are dynamically configured by asoftware algorithm which reduces the routing table to a few popular entries. After grouping these entries by output port, each correlator is configured to match to a particular groupby recognizing only a subset of the bits in the destination address. Packets that fail to match any of the correlators are routed via conventional electronics.

Fig. 3. (a) A set of micro-heaters are deposited on a single long out-of-band FBG, creating small, individually-tunable gratings out of the larger FBG. Each is spaced by 1 cmto correspond to 1

2 a bit time at 10 Gbit/s. (b) The optical spectrum of the FBG in transmission showing one of the smaller gratings tuned .6 nm away from the central reflec-tion spectrum via thermal tuning

Page 32: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

32 IEEE LEOS NEWSLETTER OCTOBER 2002

report details a new sampled FBG structure that can reduce thelength of sampled FBGs while maintaining high reflectivity,enabling the application of this correlation technique to high-bit-rate systems [14].

An interleaved sampled FBG correlator was constructed andconfigured to match a header of “1010.” In the “ones” correlator,the first grating was tuned via stretching to partially reflect, andthe third tuned for full reflection, while in the “zeros” correlator,the second grating was tuned to partially reflect, and the fourthgrating tuned for maximum reflectivity. Packets on two WDMchannels, each at 10 Gbit/s, were correlated by the gratings andsent to individual packet-rate decision circuits as described in theabove section. The resulting match/no-match signals for each chan-nel were used to control an optical switch fabric where header-matched packets were routed to one output port, and all otherpackets to an alternate output.

V. ConclusionWhile optical correlators currently see limited commercial applica-tion, the frontier of all-optical networking is rapidly approaching,aided by the ever-increasing demand to transmit more bandwidthover the network core. For applications such as header recognition,technologies that can implement huge arrays or banks of correla-tors to efficiently test incoming signals against all possible bitsequences may be needed. However, it may also be possible (withinthe network core) to use small, presently achievable banks of opti-cal correlators to recognize a small subset of the 32 bits in thestandard IP header, and thus route a significant percentage of thetraffic optically, using electronics only when all available opticallookup techniques fail to produce a matching condition. In addi-tion, optical header recognition techniques that can act simultane-ously on multiple wavelength channels in a WDM system mayresult in significant savings. While reaching these goals presents asignificant engineering challenge, progress is being made, andoptical header recognition using optical correlators offers greatpotential in making the all-optical switched network of the futurea reality.

References1. J. Bannister, J. Touch, P. Kamath, and A. Patel, “An optical booster

for internet routers,” Proc. Eighth Int’l Conf. on High PerformanceComputing, Dec. 2001

2. N. Wada and K.-I. Kitayama, “Photonic IP routing using opticalcodes: 10 Gbit/s optical packet transfer experiment,” Conf. onOptical Fiber Communications (OFC) 2000, paper WM51, pp. 362-364, 2000.

3. D. B. Hunter and R. A. Minasian, “Programmable high-speed opti-cal code recognition using fibre Bragg grating arrays,” ElectronicsLett., vol. 35, no. 5, 412-414, 1999.

4. D. Gurkan, M. C. Hauer, A. B. Sahin, Z. Pan, A. E. Willner, K. R.Parameswaran, and M. M. Fejer, “Demonstration of multi-wave-length all-optical header recognition using a PPLN and optical cor-relators,” European Conf. on Optical Communications (ECOC) 2001,paper We.B.25 (Invited), Amsterdam, 2001.

5. J. S. Wey, J. Goldhar, D. L. Butler, and G. L. Burdge,“Investigation of dynamic gratings in erbium-doped fiber for opti-cal bit pattern recognition,” Proc. Conf. on Lasers and Electro-Optics(CLEO) 1997, paper CThW1, 443-444, 1997.

6. J. Widjaja, N. Wada, Y. Ishii, and W. Chijo, “Photonic packetaddress processor using holographic correlator,” Electronics Lett, vol.37, no. 11, 703-704, 2001.

7. J.-D. Shin, M.-Y. Jeon, and C.-S. Kang, “Fiber-optic matched filterswith metal films deposited on fiber delay-line ends for optical pack-et address detection,” IEEE Photonic Tech. Lett., vol. 8, no. 7, 941-943, 1996.

8. Y.-W. Song, Z. Pan, D. Starodubov, V. Grubsky, E. Salik, S. A.Havstad, Y. Xie, A. E. Willner, and J. Feinberg, “All-fiber WDMoptical crossconnect using ultrastrong widely tunable FBGs,” IEEEPhotonic Tech. Lett., vol. 13, no. 10, pp. 1103-1105, 2001.

9. E. R. Lyons and H. P. Lee, “An efficient electrically-tunable etchedcladding fiber Bragg grating filter tested under vacuum,” IEEEPhotonic Tech. Lett., vol. 13, no. 5, pp. 484-486, 2001.

10.R. Kashyap, Fiber Bragg Gratings, Academic Press, California,1999.

11.M. C. Hauer, J. McGeehan, J. Touch, P. Kamath, J. Bannister, E. R.

Fig. 4. Experimental routing results showing the successful recognition and switchingof packets containing the header subset pattern “xx1x01x0”. The single matched pack-et is routed to port C, non-matched packets are routed to port D.

Fig. 5. The spectrum of a sampled fiber Bragg grating showing multiple reflectionpeaks. As the grating is tuned via stretching, the spectrum shifts and the reflectivity ofeach channel decreases. By applying the right amount of stretching, the gratings can betuned such that there is almost no reflection in each channel, enabling tuning of a cor-relation sequence.

Page 33: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 33

Lyons, C. H. Lin, A. A. Au, H. P. Lee, D. S. Starodubov, and A. E.Willner, “Dynamically reconfigurable all-optical correlators to sup-port ultra-fast internet routing,” Conf. on Optical FiberCommunications (OFC) 2002, paper WM7, 268-270, 2002.

12.J. McGeehan, M. C. Hauer, A. B. Sahin, and A. E. Willner,“Reconfigurable multi-wavelength optical correlator for header-based switching and routing,” Conf. on Optical FiberCommunications (OFC) 2002, paper WM4, 264-266, 2002.

13.M. Ibsen, M. K. Durkin, M. J. Cole, and R. I. Laming, “Sinc-sam-pled fiber Bragg gratings for identical multiple wavelength opera-tion,” IEEE Photonic Tech. Lett., vol. 10, no. 6, pp. 842-844, 1998.

14.N. Yusuki and Y. Shinji, “Realization of various superstructure fiberBragg gratings for DWDM systems using multiple-phase-shifttechnique,” Conf. on Optical Fiber Communications (OFC) 2002,paper TuQ3, 110-111, 2002.

OTDM Packet Networking Devices at 100 Gbit/s and Beyond

S. A. HAMILTON and B. S. ROBINSON

Massachusetts Institute of Technology, Lincoln Laboratory, 244 Wood St., Lexington, MA 02420Telephone: (781) 981-2904, Fax: (781) 981-4129, Email: [email protected]

I. IntroductionThe enormous growth of Internet traffic in recentyears has led to a dramatic increase in demand fordata transmission capacity. Previously, networkbackbone capacity has been dominated by voicetraffic. Today, data traffic capacity requirementsare beginning to surpass those for voice. In antici-pation of this added demand, the telecommunica-tions industry is actively working to increase thetransmission capacity available in backbone net-works. Recent advancements in backbone net-work capacity can be credited, in large part, todevelopments in high-speed transmission andswitching systems. Of particular importance inthe transmission arena has been the emergence ofwavelength division multiplexing (WDM) tech-nology, which supports multiple simultaneouswavelength channels on a single fiber. Anotherpromising transmission technology, still in itsinfancy commercially, is optical time divisionmultiplexing (OTDM), which supports ultrafast data rates (i.e. > 100 Gbit/s) on a single wavelength channel. Both WDM andOTDM systems with point-to-point transmission capacity exceed-ing a Tbit/s have been demonstrated recently in the laboratory [1]-[3]. Such high data rates will place severe demands on networkelement processing speeds. Additionally, optical-electronic-optical(O/E/O) conversion in electronic routers will result in congestionand reduced efficiency in WDM and OTDM optical networks.

II. Ultrafast Optical Packet SwitchOptical routers in future packet-switched networks will allowactive packet routing within and between wavelengths whilesimultaneously providing an optical path that is transparent toboth data format and transmission rate. All-optical packet-switched networks are expected to provide many advantages com-pared to earlier architectures. First, low-level network functionali-ty, such as routing, is distributed in the optical network core,while high-level functionality, like traffic grooming and protocol

translation which requires a large amount of slow processing, ispushed to the network edges. Because payload data is transparentin an optical packet-switched network, each switch can remain inany given state for an arbitrary amount of time and a wide range ofservices from best-effort datagram, to virtual circuit connections,to guaranteed bandwidth via a dedicated lightpath can be providedsimultaneously. Optical packet-switched networks may also pro-vide scalability and flexibility to efficiently handle time-dependentnetwork traffic patterns.

Here, we present a possible architecture for a 16x16 opticalpacket switch based on ultrafast serial processing achievable withultrafast Boolean optical logic gates. A generalized optical packetswitch [4],[5] consists of an input interface, header processor,switching matrix, and output interface as shown in Figure 1. Theinput interface provides synchronization for packet delineation andphase alignment at the switching matrix input. The header proces-sor reads the packet header and provides the signals used to config-ure the switching matrix. The switching matrix routes the packetto the desired port and resolves port contention. The output inter-

Figure 1 – Generalized 16x16 optical packet switch architecture consists of an input interface, header proces-sor, switching matrix, and output matrix.

Page 34: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

face performs packet regeneration and synchronization to providehigh-quality transmission to the next node in the network.

The packet architecture is an important feature for ultrafastoptical packet-switched networks. We will consider a fixed-lengthoptical packet architecture demonstrated previously by our groupfor use in a slotted OTDM network testbed that is capable of oper-ating at line rates of 112.5 Gbit/s [6]. In this demonstration, a sin-gle optical packet is 100 ns long and fully-loaded with synchronous12.5 GHz network clock pulses interleaved between 100 Gbit/srate multiple-bit address header and payload data. The first clock

pulse in each packet is removed to pro-vide a packet marker. In order to achievepacket switching at rates beyond 100Gbit/s, the packet marker is used to syn-chronize the packet at the network inter-face and the address header is opticallyprocessed for routing information used toforward the packet through the all-opticalswitching matrix.

