presented ucsb - solid state technology review santa ... talks pdf/alferness.pdfpresented ucsb -...

35
Rod C. Alferness Bell Labs Research Optical Networking Research Division Senior Vice President Presented UCSB - Solid State Technology Review Santa Barbara, California November 19, 2002

Upload: dangdang

Post on 05-Jun-2019

215 views

Category:

Documents


0 download

TRANSCRIPT

Rod C. AlfernessBell Labs ResearchOptical Networking Research DivisionSenior Vice President

PresentedUCSB - Solid State Technology Review

Santa Barbara, CaliforniaNovember 19, 2002

SKK 2002. Source:

Lightwave CommunicationCelebrates 25 Years of Service

AT&T / Illinois Bell, ChicagoMay 11, 1977

AT&T / Illinois Bell, ChicagoMay 11, 1977

44.7 Mb/s x 2.4 kmMultimode Fiber & LED

Adapted from: Landers, today@lucent (http://today.lucent.com), May 9, 2002.

107

106

105

104

103

102

101

1

Cap

acity

-Dis

tanc

e (G

b/s

•km

)

1975 1980 1985 1990 1995 2000Year

7C28360.001ppt-Kogelnik-7/97

Optical Amplifiers1.5 µm SF Laser1.3 µm SM Fiber0.8 µm MM Fiber

FourthGeneration

ThirdGeneration

SecondGeneration

FirstGeneration

Doubles Every Year

Progress In Optical Systems- Research Enabled

Lucent Technologies – Proprietary

• Original Application Driver for WDM Technology

• Initially Justified Solely for Capacity Upgrade Over EmbeddedFiber Base

• Announced Systems with 40, 80…..Wavelengths

• Value Proposition− Single Amplifier for All Wavelength Channels− Utilize Embedded Fiber Base

Data In

WDM Point-to-Point

Data Outλ1

λ2

λN

λ1

λ2

λN

OMUX

ODMUX

OA OA OA OAXMTR

XMTR

RCVR

RCVR

RCVR

•••••

•••••

XMTR

WDM – Driving Down the Cost of Transmission and the Network

SKK 2002. Source:

State of the Telecomm Market

Ryan, CIBC, January 16, 2002.

“The Bubble”“The Bubble”

22 22 22 23 2331 33

3846

66

92North American Telco

Capital Expenditure (US$B)(EIXCs, ILECs, IXCs, CLECs) 66

53

6170

81

1990 1992 1994 1996 1998 2000 2002 2004

The Good News- Research/Technology

• WDM has had a Profound Impact on Transforming the Backbone Network both with Lower-Cost, Highly Scalable Transmission and Also in Networking

• The Key Elements of Optical Layer Networking- Ultra-Long Haul Systems, Optical XC and Optical ADM are Products

• Tremendous Advances in Optical Technology- Components and Transmission Fiber

• The Value that Optics Brings to the Network, Not Hype is Valued

• The Start-Up Frenzy has Provided Many in the Field with a Wide Range of Experiences and Skills which will be Valuable in the Rebuild

Despite the Current Setback, Optics has Taken its Place as a Major, Mainstream Technology in the World Economy

..and Even for the Market !• Continued Capacity Demand Growth

• Global Situation is Better than North America

• Metro and Access Remain Opportunities- Cost, Granularity of Bandwidth are Key Issues

• Inventories Have Been Significantly Reduced

• On Some Routes Bandwidth Margins are Quite Small

• Globally a Growing Push for Broadband Access

• Wetted Appetites for Broadband Services

…and Network Demand Continues to Grow-So Where Do We Go From Here?

• With a View of End-User Needs, Continue to Refine Our View of the Network Evolution

• Understand the Critical Network Element Functions Required for That Network Evolution Vision

• Examine the Functional Component Requirements to Realize Those Network Elements in the Most Cost-Effective, Scalable Way

• Leverage the Advantages and Attributes of Optics- Those We Currently Know as Well as Those Research Can Reasonably Make Happen

• Constantly Ask the Question? What is the Value Proposition That Optics Brings to This Network Solution?

Higher TDM rates(40Gbps)

Broader Amplification BandsAmplifier Bandwidth Wavelengths

12 nm 16

35 nm 80

80 nm 200

First Generation First Generation

Wideband Wideband

Ultra-WidebandUltra-Wideband

MEMS - All OpticalSwitches and OADMs

Dynamic GainEqualizing Filters

1530 1535 1540 1545 1550 1555 1560 1565

-10

-5

0

Wavelength (nm)

Pow

er (d

Bm

)

1530 1535 1540 1545 1550 1555 1560 1565

-25

-20

-15

Wavelength (nm)

Pow

er (d

Bm

)

Before equalization

After equalization

Optical Technology- Key Enabler When Applied to Real Network Issues

Cost per bit

1

10

100

1995 2000 2005

Syst

em C

ost (

$K/G

b/s)

Transmission Fiber

1550 signal(s)

1450/ 1550 nm WDM

ErAmplifier

Transmission Fiber

1453 nmPump

RamanAmplification

SKK 2002. Source:

Key Market Driver - Lower Cost, Scalable Networks That Improve Life Quality

Historically, cost of transport has fallen 35%

per year

Historically, Historically, cost of transport cost of transport has fallen 35% has fallen 35%

per yearper year

Gawrys, NFOEC 2001.