At the packet switch input and outputinterfaces, both fine and coarse packetsynchronization are required. Fine syn-chronization is achieved at each inputport using a single all-optical logic gate.In this case, the optical NAND gate per-forms a bitwise comparison between theglobal clock pulses in the network packetand clock pulses generated locally in theinput interface. Because the optical logicgate is biased for NAND operation, a sin-gle synchronization pulse is switched outof the gate only when the missing packetmarker pulse is present in the networkpacket. The packet marker can be used forboth fine and coarse packet synchroniza-tion at the packet switch input interface

as shown in Figure 2. In order to achieve fine synchronization, thesynchronization pulse is used in conjunction with an optoelectronicdithering phase-locked loop to maintain the temporal overlap of theoptical pulses in the locally generated clock train with the globalclock pulses in the network packet. Coarse packet synchronizationcan be achieved if the synchronization pulse arrival time for eachinput port is measured via a counter and then compensated using avariable optical delay line on each input port. As shown in Figure2, each packet can be temporally aligned prior to entering theswitch matrix after fine- and coarse-synchronization at the packet

OpticalClock

BitSynch.

PacketSynch.

BitSynch.

PacketSynch.

BitSynch.

PacketSynch.

T2

TN

Payload Header

Payload Header

Payload Header

T2

TN

Payload Header

Payload Header

Payload Header

2

1

InputInte

r

.

face OpticalClock

BitSynch.

PacketSynch.

BitSynch.

PacketSynch.

BitSynch.

PacketSynch.

T2

TN

Payload HeaderPayload Header

Payload HeaderPayload Header

Payload HeaderPayload Header

T2

TN

Payload HeaderPayload Header

Payload HeaderPayload Header

Payload HeaderPayload Header16

2

1

InputInterface ∆

.

.

Figure 2 – The input interface consists of an optical clock source, bit synchronizer, counter, and packet synchronizer. Fineand coarse packet synchronization is achieved with an ultrafast optical NAND gate contained in the bit synchronizer.

Figure 3 – Header processor for an ultrafast 2x2 switch in the Banyan matrix. Optical routing information (Ci) is generated for the ith column of the Banyan matrix by com-paring the ith header bit in packet 1 (A1

i) and packet 2 (A2i) with the empty/full check bit for packet 1 (E1

i).

34 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 35: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 35

switch input interface. In our proposed 16x16 packet switch archi-tecture, synchronization at the input and output interfaces can beachieved with 16 optical logic gates.

The 16x16 switching matrix consists of an input queued Banyanconfiguration. An NxN Banyan network implemented using inter-connected 2x2 switches offers advantages of relatively low switchcount (0.5Nlog2N) and straightforward address header processing(see for example [7]). Another key advantage that can be realized ifthe Banyan matrix is composed of ultrafast optical 2x2 switches [8]is picosecond reconfiguration times. In this example, our 16x16Banyan switch matrix will require 32 interconnected 2x2 opticalswitches and each path length must be closely matched in order tomaintain packet synchronization. Monolithic integration is onepotentially viable means of meeting this precise interconnectionrequirement.

A requirement of the input buffer queue is that it provides com-plete packet contention resolution prior to the Banyan switchmatrix. Most optical buffers proposed today require fiber delay linesand storage time variability is provided using the temporal, spatial,or wavelength domain (see for example [9]-[11]). Although head-of-line blocking in an input queued packet switch has been shownto limit throughput to 58.6% [12], it is possible to use virtual out-put queueing via more complex input buffer queue architectures toachieve 100% contention resolution [13],[14]. If port contention iscompletely resolved prior to the Banyan switch matrix, this condi-tion will significantly reduce the logic density required in the head-er processing block of our proposed optical packet switch.

Header processing can be implemented with ultrafast opticallogic gates to minimize propagation delays through the packetswitch and allow serial processing at line rates beyond the abilitiesof available electronics. Because current optical logic gate technolo-gy cannot easily provide the complex processing and high level ofintegration achieved by the electronics industry, special care mustbe taken to simplify the header processing procedure in an opticalpacket switch. A single optical AND gate has been demonstratedfor processing multiple-bit address headers at >100 Gbit/s linerates [6]. In this experiment, the optical AND gate performed abitwise comparison of 4-bit network and local receiver addressheaders with no restrictions on the allowable keyword space. The16x16 Banyan switch proposed in this paper consists of 4 columnseach composed of 8 ultrafast 2x2 switches and each packet headercontains 4 address bits. In a Banyan matrix, only the ith addressmust be processed to correctly route the packet through the ith

switch column as shown in Figure 3. The final piece of routinginformation required is obtained after performing an empty/fullcheck on each packet. If packet contention is completely resolvedby the input buffer queue, header processing at each optical 2x2 inthe Banyan switch matrix is simplified to just two optical logicgates and a power combiner as shown in Figure 3. In our proposed16x16 packet switch architecture, header processing can beachieved with 64 optical logic gates.

III. ConclusionAll-optical packet switching will be required to eliminate theO/E/O conversion bottlenecks that limit capacity growth in today’snetworks. One potential means of implementing a 16x16 all-opticalpacket switch is to exploit the ultrafast serial processing provided byoptical logic gates and memory buffers. We have presented a 16x16

optical packet switch architecture that is capable of achieving packetsynchronization, header processing, and routing at rates in excess of100 Gbit/s and requires approximately 100 optical logic gates.

IV. References1. K. Fukuchi, T. Kasamatsu, M. Morie, R. Ohhira, T. Ito, K. Sekiya,

K. Ogasahara, and T. Ono, “10.92-Tb/s (273 x 40 Gb/s) triple-band/ultra-dense WDM optical-repeatered transmission experi-ment,” in Optical Fiber Communications Conference, Anaheim, CA,PD24 (2001).

2. S. Bigo, Y. Frignac,G. Charlet, W. Idler, S. Borne, H. Gross, R.Dischler, W. Poehlmann, P. Tran, C. Simmoneau, D. Bayart, G.Veith, A. Jourdan, J.-P. Hamaide, “10.2 Tbit/s (256x42.7 Gbit/sPDM/WDM) transmission over 100 km TeraLight fiber with 1.28bits/s/Hz spectral efficiency,” in Optical Fiber CommunicationsConference, Anaheim, CA, PD25 (2001).

3. M. Nakazawa, T. Yamamoto, and K. R. Tamura, “1.28 Tbit/s-70km OTDM transmission using third- and fourth-order simultane-ous dispersion compensation with a phase modulator,” Electron.Lett., 36, 2027-2029 (2000).

4. D. J. Blumenthal, P. R. Prucnal, and J. R. Sauer, “Photonic packetswitches: architectures and experimental implementations,” Proc.IEEE, 82, 1650-1667 (1994).

5. S. L Danielsen, P. B. Hansen, and K. E. Stubkjaer, “Wavelengthconversion in optical packet switching,” J. Lightwave Technol., 16,2095-2108 (1998).

6. S. A. Hamilton and B. S. Robinson, “100 Gbit/s synchronous all-optical time-division multiplexing multi-access network testbed,”in Optical Fiber Communications Conference, Anaheim, CA, ThEE3(2002).

7. P. R. Prucnal, “Optically processed self-routing, synchronization,and contention resolution for 1-D and 2-D photonic switchingapplications,” IEEE J. Quantum Electron., 29, 600-612 (1993).

8. G. Theophilopoulos, M. Kalyvas, C. Bintjas, N. Pleros, K.Yiannopoulos, A. Stavdas, H. Avramopoulos, and G. Guekos,“Optically addressable 2x2 exchange/bypass packet switch,” IEEEPhoton. Technol. Lett., 14, 998-1000 (2002).

9. J. D. Moores, W. S. Wong, and K. L. Hall, “50-Gbit/s optical pulsestorage ring using novel rational-harmonic modulation,” Opt. Lett.,20, 2547-2549 (1995).

10.D. K. Hunter, M. C. Chia, and I. Andonovic, “Buffering in opticalpacket switches,” J. Lightwave Technol., 16, 2081-2094 (1998).

11.D. K. Hunter, et. al., “WASPNET: a wavelength switched packetnetwork,” IEEE Commun. Mag., 120-129 (1999).

12.M. J. Karol, M. G. Hluchyj, S. P. Morgan, “Input versus outputqueueing on a space-division packet switch,” IEEE Trans. Commun..,35, 1347-1356 (1987).

13. N. McKeown, A. Mekkittikul, V. Anantharam, and J. Walrand,“Achieving 100% throughput in an input-queued switch,” IEEETrans. Commun.., 47, 1260-1267 (1999).

14.G. Thomas, “Multi-channel input-queueing for high throughputswitches,” Electron. Lett.., 33, 184-185 (1997).

This work was sponsored by the Defense Advanced ResearchProjects Agency (DARPA) under Air Force contract #F19628-00-C-002. Opinions, interpretations, recommendations, and conclu-sions are those of the author and are not necessarily endorsed by theUnited States Government.

Page 36: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

36 IEEE LEOS NEWSLETTER OCTOBER 2002

Applications of Fiber-Based Optical Parametric Amplifiers

J. HANSRYD and P.A. ANDREKSON*

CENiX Inc., 7360 Windsor Drive, Allentown, PA 18106Tel: (610) 336-5772, FAX: (610) 336-5888, email: [email protected]

* Also with Chalmers University of Technology, Göteborg, Sweden

Fundamentals and applications featuring the multi-functionalproperties of the fiber based optical parametric amplifier(FOPA), such as all-optical signal sampling, optical time divi-

sion demultiplexing, pulse generation, and wavelength conversionare reviewed and discussed.

The amplification in a FOPA is a non-linear phenomenonexploiting the light induced modulation of the fiber refractiveindex [1]. It is notable that the fiber only serves as a passive medi-um in contrast to, for instance, Raman, Brillouin or rare-earthdoped fiber amplification. Parametric gain in optical fibers is oftenreferred to as a third order parametric process as it is relying on thethird order susceptibility χ(3) of the material. This is in differentfrom the well-investigated parametric processes exploiting the χ(2)

non-linearity in optical materials such as crystals LiNbO3 and KTP. Figure 1 shows a typical FOPA configuration. A pump is com-

bined with the signal into the non-linear medium. During theamplification process photons are transferred from the pump waveto the signal wave and to an additional wave, the idler. Due to therelatively high pump powers, stimulated Brillouin scattering (SBS)suppression techniques are commonly needed for the pump wave.

The FOPA response time is limited solely by the relaxation timeof the bound electrons in the material. This constant is <10 fs thusenabling ultra fast signal processing applications. It can be shownthat when a specific phase condition is maintained throughout thefiber, the FOPA will offer maximum gain. Due to the phase match-ing condition, the FOPA does not only offer conventional incoher-ent or phase-insensitive gain but also the important potential ofcoherent or phase-sensitive parametric amplification. The phase-sensitive amplifier amplifies noise components with the same phase

as the signal while attenuating components with the oppositephase. This property has demonstrated many potential signal-pro-cessing applications e.g. pulse reshaping, quantum noise suppres-sion and for soliton communication systems: dispersive wave andsoliton-soliton interaction reduction [1]. It also has the potential foramplification with 0 dB noise figure. For the phase-insensitive (PI)FOPA, one or two pump light waves with arbitrary phase(s) willinteract with a signal light wave. A fourth light wave, the idler, isformed with a phase such that it satisfies the phase matching condi-tion. As the idler adjusts its phase to the injected light waves, thePI FOPA lacks the phase-sensitive features but requirements for itsimplementation are substantially relaxed.

Except for the obvious feature of providing gain, an FOPA has sev-eral properties important for all-optical signal processing applications:

1. High differential gain.2. Optional wavelength conversion.3. Wide optical bandwidth.4. Instantaneous gain response. 5. Operation centered at an arbitrary wavelength.

In the special case when the phase matching condition is main-tained over the whole fiber length, L, the parametric gain, G, maybe written in dB units as [2]

G = Pp L S − 6[dB]Here γ is the fiber nonlinear parameter and S = 8.4γ are the differ-ential gain in [dB/W/km]. A typical γ value for highly non-linearfiber (HNLF) is 10-15 W-1km-1 resulting in a differential gainbetween 80-130 dB/W/km. Such a steep nonlinear differentialgain may advantageously be used in applications such as highpower return-to-zero (RZ) pulse generation [3], O-TDM demulti-

Signal

Pump

1547

Idler

PumpSignal

1563 1579 nm

Pump laser at 1563 nm

SBS suppression(phase modulator)

EDFA

Input signal

Fiber

Figure 1 Schematic of a FOPA setup. An additional lightwave, the idler is created in the gain process.

Page 37: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 37

Figure 2 (a) Schematic of FOPA used for RZ-pulse generation. (b) Left: measured optical spectrum. Righ top: Generated 40 GHz pulses measured with streak camera (3.5 psresolution). Right low: Generated 40 GHz pulses measured with oscillscope (5 min. persistence).

0 3.3 6.6 9.9 13.2

Time (ps)

300 Gbit/s3.3 ps

0 5 10 15 20 25Time (ps)

310 km

(a)

(b)

A/D converterwith computer

Sampling pulse pump source

Input signalIdler

Repetition rate, f0

Repetition rate, f0/N+∆f

FOPA

Figure 3 (a) Schematic of a FOPA used as an optical sampling device. (b, left) Dispersion managed soliton measured after 310 km propagation. (b, right) 300 Gbit/s sampledeye diagram.

Page 38: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

38 IEEE LEOS NEWSLETTER OCTOBER 2002

plexing with inherent gain [2,4] and all-optical sampling [5].Figure 2 (a) shows an overview of the schematics for a RZ pulsegenerator based on a FOPA. Due to the steep gain slope it is suffi-cient to use a sinusoidally modulated pump wave while still gener-ating very narrow pulses. Figure 2 (b) shows measured results for asinusoidally modulated 40 GHz pump wave. The left part showsthe optical spectrum for the generated pulses, the upper right fig-ure shows the generated pulses measured by a streak camera with3.5 ps temporal resolution. The right lower figure shows the samepulses measured over long time persistence (5 min.) with a highbandwidth (45 GHz) optical photo-detector and a 50 GHz sam-pling oscilloscope.

Common for the above-mentioned applications is that they areusing the FOPA as a generic building block to create new func-tions. Figure 3 (a) shows a schematic for an optical sampling devicebased on a FOPA [5]. The bandwidth of the sampling device issolely limited by the pulse width of the sampling pulses, in thiscase 1.6 ps (~600 GHz). The control pump pulses are now asyn-chronous and “scanning” over the measured pulses. The high sam-pling bandwidth makes it possible to observe details previouslyimpossible to detect with a conventional electric sampling oscillo-scope. Figure 3 (b, left) shows a dispersion-managed soliton sam-pled after 310 km propagation through a dispersion managed trans-mission link. Oscillations in the pulse tail show the effect of insuffi-cient β3 compensation. The right figure shows a sampled 300Gbit/s optical eye diagram.

A key component for FOPAs operating over a wide bandwidth isthe availability of HNLF. The wide operating bandwidth of FOPAsis a direct consequence of the HNLF offering a high differentialgain and typically having a low dispersion slope (~0.03 ps/nm2km),thus offering good phase matching over a wide bandwidth. Thisfeature may be used for arbitrary and transparent wavelength con-version [6, 7]. Arbitrary wavelength conversion was until recentlyonly considered a realistic alternative in semiconductor opticalamplifiers due to the inherent narrowband operation of third orderparametric processes in conventional optical fibers.

One problem for single pumped FOPAs is the high gain ripple.Recent numerical and experimental results demonstrate that it ispossible to achieve a flat exponential gain spectrum over a widewavelength range by using a dual pumping scheme [8].

The fast saturation time of the FOPA gain have been demon-strated usable e.g. for all-optical limiting amplifiers [1,4,9].

Since the first fiber-based parametric amplifier experiments provid-ing net CW gain were only conducted a few years ago [10], there isreason to believe that substantial progress may be made in the future.The further development of “holey fibers” together with new fiber

materials may enhance the performance and practical implementationfor these amplifiers by offering even higher non-linearities [11].

References 1. J. Hansryd, P. A. Andrekson, M. Westlund, J. Li, and P.-O.

Hedekvist,” Fiber-based optical parametric amplifiers and theirapplications”, IEEE Journal on Sel. Top. in Quantum Electronics,vol.8, no.3 , pp. 506 -520, May/June 2002.

2. J. Hansryd and P. A. Andrekson, “O-TDM demultiplexer with 40-dB gain based on a fiber optical parametric amplifier”, IEEEPhoton. Technol. Lett., vol.13, no.7 , pp. 732 -734, July 2001.

3. J. Hansryd and P. A. Andrekson, “Wavelength tunable 40GHzpulse source based on fibre optical parametric amplifier”, Electron.Lett., vol.37, no.9 , pp. 584 -585, April 2001.

4. P.-O. Hedekvist, M. Karlsson, and P. A. Andrekson, “Fiber four-wavemixing demultiplexing with inherent parametric amplification”, J. Lightwave Technol, vol.15, no.11, pp.2051-2058, Nov. 1997.

5. J. Li, J. Hansryd, P. -O. Hedekvist, P. A. Andrekson, and S. N.Knudsen, “300 Gbit/s eye-diagram measurement by optical sam-pling using fiber based parametric amplification”, Optical FiberCommunication Conference 2001, vol. 4, postdeadline paper PD 31,Anaheim, USA, 2001

6. M.C. Ho, K. Uesaka, M. Marhic, Y. Akasaka, L. G. Kazovsky,“200-nm-bandwidth fiber optical amplifier combining parametricand Raman gain”, J. Lightwave Technol., vol.19, no.7, pp. 977-981,July 2001.

7. M. Westlund, J. Hansryd, P. A. Andrekson, P.A., and S. NKnudsen, “Transparent wavelength conversion in fiber with 24 nmpump tuning range”, Electron. Lett., vol.38, no.2, pp. 85 –86, Jan.2002

8. C. J. McKinstrie, S. Radic, and A. R. Chraplyvy, “Parametricamplifiers driven by two pump waves”, IEEE Journal on Sel. Top. inQuantum Electronics, vol.8, no.3 pp. 538 -547, May/June 2002.

9. A. Takada and W. Imajuku, “Amplitude noise suppression usinghigh gain phase sensitive amplifier as a limiting amplifier”,Electron. Lett., vol.32, no.7, pp. 677-679, March 1996.

10.J.Hansryd, P.A. Andrekson,. “Broadband CW pumped fiber opticalparametric amplifier with 49 dB gain and wavelength conversionefficiency”, Optical Fiber Communication Conference 2002, vol. 4,postdeadline paper PD-6, Baltimore, USA, 2000

11.K. M. Kiang, K. Frampton, T. M. Monro, R. Moore, R., J.Tucknott, D.W. Hewak, D. J. Richardson, and H. N. Rutt,“Extruded singlemode non-silica glass holey optical fibres”,Electron. Lett. , vol.38, no.12 , pp.546 –547, June 2002.

The Fortieth Anniversary of the Injection LaserIn July 1962, at the Solid State Device ResearchConference in Durham NH, groups from MIT Lincoln Labsand RCA Laboratories reported GaAs LEDs with nearly100% internal efficiency. These results initiated a frenzy ofresearch activity at a variety of labs which lead, inSeptember 1962, to the first demonstration of the GaAsinjection laser by Robert Hall and his collaborators at

General Electric’s Corporate Research Laboratory inSchenectady NY. Subsequently, groups lead by RobertRediker (Lincoln Labs), and Marshall Nathan (IBMResearch Labs) also demonstrated GaAs lasers, and NickHolonyak (GE, Syracuse) demonstrated a visible GaAsPlaser-the first alloy semiconductor device. In a future issue,we will examine these seminal events in more detail.