Near Term Challenge- Drive Down Component Costs

New Optical Technology has Reduced Cost/ Bit But Not Enough for Broadband Data

• Optical Components are Still Too Expensive

• Craft Intensive Assembly of Subsystems Makes Cost-Reduction and Ramp-Up (when needed) Difficult

• Lack of Cost-Effective Wavelength Agile Sources has Resulted in Excessive Inventories at all Stages of Component and Systems Manufacture

Automated Assembly, More Robust Components, Smarter Packaging and Increased Integration Will Be Essential for Needed Cost Targets

Strive for the Right Balance of Optics and Electronics

Optics has Demonstrated Advantages in Transmission and, Recently, in Massive High Capacity Switching, but

• Electronics Remains More Mature, More Integrable, More Functional, More Compact…..

• Cost-Effective Terminal Electronics (Coding) to Increase Reach and Enhance Capacity is Essential

• New Techniques to Achieve Cost-Effective Finer Granularity With Optics

• Integrated Opto-Electronic Circuits to Get the Best of Both Fields Remains an Important Target

Lightwave Capacity Trends

0.1

1

10

100

1000

0.01 0.1 1 10 100 1000

Data rate per channel (Gb/s)

Num

ber o

f cha

nnel

s

'80 '83 '86 '87

'89

'91'93

'95

'95

'96

'98

'98

'02

10Gb/s100Gb/s

1Tb/s

10Tb/s

TotalCapacity

'01

'01

100Tb/s

Opt

ics

Electronics

Photonic Transport Photonic Networks

– High Capacity Transmission

– Fixed Sharing Between Multiple Nodes – Passive Access of Wavelength Channels

– Automated Connection Provisioning – Flexible Adjustment of Bandwidth– Network Self-Healing/Restoration

WDM/Point-to-Point Transport

Fixed WDM/Multipoint Network

Photonic XC and WADMReconfigured WDM/Multipoint Network

Fiber Amplifier

Wavelength Multiplexer/Demultiplexer

Wavelength Add/Drop

Wavelength Cross-Connect

Wavelength Adaptation/Assignmentin Optical Networks

Wavelength Translation at Network Entry (Adaption)

λarb OTU λ1 (fixed) λ1 ε λsyst

Fix λ Source

Wavelength Translation/Assignment (Network Controlled)

λarb OTU+ λi λi ε λsyst

Wavelength Selectable/Tunable Source

Network Control & Management (NC&M)

amplifiergain section tuning mirror modulator(transfer tapers)

(dilation taper) (dilation tapers)

Tuning Curves

1548

1549

1550

1551

1552

1553

1554

1555

1556

0 1 10 100Tuning Current [m A]

Wav

elen

gth

[n

m]

30

35

40

45

50

55

60

SM

SR

[dB

]

24C 20C

operating point

4 more channels with16, 28, 32C temp. tuning

Wavelength-Selectable Laser Modulatesup to 20 Channels

• Tunable filter, all InP device• Performance demonstrated,

20 λ x 50GHz≥ 0dBm fibered power

• EA-DBR is primary path to wavelength selectable functionality

• Essential Network Element-Analogousto Digital Cross-Connect

• Hub Node on a Ring• Mesh Networks

NetworkControl

1 1 1

2 2 2

3 3 3

1

2

3

3 2 1

2 1 0

1 3 3

1

3

2

Drop/Add

• • •

Wavelength Express/Local Routing at a Branch:WDM Optical Cross-Connect

I/O Fibers

Imaging Lenses

Reflector

MEMS 2-axis Tilt Mirrors

• True optical switch fabric

• Scalable to >1000 x 1000 ports

• Strictly non-blocking

• Wavelength, data-rate, and protocol independent

• <10ms mirror switching speed

Optical Switching Advancements Scalable Optical Cross-Connect with MEMS

Lucent ProprietaryUse According to Company Instructions

Optical NetworkNavigator

The Optical Data Network

DeviceServer

DeviceServer

DeviceServer

Service Provider Servlet User Feature Applet

Session Coordinator

Optics and Software- Better Together!!

• Intelligent Control in Line Systems is Needed to Continue Reducing Cost/Bit(Smart Amplifiers, DGE, Transient Control, Raman Control,

• Automated Control Plane Critical to Extract the Full Value Proposition of Optical Networks

• IP and Optical- We Need to Understand the Convergence!