Page 39: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 39

Optical Packet Switching and Associated Optical SignalProcessing

D. J. BLUMENTHAL, JOHN BOWERS, YI-JEN CHIU, HSU-FENG CHOU, BENGT-ERIK OLSSON*, SURESH RANGARAJAN,

LAVANYA RAU and WEI WANG

University of California, Santa Barbara, CA 93106Tel: (805) 893-4168; Fax: (805) 893-5705; Email: [email protected]

* Optillion AB, Sweden

AbstractIn this article we review the application of all-optical sig-nal processing using an ultra-fast all-optical nonlinearfiber wavelength converters to optical packet switchingand ultra-fast network applications. The wavelength con-verter can be used as a basic building block for these net-works and has been demonstrated at bit rates in excess of40 Gbps. Experimental optical packet switching andOTDM add/drop multiplexer experimental results at 80Gbps are reviewed.

IntroductionWithin today’s Internet, packets (the basic unit of Internetdata) are directed to their final destination using electronicrouters. These packets are moved from router to routerusing optical fiber transmission and wavelength divisionmultiplexing (WDM) systems where data is transportedover different wavelength (colors) of light that are combinedonto the same fiber. Today’s fiber systems carry a typical 32-80 wavelengths modulated at 2.5 Gbps (1 Gbps = 109 bitsper second) to 10 Gbps per wavelength while routers arerequired to handle almost 1 Terabits (1012) per second.Things become interesting when we consider that the data carryingpotential of optical fibers continues to double every 8-12 monthswith state-of-the-art single fiber capacity exceeding 10 Tbps.Comparing this increase with that of electronic processor speeds thatdoubles every 18 months (Moore’s Law) and comes at the expense ofincreased chip power dissipation we see that there is a potentialbandwidth mismatch in handling capability between fiber transmis-sion systems and electronic routers and switching systems.

The story is more complex when we consider that future routersand switches will potentially terminate hundreds or thousands ofoptical wavelengths and the increase in bit-rate per wavelength willhead out to 40 Gbps and beyond to 160 Gbps. Additionally, elec-tronic memory access speeds only increase at the rate of approxi-mately 5% per year, an important data point since memory plays akey role in how packets are buffered and directed through therouter. It is not difficult to see that the process of moving a massivenumber of packets per second (100 million packets/second andbeyond the 1 Billion packets/second mark) through the multiplelayers of electronics in a router, can lead to router congestion andexceed the performance of electronics and the ability to efficientlyhandle the dissipated power.

In this article we review research at the University of California,Santa Barbara in fast optical signal processing as it applies to trans-

mission, time division multiplexed and packet switched networks.We will also describe how the use of optical signal processing tech-niques can be used to alleviate the bottlenecks in transmission androuting as described above.

Synchronous and Asynchronous NetworksTwo basic approaches exist to carry packet data over a network andare shown in Figure 1. The asynchronous approach in Figure 1aallows a router to take packets arriving at random times at itsmultiple inputs and redirect them to various outputs withoutoverlapping. This function requires memory (buffers) inside therouter to temporarily hold packets so they can be delayed withrespect to other packets appropriately and merged at the outputs.This is analogous to the on-ramp of a highway where or a mergelane where cars must adjust their speed and wait time in order tomerge without colliding. Packets that arrive on different wave-lengths will most likely have to be merged onto a common output.The synchronous type network shown in Figure 1b is more analo-gous to loading up a railroad car and transporting the cars on atrack. Before packets can be transported on the synchronous net-work, they are loaded up into “frames.” The network then switchesand routes based on these frames and not the packets inside them.

Figure 1. Two methods to transport packets on a network (a) using asynchronous multiplexing and(b) using synchronous multiplexing.

Page 40: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

40 IEEE LEOS NEWSLETTER OCTOBER 2002

Ultra-Fast Optical Wavelength Converter forSignal Processing and Network FunctionsWe have utilized an all-optical fiber wavelength converter as thebasic building block for both asynchronous and synchronous pack-et switched networks [1]. This wavelength converter can be usedto imprint data from one optical wavelength onto a new opticalwavelength without passing the data through electronics. In addi-tion to wavelength conversion, it can also be used to regenerate thebits in a digital signal and implement higher-level functions forasynchronous and synchronous packet networks. This approach isespecially useful when the data rate exceeds 40 Gbps where elec-tronics is not readily available. The wavelength conversion processis based on cross-phase modulation (XPM) in dispersion shifted

optical fiber and is shown in Figure 2 along with conver-sion of return-to-zero (RZ) 80 Gbps data stream. Theconverter operation is based on the principle of XPM in anon-linear fiber, such as a dispersion-shifted fiber (DSF).A CW signal or a pulse train at a new wavelength λj iscombined with an intensity modulated pulse train atwavelength λj. The incoming data imposes a phase mod-ulation of the CW signal or the pulse train due to XPM.This phase modulation causes a spectral broadening ofthe CW signal or the pulse train thereby generating side-bands. One of these sidebands is filtered to convert phasemodulation to amplitude modulation. The filter arrange-ment consists of a fiber Bragg grating (FBG) and a tun-able band-pass filter (BPF). The FBG notches out theoriginal data signal and the non spectrally broadenedpart of the new signal and lets only the desired sidebandthrough. This improves the extinction ratio of the con-verted signal. This method of wavelength conversion isin principle very fast since non-linear processes arealmost instantaneous and thus can be used to wavelengthconvert very high bit-rate data. The wavelength convert-er also acts as a 2R regenerator as seen by the smoothingout of bit amplitude fluctuations at the output.

Optical Packet Switching and Label Swappingfor Asynchronous NetworksAll-Optical Label Swapping (AOLS) is a type of optical packetswitching that is intended to solve the potential mismatchbetween fiber capacity and router packet forwarding capacity.AOLS imparts the functionality to direct packets through an opti-cal network without the need to pass these packets through elec-tronics whenever a routing decision is necessary [2-6]. Inherent tothis approach is the ability to route packets independently of bit-rate, packet or coding format and packet length.

An example AOLS network is illustrated in Figure 3. Internet

Figure 2. All-optical fiber optic cross-phase modulation (XPM) wavelength converter.

Figure 3. An optical label-swapping network and example of AOLS with 80 Gbps packets.

Page 41: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 41

Protocol (IP) packets enter the network through an “ingress” nodeand are encapsulated with an optical label and then re-transmitted ona new wavelength. Once inside the network, only the optical label isused to make routing decisions and the wavelength is used to dynam-ically redirect (forward) packets. At the internal nodes, labels are readand optically erased, then a new label is attached to the packet andthe optically labeled packet is converted to a new wavelength usingall-optical wavelength conversion. Throughout this process, the con-tents (e.g., the IP packet header and payload) are not passed throughelectronics and are kept intact until the packet exits the optical net-work through the “egress” node where the optical label is removedand the original packet is handed back to the electronic routing hard-ware. For packet networks where the bit rate can exceed 40 Gbps,ultra-fast signal processing techniques have been used to perform thefunctions of (i) optical label removal, (ii) optical wavelength conver-sion, (iii) optical signal regeneration and (iv) optical label replace-ment. We have demonstrated that the XPM fiber optic wavelengthconverter can be used to perform these functions for packet bit-ratesas high as 80 Gbps with the potential to scale to data rates in excessof 160 Gbps [7]. Figure 3 also shows results of AOLS with 80 Gbpspackets and labels running at 10 Gbps.

Synchronous Ultra-Fast WDM/OTDM NetworksWe have demonstrated that the XPM fiber wavelength convertercan be used to integrate WDM and OTDM networks all-optically.Various functions required by a hybrid WDM-OTDM networkincluding the ability to a) multiplex several low bit-rate DWDMchannels into a single high bit-rate OTDM channels, b) demulti-plex a single high bit-rate OTDM channel into several low bit-rateDWDM channels c) add and/or drop a time-slot from an OTDMchannel d) wavelength route OTDM signals [8-11]. In addition tothese basic functionalities the ability to multicast high bit-rate

signals can be a very useful feature. The advantages of performingthese functions all-optically are scalability and potential lowercosts by minimizing the number of O-E-O conversions. An exam-ple of a WDM-OTDM is presented in Figure 4. One example isthe all-optical OTDM add/drop mutiplexer shown in Figure 5.This building block has the capability of removing bits in one slotof the OTDM time frame and replacing new bits without passingthe high-speed bus through electronics.

AcknowledgementsThis work was supported by the DARPA supported multidiscipli-nary optical switching technology center (MOST), a DARPA spon-sored NGI award, a DARPA DURIP award, and funding fromCisco Systems.

References1. “All-Optical Demultiplexing using Fiber Cross-Phase Modulation

and Optical Filtering,” B. E. Olsson and D. J. Blumenthal, IEEEPhotonics Technology Letters, 13 (8), pp. 875-877, August (2001).

2. “Routing Packets with Light,” D. J. Blumenthal, Scientific American,January (2001).

3. B. Bostica, A. Cappellari, M. Burzio, B. Vercellone, C. Guillemot,A. Gravey, P. Gravey, F. Masetti, M. Sotom, M. Renaud, “A NovelPacket Switching network Adopting Transparent Optical Packets,”CSELT Technical Reports, Vol. 24, No. 6, CSELT, pp. 1049-56, Dec.1996.

4. D. J. Blumenthal, B-. E. Olsson, G. Rossi, T. Dimmick, L. Rau, M.Masanovic, O. Lavrova, R. Doshi, O. Jerphagnon, J. E. Bowers, V.Kaman, L. A. Coldren and J. Barton, “All-Optical Label SwappingNetworks and Technologies,” IEEE Journal of Lightwave Technology,Special Issue on Optical Networks, Dec. 2000.

Figure 4. WDM/OTDM network based on ultra-fast all-optical wavelength converters.

Page 42: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

42 IEEE LEOS NEWSLETTER OCTOBER 2002

5. Y. M. Lin, W. I. Way, and G. K. Chang, “A Novel Optical LabelSwapping Techniques using Erasable Optical Single0SidebandSubcarrier Label,” IEEE Photonic Technology Letters, Vol. 12, pp.1088-90, Aug. 2000.

6. A. Carena, M. D. Vaughn, R. Gaudino, M. Shell and D. J.Blumenthal , “OPERA: An Optical Packet Experimental RoutingArchitecture with Label Swapping Capability,” IEEE Journal ofLightwave Technology, Special Issue on Photonic Packet Switching, Vol.16, No. 12, pp. 2135-45, Dec. (1998).

7. L. Rau, S. Rangarajan, D. J. Blumenthal, H.-F. Chou, Y.-J. Chiuand J. E. Bowers, “Two-Hop All-Optical Label Swapping withVariable Length 80 Gb/s Packets and 10 Gb/s Labels usingNonlinear Fiber Wavelength Converters, Unicast/Multicast Outputand a Single EAM for 80- to 10Gb/s Packet Demultiplexing,”Conference on Optical Fiber Communications, Postdeadline paperFD1, Conference on Optical Fiber Communications (OFC ’02), Anaheim,CA, Mar 19-22 (2002).

8. “WDM to OTDM Multiplexing using an Ultra-fast All-OpticalWavelength Converter,” B. E. Olsson, L. Rau and D. J.Blumenthal, IEEE Photonics Technology Letters, 13 (9), September,(2001).

9. B. E. Olsson and D. J. Blumenthal, “80 to 10 Gbit/sDemultiplexing using Fiber Cross-Phase Modulation and OpticalFiltering,” Proceedings of the IEEE/LEOS 13TH Annual Meeting,(LEOS ’00) Puerto Rico, Paper TuB 4, pp. 159-160, November 13– 16, (2000).

10.L. Rau and D. J. Blumenthal, “Wavelength Multicasting Using anUltra High-Speed All-Optical Wavelength Converter,” TechnicalDigest of the Optical Fiber Communication Conference (OFC ’01),Anaheim, CA., March 17-23, (2001).

11.“All-optical add-drop of an OTDM channel using an ultra-fast fiberbased wavelength converter,” L. Rau, S. Rangarajan, W. Wang andD. J. Blumenthal, Conference on Optical Fiber Communications (OFC’02), Anaheim, CA, Mar 19-22 (2002).

Figure 5. Example of an all-optical OTDM add/drop multiplexer using the XPM fiber wavelength converter.

Page 43: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

Ultrahigh-Speed Optical Signal Processing with Symmetric-Mach-Zehnder-Type All-Optical Switches

K. TAJIMA, S. NAKAMURA, and Y. UENO*

Networking Res. Labs. NEC Corporation, 34 Miyukigaoka, Tsukuba, Ibaraki 305-8501, JapanTel: +81-298-50-1126 Fax: +81-298-50-1106 E-mail: [email protected]

* Present address: The University of Electro-Communications, Chofu, Tokyo, Japan

1. IntroductionAll-optical signal processing technology (AOSPT) is expected toimprove the efficiency, flexibility, and capacity of optical networks.It is also expected to be useful in reducing the cost, size, andpower consumption of optical components used in these networks,primarily because the conversion between optical and electricalsignals can be avoided. For these reasons, this technology is beingactively pursued not only in the ultrafast regime, but also at speedsof a few tens of Gb/s.

One of the key devices for AOSPT is an all-optical switch. Thisarticle reports ultrafast all-optical switches and all-optical signal pro-cessing experiments.

2. Symmetric Mach-Zehnder-type all-opticalswitchesAll-optical switches must satisfy various requirements simultane-ously in order to make AOSPT practical and competitive. One isthat the control-light-pulse energy must be much less than 1 pJ,otherwise the control light power is too high even at a few tens ofGb/s. To satisfy this requirement, semiconductor optical amplifiers(SOAs) are used for nonlinear elements to enhance the bandfillingnonlinearity in semiconductors by stimulated emission. Thisenables true femtojoule switching, but the relaxation time of SOAslimits the operating speed to less than a few tens of Gb/s. Apromising way to mitigate this problem is to use a differentialphase modulation (DPM) scheme. A representative DPM device isthe Symmetric Mach-Zehnder (SMZ, some-times called an SOA-MZI) all-optical switch[1]. In SMZs, the differ-ence in nonlinear phaseshifts (φNL) opticallyinduced in the two non-linear elements definesthe state of the switch, sothe effect of slow relax-ation is circumvented, asillustrated in Fig.1. TheDPM scheme also real-izes a rectangle-likeswitching window,which is highly desirablein many applications.

There are two variantsof the SMZ switch: a

polarization-discriminating SMZ [2,3] (PD-SMZ, often called UNI)and a delayed-interference signal-wavelength converter (DISC) [4].These devices can be configured in inherently stable forms suitablefor bulk optics implementation. The original SMZ switch on theother hand can be unstable unless integrated. Hybrid-integrated(HI-) [5] and monolithic [6,7] SMZ-type switches have beenreported.

3. Ultrafast all-optical signal processing experimentsTransmission experiments faster than 100 Gb/s per wavelengthhave been reported [8,9]. In these experiments, all-optical switcheswere used as demultiplexers to achieve very high bit rates. For bit-rates faster than 200 Gb/s, however, all-optical switches based onthe Kerr effect in optical fibers have been used, but compact andless-power-hungry semiconductor devices are desirable. Recently,we have achieved error-free demultiplexing from 336 to 10.5 Gb/s(multiples of 82 MHz for synchronization with a streak cameraetc.) using a HI-SMZ switch [10].

In the experiment, 336-Gb/s signal pulses (1561 nm) and 10.5-GHz control pulses (1546 nm) were input into the HI-SMZ switch.The durations of the control and signal pulses are important parame-ters. Our analysis of the crosstalk between the demultiplexed chan-nel and adjacent channels indicates that a 1.5-ps control pulse canopen a 3.0-ps switching window for 336-Gb/s demultiplexing. It isnot necessary to use an extremely short control pulse, which is also

advantageous in reducingthe effect of carrier heat-ing in SOAs, which caus-es degradation of theextinction ratio [11]. Onthe other hand, the signalpulse-width was set to0.7 ps, because thecrosstalk is more sensitiveto it. Fig. 2(a) shows thewaveform of the inputsignal pulses measured bya cross-correlation tech-nique. As shown in Fig.2(b), measured bit errorrates (BERs) of thedemultiplexed signalpulses reached an error-free level where BERswere less than 10-9.

Signal pulses

Switch-oncontrol pulse

Switch-offcontrol pulse

T

φNL in arm 1

φNL in arm 2

T

Nonlinearwaveguide 1

Nonlinearwaveguide 2

T Canceling

Time

Fig. 1. Schematic illustration of a Symmetric-Mach-Zehnder all-optical switch. The differential phasemodulation scheme is shown in the inset.

OCTOBER 2002 IEEE LEOS NEWSLETTER 43

Page 44: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

44 IEEE LEOS NEWSLETTER OCTOBER 2002

In the above demultiplexing experiment, an all-optical switchwas driven by a regular-pulse-train at an electronic bit-rate.However, other applications such as wavelength conversion and 3Rregeneration require all-optical switches to operate randomly at ahigher transmission bit-rate, which is a tougher situation for opticalswitches. Here, we describe 3R operation at 84 Gb/s and wave-length conversion at 168 Gb/s.

The “3R” regeneration refers to re-timing, re-shaping, and re-amplification of signal pulses that are distorted after transmission.In the experiment [12], a PD-SMZ switch was used as a pulseregenerator. Signal pulses carrying 84-Gb/s data (1560 nm, 2.1 ps)drove the PD-SMZ to encode 84-GHz clock pulses (1547 nm, 2.8ps) so that the 84-Gb/s data was copied onto the clean clock pulsesat 1547 nm. The pattern effect of the switch was suppressed by set-ting the average power of the clock pulses higher [13]. The averagepowers of the signal and clock pulses at the input of the SOA mod-ule in the PD-SMZ switch were -1 and +5 dBm, respectively. Fig.3(a) shows the measured BERs after 84-Gb/s pulse regeneration bythe PD-SMZ switch and demultiplexing back to 10.5 Gb/s by theHI-SMZ switch, which coincided with the curve for the case with-out regeneration. This shows that the power penalty due to theregeneration was negligibly small. We also evaluated the retimingcapability. Fig. 3(b) shows the variation in BER caused by inten-tionally displacing the timing between the signal and clock pulses,

while other parameters were kept unchanged. Within the range of2.3 ps, the degradation of the BER was below 1 digit. This charac-teristic stems from the nearly rectangular switching window of theSMZ-type switch [1].

In the wavelength conversion experiment [14], 168-Gb/s signalpulses (1564 nm, 1.8 ps) drove a DISC to modulate a CW light(1547 nm). The average powers of the signal pulses and the CWlight at the input of the SOA module were +10 and +16 dBm,respectively. The DISC generated 168-Gb/s, 2.0-ps pulses at 1547nm. As shown in Fig. 4(a), a clear eye opening was obtained in theeye-diagram of the signal pulses after demultiplexing to 10.5 Gb/s.Fig. 4(b) shows the results of BER measurement, indicating error-free operation. The results confirm that the SMZ-type switches canbe driven by data-modulated 160-Gb/s pulses.

4. ConclusionWe have demonstrated various types of ultra high-speed opticalsignal processing with Symmetric-Mach-Zehnder-type all-opticalswitches: error-free 336-Gb/s demultiplexing, penalty- and error-free 84-Gb/s pulse regeneration, and error-free 168-Gb/s wave-length conversion. These are the fastest experiments reported todate in each category. This work was performed partially under themanagement of the FESTA supported by the NEDO.

Fig.2. 336 Gb/s demultiplexing with HI-SMZ switch. (a) Cross-correlation trace of input signal pulses. (b) Results of BER measurements.

MUX+3R + DEMUX

-40 -35 -30 -25 -20

log(

BE

R)

Received power (dBm)

-4

-5

-6

-7

-8

-9

-10-11-12

MUX+DEMUX

PRBS 231-110.5-Gbpsbaseline

-3 -2 -1 0 1 2 310-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

BE

R

Time difference (ps)

2.3 ps

(a) (b)

Fig. 3. 84-Gb/s pulse regeneration with PD-SMZ. (a) Results of BER measurement. (b) BER variation with the relative timing change between the signal and clock pulses.

Page 45: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

References1. K. Tajima, Jpn. J. Appl. Phys. 32, L1746 (1993).2. K.Tajima, S. Nakamura, and Y. Sugimoto, Appl. Phys. Lett. 67,

3709 (1995).3. S. Nakamura, Y. Ueno, and K. Tajima, IEEE Photon. Technol. Lett.