WDM Applications Evolution

Residential Access(WDM PON)

LAN(Interconnect, Routing)

Long Distance

Metro/Business Access

Cable TV

High SpeedSwitch/Router Optical/Wireless

WDM Metro Ring: Potential Features• Ultra-High Capacity Provided by DWDM• Per Wavelength Protection• Provide Access (Hubbed) or Inter-Office (Mesh-like) Connectivity• Wavelength Channels Provide Format Independence• Provide Direct Connection to High-Speed IP Router Interfaces

WDM Fiber Ring Networking

OpticalA

/DO

ptic

alA

/D

OpticalX-

connectATM

switchSONET

A/D

OpticalA/D

OpticalA/D

OpticalA/D

OpticalA/D

IP Router

IP Router

ATMswitch

SONETA/D

Integrated Fully Reconfigurable Wavelength Add/Drop

Thermooptic phase shifters

Waveguide gratings with interleave chirp

10 cm

1544 1546 1548 1550 1552 1554 1556-50

-40

-30

-20

-10

0

Wavelength (nm)

Tran

smis

sivi

ty (d

B)

• Compact

• Low insertion loss

Line

Add

Through

Drop

Total on-chip loss < 16 dBInter-channel crosstalk < -20 dBDimensions 8x12 mm2

Total on-chip loss < 16 dBInter-channel crosstalk < -20 dBDimensions 8x12 mm2

OXC1

2

1

2

in out

XX

XX

4 λ 2x2 OXC with Dilated Switches

M. Smit

Does Fast Optical Switching Have a Role?

Logical Architecture of Multi-Shelf Switches

• Data are framed into 64-byte envelopes and transmitted to fabric– Small envelopes can lead to low latency for small packets

• Fabric stores data in Virtual Output Queues and switches through the cross-bar• Fragmentation effect due to variable size packets

– IP packets are not always integral multiples of 64-byte envelopes– Speed up required

• Power consumption is high– Double laser receivers/transmitters– Buffers on the fabric– Electronic crossbar

LocalBuffers

Scheduling

LocalBuffers

Scheduling

Crossbar

LocalBuffers(VOQ)

LocalBuffers(VOQ)

Scheduler

Switch Fabric System19’’ x 36’’ , >2KW power, expensive high-speed backplane

Laser

Framer

Framer Framer

Framer

Line Card Line Card

Line Card Line Card

Laser

Laser

Laser

Laser

Laser

Laser

Laser

What if we could design an optical packet fa

TunableTransmitter

AWG

Scheduler

TunableTransmitter

TunableTransmitter

Burst-ModeReceiver

Burst-ModeReceiver

Burst-ModeReceiver

32 In

put L

ine

Car

ds

32 Output Line C

ards

0 10 20 30 40 500

10

20

30

40

50

60

Num

ber o

f ocu

rren

ces

Switch time (ns)

High-Speed Packet Optical Switch Fabric

Totally passive core Laser switching time < 50 ns

Based on commodity components Phase recover within a few ns

Broadband Access -Is Fiber-to-the Home Necessary?

Can We Afford it?

OperationsComplex

Passive

λ 3

λ N

λ Router &Splitter

ONU

λ 2

λ 1RPRP

ONU

ONU

SmallBusiness

Control

Distr.Intfc.

Inter-conn.

Ntwk.Intfc.

Host TerminalComplex

ServicesComplex

Vision Distribution Architecture

Photograph of SiOB Optical Transceiver Chip

1.3 / 1.5 um

1.3 um

1.5 um

1.3umMONITOR

PIN

1.3umRECEIVER

PIN1.3um

XB-LASER

1.5um REVERSINGELEMENT

1.3um / 1.5umMACH-ZENDER WDM

INTEGRATED V-GROOVEfor Passive

FIBER ALIGNMENT

1.3umSPLITER

Gates et al., ECTC (1998)

What Will Be the Next Big Thing?

I Don’t Know, But I Do Know There Will Be One!!!

SiliconSilicon substrateSilica

Thermally grown silicaSample surface

SOI Based Photonic Crystal Waveguides

J. Arentoft and M. Kristensen

NanophotonicsResonantly Enhanced Nonlinearity

ChalcogenideResonator Cavity

Cavity round trips = 30Nonlinearity = x1000!

Small cavity

Large bandwidth

4 mm dia.Disk

Planar Waveguide

Resonantly Enhanced Interferometer

1990 1995 2000 2005 2010 2015 2020Predicted date -to-market

Waveguide in Ultra Compact Technology

1 µm1 µmVery CompactTechnology prototype(COBRA, 2000)

1

10

100

1000

SSI

MSI

LSI

Inte

grat

ion

Scal

e1.5 mm

Compact Technology

Very Compact Technology

Ultra Compact Technology

Compact Technology prototype(COBRA, 1999)

2 mm2 mm

M. Smit

Summary

• The Rapid Evolution and Deployment of Optical Networking Systems Has Been Driven by Innovation in Fiber, Optical Components and Transmission Systems Techniques

• Scalable Optical Networks Employing Optical Switching and Routing Require Highly Functional and Dense Optical Components and Circuits that Must be Cost-Effective

• Efforts to Cost, Size and Power Reduce Optical Subsystems is Ongoing. Integration Technologies-both Hybrid and Monolithic Will Be Key

• Increasing the Domain of Application of Fiber Optics to Increase Component Volumes is Essential to Continue to Drive Down Cost