10, 1575 (1998).4. Y. Ueno, S. Nakamura, K. Tajima, and S. Kitamura, IEEE

Photonics Technol. Lett. 10, 346 (1998).5. K. Tajima, S. Nakamura, Y. Ueno, J. Sasaki, T. Sugimoto, T. Kato,

T. Shimoda, M. Itoh, H. Hatakeyama, T. Tamanuki, and T. Sasaki,IEE Electron. Lett. 35, 2030 (1999).

6. R. Hess, M. Caraccia-Gross, W. Vogt, E. Gamper, P. A. Besse, M.Duelk, E. Gini, H. Melchior, B. Mikkelsen, M. Vaa, K. S.Jepsen, K. E. Stubkjaer, and S. Bouchoule, IEEE Photon.Technol. Lett. 10, 165 (1998).

7. J. Leuthold, C. H. Joyner, B. Mikkelsen, G. Raybon, J. L.Pleumeekers, B. I. Miller, K. Dreyer, and C. A. Burrus, IEEElectron. Lett. 36, 1129 (2000).

8. M. Nakazawa, ECOC 2001, Tu.L.2.3, Amsterdam 2001.9. U. Fieste, R. Ludwig, C. Schubert, J. Berger, S. Diez, C.

Schmidt, H. G. Weber, B. Schmauss, A. Munk, B. Buchold,

D. Briggmann, F. Kueppers, and F. Rumpf, IEE Electron. Lett. 37,443 (2001).

10.S. Nakamura, Y. Ueno, and K. Tajima, OFC 2002, FD3(PD),Anaheim 2002.

11.S. Nakamura, Y. Ueno, and K. Tajima, Appl. Phys. Lett. 78, 3929(2001).

12.Y. Ueno, S. Nakamura, and K. Tajima, IEEE Photonics Technol.Lett. 13, 469 (2001).

13.R. J. Manning, A. D. Ellis, A. J. Poustie, and K. J. Blow, J. Opt.Soc. Am. B14, 3204 (1997).

14.S. Nakamura, Y. Ueno, and K. Tajima, IEEE Photon. Technol. Lett.13, 1091 (2001).

OCTOBER 2002 IEEE LEOS NEWSLETTER 45

Fig. 4. 168-Gb/s wavelength conversion with DISC. (a) Eye-diagram of wavelength-converted pulses after demultiplexing. (b) Results of BER measurement.

The following individuals were elevatedto Senior Member membership gradethru September 2002:

New SeniorMembers

Stephen J. Clements

Hossein Eslambolchi

Israel Greiss

Daniel K. Lau

Reginald J. Perry

Roberto Sabella

Tadashi Saitoh

Chester C. Shu

Leif Sornmo

Andy Turudic

Page 46: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

2002 International Symposium on CompoundSemiconductors (ISCS 2002)Conference Dates: 7-Oct-2002 to 10-Oct-2002Hotel Alpha, Lausanne SwitzerlandConference URL: http://iscs2002.epfl.ch

International Conference on Optics-Photonics Design & Fabrication (ODF 2002)Conference Dates: 30-Oct-2002 to 1-Nov-2002National Museum of Emerging Scienceand Innovation, Tokyo JapanConference URL: http://annex.jsap.or.jp/OSJ/meet/ODF2002/index.html

International Topical Meeting onMicrowave Photonics (MWP2002)Conference Dates: 5-Nov-2002 to 8-Nov-2002Awaji Yumebutai Int’l Conference Center,Hyogo JapanConference URL: www.casjo.org/mwp2002/

IEEE LEOS 15th Annual Meeting (LEOS 2002)Conference Dates: 11-Nov-2002 to 14-Nov-2002Scottish Exhibition and ConferenceCentre, Glasgow , Scotland UKConference URL: http://www.i-leos.org/info/calendar2002.html

Conference on Optoelectronic andMicroelectronic Materials and Devices(COMMAD 2002)Conference Dates: 11-Dec-2002 to 13-Dec-2002University of New South Wales, Sydney AustraliaConference URL: http://www.commad.unsw.edu.au

Sixth International Conference onOptoelectronics, Fiber Optics andPhotonics (PHOTONICS) Conference Dates: 16-Dec-2002 to 18-Dec-2002Tata Institute of Fundamental Research,Mumbai IndiaConference URL: http://ph2002.tifr.res.in/

2003 Optical Fiber CommunicationConference and ExpositionConference Dates: 23-Mar-2003 to 28-Mar-2003Georgia World Congress Center, Atlanta,GA USAConference URL: www.ofcconference.org

Workshop on Interconnections within HighSpeed Digital SystemsConference Dates: 4-May-2003 to 7-May-2003La Posada de Santa Fe, Santa Fe, NMUSAConference URL: http://www.ieee.org/organizations/society/leos/LEOSCONF/HSD03/santafe03.htm

International Conference on IndiumPhosphide and Related Materials (IPRM2003)Conference Dates: 12-May-2003 to 16-May-2003Fess Parker’s Doubletree Resort, SantaBarbara, CA USAConference URL: www.ece.ucsb.edu/IPRM03/

Quantum Electronics & Laser ScienceConference (CLEO/QELS 2003) colocatedwith CLEO 2003Conference Dates: 1-Jun-2003 to 6-Jun-2003Baltimore Convention Center, Baltimore,MD USAConference URL: www.cleoconference.org

Conference on Lasers and Electro-Optics(CLEO/QELS 2003) colocated with QELS2003Conference Dates: 1-Jun-2003 to 6-Jun-2003Baltimore Convention Center, Baltimore,MD USAConference URL: www.cleoconference.org

Conference on Lasers and Electro-Optics/Europe (CLEO/Europe-EQEC 2003)co-located with EQEC 2003Conference Dates: 23-Jun-2003 to

27-Jun-2003Munich ICM, Munich GermanyConference URL: www.cleoeurope.org

European Quantum ElectronicsConference (CLEO/Europe-EQEC 2003)co-located with CLEO/EuropeConference Dates: 23-Jun-2003 to 27-Jun-2003Munich ICM, Munich GermanyConference URL: www.cleoeurope.org

Pacific Rim Conference on Lasers andElectro-Optics (CLEO/Pacific Rim 2003)Conference Dates: 22-Jul-2003 to 26-Jul-2003Taipei International Convention Center,Taipei Taiwan, R.O.C.Conference URL:http://cleo2003.ee.ntu.edu.tw/

International Symposium on CompoundSemiconductors (ISCS 2003)Conference Dates: 25-Aug-2003 to 27-Aug-2003University of California San Diego -, La Jolla, CA USAConference URL: www.I-leos.org

European Conference on OpticalCommunicationConference Dates: 21-Sep-2003 to 25-Sep-2003Rimini Fiera, Rimini ItalyConference URL: www.ecoc.it

IEEE LEOS 16th Annual Meeting (LEOS2003)Conference Dates: 26-Oct-2003 to 30-Oct-2003Sheraton El Conquistador, Tucson, AZ USAConference URL: www.I-leos.org

2004IEEE LEOS 17th Annual Meeting (LEOS2004)Conference Dates: 7-Nov-2004 to 11-Nov-2004Westin Rio Mar Beach Resort, CountryClub & Ocean Villa, Rio Grande, PR USAConference URL: www.I-leos.org

Conference Calendar www.i-leos.org

ICO Prizes and Awards : Call for Nominations

The October 2002 (issue number 53) of the ICONewsletter has been posted on the ICO Websitewww.ico-optics.org.

ICO, the International Commission for Optics, hasestablished three awards: the “ ICO Prize”, the ICOGalileo Galilei Award and the ICTP/ICO Award. The

latter was established jointly with ICTP, theInternational Commission for Theoretical Physics. Thedeadline for the ICTP/ICO Award is November 10,2002, while for the two former awards it is March 15.Rules are reproduced from the ICO web site,www.ico-optics.org, where complete lists of previousrecipients can also be found.

46 IEEE LEOS NEWSLETTER OCTOBER 2002

Page 47: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

Changes in 2003 Membership Renewal

In the membership-renewal informationyou recently received, you will find anannouncement of the new “Lasers and

Electro-Optics Society Digital Library” onpage three and again under the LEOS entry.Whereas members have had free access toJQE, PTL, JSTQE, and JLT volumes from1988 forward via IEEE Xplore™, the“backfiles” from the LEOS Digital Archivesfor earlier issues of JQE (1965 - ) and JLT(1983 - ) are being added to Xplore. Over40 conference proceedings will also beincluded in this new Digital Library, whichremains free to members.

How does the new Digital Library differfrom the existing Digital Archives? Thetwo most-noticeable differences are the userinterface and the new conference content.

First, the Digital Archives reside on theLEOS server at www.i-LEOS.org and use ahighly intuitive GUI. The Digital Libraryis available at http://ieeexplore.ieee.org/,although there are several quick links toXplore on the LEOS portal.

Second, although the conference proceed-ings were already part of the IEL (IEEE/IEEElectronic Library), they were not availableto individual users unless your company oruniversity subscribed to the IEL and deliv-ered it over their LAN. We have changedthose access arrangements for members to getLEOS-only (vs. LEOS-cosponsored) confer-ences, and we are examining possibly addingproceedings from LEOS-cosponsored confer-ences in the future. None of this content is inthe Digital Archives.

There are less-obvious differences youwill notice as you use each service. For

instance, there are no Rapid Postings ofaccepted papers in the Digital Archives,and the Digital Archives may run an issueor two behind Xplore. Also, referencelinks, bibliographic links, and searches inXplore are more extensive because theyspan all the IEL content, including jour-nals, magazines, and IEEE and IEE confer-ences. Equivalent features in the DigitalArchives span only LEOS journal publica-tions. These and other comparisons aresummarized in the table above.

The other important change is“unbundling” of the annual LEOS CD-ROM Journal Collection. From 1995 until2000, it was offered to members for$25.00. In 2000 it was provided free tomembers. Now IEEE is closely examiningthe cost components of membership, and

encouraging the Societies to structure theirdues to cover the variable costs. In June,the Societies endorsed this concept. Thechoices are either to raise dues or unbundle“free extras.” Therefore, a decision wasmade to keep our dues at their current val-ues and allow members to continue receiv-ing the CD-ROM as an optional $5.00add-on. When priced at $25.00, fewer than5% of LEOS members selected the option.This suggests that many of the free CD-ROMs are not used, which effectively raisesthe expense for providing them to thosewho do. By unbundling the CD-ROM, weexpect the production run, and associatedexpense, to fall, allowing the funds to beapplied for other benefits more ubiquitous-ly valued such as the Digital Library. Wewill examine the financial results of this

LEOS Digital LEOS DigitalFeature Library Archives

Free to LEOS members 4 4

LEOS conferences (1988 - ) 4 No

JQE, PTL, JSTQE, JLT (1965 - ) 4 4

Rapid Postings (PTL, JLT) 4 No

Full-text search capability 4 4

Links to all IEEE, IEE content 4 No

Linked reference terms 4 4

Author bibliographies 4 4

Author biographies No 4

Easy identification of Special Issues No 4

Annual indexes, announcements No 4

Copy&Paste IEEE-formatted references No 4

Easy-to-navigate interface No 4

Four Elected to BoG

Roel Baets Ghent University, Belgium

Nadir Dagli University of California at Santa Barbara, US

Mary Y.L. Wisniewski IBM TJ Watson Research Center, US

Yuzo Yoshikuni NTT Photonics Labs, Japan

The following four candidates were elected for three-year terms to the LEOS Board of Governorsbeginning 1 January, 2003:

The twelve elected members of the Board of Governors have special responsibilities for determiningpolicy and direction for the Society. Please contact any of them with your ideas for future directions ofthe Society.

OCTOBER 2002 IEEE LEOS NEWSLETTER 47

Page 48: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

48 IEEE LEOS NEWSLETTER OCTOBER 2002

Chapter Highlights

IEEE LEOS TURKEY CHAPTERIn July 1998, the IEEE LEOS presidentDr. J. Gary Eden invited the LEOSTurkish members to form a Chapter inTurkey. After a year of membershiprecruitment, the IEEE LEOS TurkeyChapter was officially approved inNovember 1999. Since then the chapterhas organized numerous technical activitiesto foster the development of opticsresearch in Turkey. In particular, work-shops on Electro-Optics (Fig. 1) have beenorganized annually since 1999 in order tofacilitate the exchange of ideas amongTurkish researchers and engineers workingin the field of optics. IEEE LEOS TurkeyChapter took an active part as a cosponsorof these events by providing support forpublicity, and travel money for studentparticipants. The number of attendees ofthe Electro-Optics workshops has beensteadily increasing. In addition, IEEELEOS Turkey Chapter held numeroustechnical seminars in Istanbul and Ankara.Distinguished speakers of IEEE LEOSwere also hosted to give talks during theseevents (Fig. 2). Recently, our chapter wasawarded the 2001 largest membershipincrease prize which was shared with theUkraine Chapter.

In 2002, several new types of activitiesbesides technical seminars and workshopswere started. In particular, two shortcourses were organized in January andMay 2002 on high-speed photodetectorsand ultra-resolution optical microscopy.These were delivered by Prof. Selim Unlufrom Boston University. Second, anindustrial training program was initiatedto give educational seminars or shortcourses to companies or research insti-tutes about current topics of importancein photonics. Volunteer chapter memberswill be responsible for contacting the

companies and organizing these seminars.The first one, an educational seminarabout mode locking of lasers, was deliv-ered by Alphan Sennaroglu at theNational Metrology Institute in Gebze,Turkey in January 2002.

Finally, the IEEE LEOS Turkey chapterhas been actively involved in recruiting newstudent members to IEEE LEOS. Studentsfrom various Turkish universities becameIEEE LEOS members in 2002. In order toincrease undergraduate students’ familiaritywith state-of-the-art optical technologies,our chapter has also been involved in thepreparation of optics demonstrations. As anexample, a laser voice transmitter (Fig. 3),was built between January and June 2002.Recently recruited LEOS student memberswere actively involved in the design andconstruction.

For more information about the IEEELEOS Turkey Chapter, visit our web pageat http://home.ku.edu.tr/~ieee-leos/

Alphan SennarogluIEEE LEOS Turkey Chapter

Figure 1: Attendees of the Second Turkish Workshop on Electro-Optics. October 20, 2000, Koç University,Istanbul, Turkey.

Figure 3: Laser Voice transmitter built by the IEEELEOS Turkey Chapter in Spring 2002. (UmitDemirbas appearing in the picture is an IEEE LEOSstudent member)

Figure 2: Hugo Thienpont, a LEOS distinguished speaker, giving a lecture at the IEEE LEOS Turkey Chapter.(May 2002)

IEEE LEOS German Chapter The LEOS German Chapter was foundedin 1997 with a membership of 160. Thefirst elections took place in 1998 andProfessor Dieter Jäger (Duisburg) andProfessor Freude (Karlsruhe) have beenelected as chair and cochair, respectively,

and approved again in 2001 for the sec-ond term. As first activities of theChapter it was agreed to increase themembership and to appear in public bysponsoring different conferences as wellas to set up networks between industry,research institutes and universities.

Regular ActivitiesSince 1998 the German Chapter wasfrequently and regularly engaged incosponsorship of different conferences,such as “IEEE International TopicalMeeting on Microwave Photonics”,“Optics in Computing”, “Microoptics

Page 49: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 49

Chapter Highlights (cont’d.)

2000”, “European Conference onIntegrated Optics”, “InternationalConference on Terahertz Electronics”and “Microcavity Light Sources” or“Microwaves and Optronics”.Additionally, several Seminars andColloquia have been organized, forexample, an evening lecture on“Photonics – Key Technology of the21st Century” which has been attendedby about 200 students and the public.

In 2000 we started to provide an infor-mation channel for the members compos-ing an annual e-Newsletter. This newslet-ter is intended to feature the differentactivities of the Chapter and to highlightalso technical innovations relevant toLEOS fields. Recent reports addressed, forexample, the topics microwave photonicswith emphasis on fiber radio communica-tions, optoelectronic antenna arrays forwireless access and recent achievements inmicrooptics. The Chapter is currentlygoing to set up a web site.

The activities of the Chapter have fur-ther been presented at different fairs andexhibitions, in particular at LASER 2001in Munich, the Hannover Fair 2002, theOPTATEC 2002 in Frankfurt, and MED-ICA 2002 in Düsseldorf. On a muchsmaller scale but also very important, theChapter has been organizing technicaland educational seminars in cooperationwith companies, such as Agilent,Keithley and Hewlett-Packard.

Competence Centers inOptics and OptoelectronicsThe German Chapter has recently beenactive in establishing a close cooperationbetween science and industry. Severalworkshops, seminars and round tableplatforms have been organized in order tofoster the discussion and interactionbetween Universities and companies.Several questionnaires have been circulat-ed and are being evaluated at this time.In parallel, a national program aiming atthe promotion of so-called “CompetenceCenters for Optical Technologies” hasbeen initiated by the Federal Ministry ofEducation and Science in Germany and,as a result, eight centers have been

formed up to now, several of them underthe leadership of LEOS members. In con-sequence of these activities, a science cen-ter on optoelectronics, called “WiNO”,and the network “OpTech-Net” havebeen formed in the State of NorthrhineWestfalia working closely together withthe LEOS German Chapter.

European and InternationalActivities The German Chapter has particularlybeen engaged in forming cooperationswith other countries. We have cospon-sored bi-tateral symposia between Chinaand Germany on “Opto- andMicroelectronic Devices and Circuits”(SODC) in Nanjing in 2000 andStuttgart 2002. A corresponding jointresearch program has recently beenestablished. European competence cen-ters are also being promoted in order topave the way for international coopera-tions. In September the German Chapterhas taken part in the organization of ainternational workshop on “Optics-Microwave Interactions” and a EuropeanWorkshop on “High-Frequency OpticalSignal Generation and Processing forFuture Broadband Fiber Radio Systems”.

Education and Students The number of students in electricalengineering has dropped dramaticallyduring the years from 1993 to about2000. It is, however, gradually increas-ing now. On the other hand there is agreat lack of engineers in the industry,especially in optical and related areassuch as optoelectronics and photonics.The German Chapter is involved in thedefinition of new master courses in opti-cal technologies at Universities and informing new study programs withemphasis in optoelectronics and photon-ics. Courses for start-up companies havealso been developed.

The Future The membership of the LEOS GermanChapter is permanently increasing and

has reached a number of about 260 thissummer. In the following years we hopeto foster the continued growth of ourcommunity and to create an effectiveinfrastructure for education, research andindustry. We will further promote closecollaborations with and between existingregional, national and internationalresearch institutes and companies and weare going to proceed with the formationof relevant networks that will facilitateand encourage the sharing of knowledgeand expertise and the enabling of innova-tions in topic LEOS fields. Another ideawe have in mind is to start a workshopon a specific subject which would rotatebetween different LEOS chapters inEurope, similar to the WFOPC(Workshop on Fiber Optics PassiveComponents) which was initiated in Italyand is now organized by the ScottishChapter. We are further looking ahead tojoint Workshops, in particular with theMTT and ED German Chapters toarrange meetings on MicrowavePhotonics and Optoelectronic Devices,respectively.

Photos and contact addresses:

Prof. Dr. Dieter Jäger (Chair)ZHO-OptoelektronikGerhard-Mercator-Universität Duisburg47048 Duisburg/Germanye-mail: [email protected]

Prof. Dr. Wolfgang Freude (Vice-Chair)High-Frequency andQuantum ElectronicsLaboratoryUniversität Karlsruhe76128Karlsruhe/Germanye-mail: [email protected]

Page 50: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

50 IEEE LEOS NEWSLETTER OCTOBER 2002

LEOS Graduate Student Fellowship Winners Announced

The IEEE Lasers & Electro-Optics Society announced the win-ners of the LEOS Graduate Student Fellowships for 2002. TheFellowships will be presented during the Awards Ceremony atthe LEOS Annual Meeting to Matthew Emsley, CristianoGallep, Ronald Holzlohner, Michael Mielke, Michael Venditti,Lianshan Yan, Sebastian Weiczorek, Marc Sciamanna, MarcoPasserini, Pascual Munoz, Ju Han Lee, and Charlotte Marra.

The LEOS Graduate Student Fellowship Program was estab-lished to provide Graduate Fellowships to outstanding LEOS stu-dents members pursuing graduate education within the LEOSfield of interest.

For more information on the program, check the Student siteon the LEOS Home Page (www.i-LEOS.org)

President’s ColumnContinued from page 1

with excuses for non-performance. Howwould people grow in their ability andskills otherwise, or be happy when theyknow deep inside they are not being pro-ductive? So, I am talking about a profes-sionally demanding organization wherepeople derive their good feelings frombeing able to perform to their full poten-tial. It is only when we know we are doingwell at things that are meaningful and use-ful that we can develop a positive self-image, feel good about ourselves, andtherefore happy. That’s the core of whatleaders have to “protect.”

Ultimately the company succeedsbecause its employees and its leader havesucceeded. So let us examine what bringssuccess and look for commonality to alignthe goals of all parties involved. In order forthe company to succeed, employees must beproductive and totally professional to pro-vide the best “services” to customers. For anindividual employee, based on Maslow’shierarchy of needs, what is of ultimateimportance is “self-actualization,” which isto fulfill one’s potential to “be all that onecan be.” Given that we spend the most pro-ductive time of day at our work, and agroup can take on many more significantprojects than can one individual, the work-place has the opportunity to become thebigger self for employees. From that stand-point, setting high expectations, providingneeded resources so people can succeed, andtraining people to take on greater responsi-bilities perfectly align the interests of thecompany and employees to create a win-win. That’s what great leaders do. Not only

continued on page 54

Page 51: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 51

Awards and Recognitions at LEOS 2002

A special Awards Ceremony will beheld on Monday evening at the LEOSAnnual Meeting. The following awardswill be presented during the ceremony:

The IEEE/LEOS William StreiferScientific Achievement Award will bepresented to James Fujimoto “for pio-neering contributions to optical coher-ence tomography.” Jim Hsieh will bepresented the Aron Kressel Award “forpioneer development of 1300nmGaInAsP diode lasers for fiber opticcommunications application and entre-preneurial leadership” Timothy Day willbe presented the EngineeringAchievement Award “for contributionsto the development and commercializa-tion of external cavity tunable diodelasers for telecommunications.” Thisyear’s LEOS Distinguished ServiceAward will be presented to ConcettoGiuliano “for continuous professionalservice to LEOS, with special contribu-

tions to conferences and meetings andthrough his leadership as President.”

The Moscow Chapter has been voted“Chapter-of-the-Year”. Other LEOS chap-ter awards will be presented to Turkey for“Most Improved Chapter”, ItalianChapter for “Most Innovative Chapter”,and the Ukraine Chapter for “Chapterwith Largest Membership Increase.”

The 2001-2002 DistinguishedLecturers: James Baird, Milton Chang,Kent Choquette, Dennis Deppe, RickTrebino will be recognized for their ser-vice to the Society. Recognition as elect-ed members of the LEOS Board ofGovernors will be given to PeterDelfyett, Ursula Keller, Steve Newton,Katsunari Okamoto.

LEOS 2001 Best Student PaperAward will be presented to Oleg Sinkinand Alan P.S. Chang.

Our congratulations to all therecipients!

Applications and nominations areinvited for the position of DepartmentHead in Electrical and ComputerEngineering, beginning in August2003. Candidates should have aPh.D. in Electrical Engineering or aclosely related field, a distinguishedrecord of academic scholarship, acommitment to excellent instructionand graduate research, strongadministrative and interpersonal abili-ties, and an established record ofuniversity and professional serviceappropriate for appointment as aFull Professor of Electrical andComputer Engineering.

The Head will lead development ofthe department’s programs and ini-tiatives. The successful candidate willbe expected to build upon thedepartment’s strengths, recruit out-standing new faculty, and promotethe Department to internal and exter-nal constituencies. The Departmentoffers BS, MS, and Ph.D. degrees, andhas programs in communications,computer engineering, lasers, opto-electronics, signal processing, radarand microwave systems, and systemsand control. Further informationabout the position and the ECEDepartment is available atwww.engr.colostate.edu/ece/.

For full consideration, complete appli-cation materials including a cover let-ter with vision statement, resume, andnames of five references should bereceived by October 28, 2002.However, the search will continueuntil the position is filled. Applicationsand nominations should be emailedto: [email protected] (pre-ferred method) or sent to Dr. AllanKirkpatrick, Head, Department ofMechanical Engineering, ColoradoState University, Fort Collins, CO80523-1374. Colorado State Universityis an EEO/AA employer.

DepartmentHeadAnnouncementHead, Electrical andComputer Engineering,Colorado State University

Iga and Asbeck Receive IEEE FieldAwards

IEEE Technical Field Awards were established to provide special recognition for out-standing achievements in special fields of electrical and electronics engineering. Ourcongratulations to Kenichi iga and peter m. Asbeck, LEOS Members!

IEEE Daniel E. Noble Award to Kenichi Iga, Executive Director, Japan Society for thePromotion of Science, Tokyo, Japan.

“For pioneering developments fo surface emitting semiconductor lasers and arrays.”

IEEE David Sarnoff Award to Peter M. Asbeck, Professor, University of California, San Diego, CA.

“For development and applications of GaAsbased heterojunction bipolar transistors.”

IEEE Technical Field Award InformationLink to: http://www.ieee.org/portal/index.jsp?pageID=

corp_level1&path=about/awards&file=tfalist.xml&xsl=generic.xsl

Page 52: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

52 IEEE LEOS NEWSLETTER OCTOBER 2002

The 15th Annual Meeting of the IEEE Lasers & Electro-Optics Society

Conference ChairKatherine HallPhotonEx CorporationBedford, MA, USA

Program ChairFranz KaertnerMassachusetts Institute of TechnologyCambridge, MA, USA

Member-at-LargeChung-En ZahCorning IncorporatedCorning, NY USA

SECCScottish Exhibition & Conference CentreGlasgow, Scotland

10 – 14 November 2002

www.i-leos.org

Page 53: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 53

Page 54: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

54 IEEE LEOS NEWSLETTER OCTOBER 2002

will employees be happier, they will havejob security consistent with today’s environ-ment, which is to have transferable skills.

What makes a leader successful? Short ofbeing a superman/woman, it is no differentfrom what makes anyone successful, whichis having the self-confidence to admit one’sshortcomings and to get help. Great leadershave good people around them, help growtheir ability, and instill self-confidence inthem. They believe in their people, helpthem see their full potential, coach them,help them identify and remove obstacles,but never settle for less than excellence.They also accept mistakes as inherent to theprocess of learning, and have the courage toface and deal with performance issues hon-estly but with compassion. In that scenariothere will be open ended “resource gather-ing” at every level so that there is nothingthe company cannot succeed at. I can getexcited just thinking about it.

Milton ChangIncubic

(650) 279-2489www.incubic.com

http://www.incubic.com

President’s ColumnContinued from page 50

Page 55: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

OCTOBER 2002 IEEE LEOS NEWSLETTER 55

Physical-Layer Division200 Executive BoulevardOssining, NY 105621-914-923-2164

Network-Layer Division19 Christopher WayEatontown, NJ 077241-732-380-2620

Web Address:www.rsoftdesign.com

Global Distribution

BeamPROPTM • FullWAVETM • BandSOLVETM • GratingMODTM • LinkSIMTM • MetroWANDTM • BroadbandSWATTM • WirelessSWATTM

NETWORK MODELING

COMPONENT DESIGN

SYSTEM SIMULATION

STRATEGIC ANALYSIS

THE ONLY COMPANY TO OFFER A FULL RANGE

OF SIMULATION AND PLANNING SOFTWARE

AND SERVICES THAT SPAN THE ENTIRE

COMPONENT TO NETWORK-LEVEL HIERARCHY

NEED TO REDUCE YOUR

DESIGN & DEPLOYMENT COSTS...?HIT THE RIGHT KEY WITH RSOFT DESIGN GROUP

Page 56: LEOS - Aristotle University of Thessalonikiusers.auth.gr/npleros/Publications/Journal Pubs_pdfs/J.3_LEOS_Newsletter.pdf · OCTOBER 2002 IEEE LEOS NEWSLETTER 1 G reetings! I encourage

LEOS Mission StatementLEOS shall advance the interests of its

members and the laser, optoelectronics, andphotonics professional community by:• providing opportunities for information

exchange, continuing education, and pro-fessional growth;

• publishing journals, sponsoring confer-ences, and supporting local chapter andstudent activities;

• formally recognizing the professional con-tributions of members;

• representing the laser, optoelectronics, andphotonics community and serving as itsadvocate within the IEEE, the broader sci-entific and technical community, and soci-ety at large.

LEOS Field of InterestThe Field of Interest of the Society shall belasers, optical devices, optical fibers, andassociated lightwave technology and theirapplications in systems and subsystems inwhich quantum electronic devices are keyelements. The Society is concerned withthe research, development, design, manu-facture, and applications of materials,devices and systems, and with the variousscientific and technological activitieswhich contribute to the useful expansionof the field of quantum electronics andapplications.

The Society shall aid in promoting close

cooperation with other IEEE groups andsocieties in the form of joint publications,sponsorship of meetings, and other formsof information exchange. Appropriatecooperative efforts will also be undertakenwith non-IEEE societies.

Visit the LEOS Home Pagehttp://www.i-LEOS.org

THE INSTITUTE OF ELECTRICAL & ELECTRONICS ENGINEERS, INC.445 Hoes LanePiscataway, NJ 08855-1331, USAwww.i-LEOS.org

ADVERTISING RATESSix Times

Ad Size One Time (Save 15%)Black & White

Full Page $525 $2675Half Page $340 $1735Third Page (vert) $285 $1450Quarter Page Member $225 $1150

Color AdvertisingFull Page $825 $4475Half Page $640 $3535Third Page (vert) $585 $3250Quarter Page Member $525 $2950

Ad Size DimensionsFull Page 9.50" H x 7.50" WHalf Page 4.75" H x 7.50" WThird Page (vert) 9.50" H x 2.25" WQuarter Page 4.75" H x 4.00" W

The Newsletter is published bimonthly in February, April, June,August, October, and December. All copy, except memberplacement ads, must be camera-ready and received on the firstday of the month preceding the issue month. Other details canbe obtained from the Staff Editor at 732.562.3892.