Submarine adopts 40G and 100G - North American … · © Ciena Confidential and Proprietary...

© Ciena Confidential and Proprietary

Submarine adopts 40G and 100G

Per HansenSr. Dir. Submarine Solutions

NANOG50, Oct. 5, 2010

2 © Ciena Confidential and Proprietary

Capacity Demand is VoraciousHigh bandwidth-growth geographic areas

CAGR 2010-13: >100%: 80% - 100%: 60% - 80%: 50% - 60%

Trans-Atlantic

(31% / 24Tbps)ME/Passage to India

(36% / 6.2Tbps)Intra Asia

(47% / 21Tbps)

Trans-Pacific(31% / 15Tbps)

US-LA(43% / 12Tbps)

DBW by traffic route and country int. bandwidth (CAGR 2010-13)

Ref.: Global Bandwidth Forecast Service, Telegeography, 2010


Increased design, component complexity & costComponents need to operate at faster speeds -> Increased system costLimited supply of components -> Yield and Reliability issues

Inferior performance to 10G 75% reach lost at 40G, 90% lost at 100GAddition of RAMAN amplifiers and/or REGENs required (CAPEX)

Cannot operate over existing optimized networkReplace some ROADMs with Regens (CAPEX)Increase cost and time to upgrade

Increased sensitivity to chromatic dispersion16 x worse at 40G, 100 x worse at 100GAddition of compensator equipment, amplifiers (CAPEX) and more complex engineering (OPEX)

Increased sensitivity to Polarization Mode Dispersion 4 x worse at 40G, 10 x worse at 100G Rip and replace bad fiber, addition of compensator equipment (CAPEX), and complex eng. (OPEX)

1

2

3

4

5

Challenges Scaling TDM to Higher Data Rates


Spectral Optimization

Four mechanisms to grow capacity and increase spectral density

Challenge the baud rateChallenge the bit/symbolChallenge carrier spacingPolarization diversity

Bits

per

sym

bol

Symbols per second

QAM, M-ARY

freq

Df


Encoding Schemes

Re

Im

“1”“0”

Re

Im

“1”

“0”

90o

Re

Im

“1,1”

“0,0”

90o

45o

“1,0”

“0,1”

1 bit/symbol 1 bit/symbol 2 bits/symbol

Electrical Field

Intensity

Time Time Time

Electrical Field

Intensity

Electrical Field

Intensity

OOK BPSK QPSK


Enhancing Spectral Density while retaining propagation characteristics

Rx Data After DSPRx Data Before DSP

QPSK X - polarization

QPSK Y- polarization

(1,1)(0,1)

(0,0) (1,0)

QI

(1,1)(0,1)

(0,0) (1,0)

IQ

2xVertical

Polarization

Horizontal Polarization

Dual Polarization

Dual Polarization• Rx Data

before DSP

• Rx Data after DSP

BPSK

CoFDM

• Two subcarriers @ ~20 GHz spacing

• 50 GHz channel spacing

Modulation Format

Polarization Mux.

Sub-Carrier Mux.

Key requirements:Grid compatibilityPropagation characteristics like 10GCost efficient hardwareOperational simplicity


Benefits of Coherent Detection

Linear frequency translation from optical to baseband

Correction of impairmentsAllows DSP to process received signal, correcting:

Electrical transfer functions and path delaysOptical filteringChromatic dispersionPolarization statePolarization mode dispersion (PMD)Polarization dependent loss (PDL)Transients

Channel monitoringAccess to amplitude, phase & polarization information

Spectral selectivityCo-FDM; colorless optical network egress

Optical gainReceiver sensitivity


Transmission PropertiesTransmitter and wet plant

Low optical launch powerAccommodates high number of channels

Minimal deleterious effects on other channels (new and old)

Allows addition via monitor ports

Dispersion maps Tolerates poor maps relatively well

Typical maps with zero-dispersion near center of band has most challenged channel where noise performance in usually very good providing betterusage of total spectrum

Optimal map has high dispersion across band letting the receiver compensate for a high amount of dispersion at end

4,000 km8,000 km2,000 km

40Gb/s 2P-QPSK 4,000 km40Gb/s 2P-BPSK 8,000 km100Gb/s 2P-QPSK 2,000 km


Transmission propertiesReceiver

Adaptive receiverLower performance variation reduced margin requirements(tracking polarization, residual dispersion, and received power)

Continuous monitoring and reporting of optical layer metrics (e.g. actual chromatic dispersion and PMD)

Easy card installation and operational simplicity

8.0

8.5

9.0

9.5

10.0

10.5

11.0

0 180 360 540 720

Q [d

BQ]

Time [h]

30 6 129

Std. dev.: 0.03 dBQ

FEC limit


1535 1540 1545 1550 1555 1560 1565-50

-45

-40

-35

-30

-25

-20

-15

-10

-5

Wavelength(nm)

Pow

er (d

Bm

/0.1

nm

)

Rx Spectrum

1557.5 1558 1558.5 1559 1559.5 1560 1560.5-36

-34

-32

-30

-28

-26

-24

-22

-20

-18

-16

Wavelength(nm)

Pow

er (d

Bm

/0.1

nm

)

Rx Spectrum

40GDP-QPSK@50GHz

100GDP-

QPSKCoFDM

100G CoFDM DP-QPSK Submarine TrialUn-announced trial partner/cable (Gen 1): 2,100 km

Idlers

Maximum channel count: 48Full-fill capacity with 100G: 4.8 Tb/s


Deployment and Trial Milestones

2008

2009

2010

N x 40G WDM upgrade on repeatered submarine cable- 50 GHz spacing augmenting existing 10G wavelengths- Commercial operation

N x 40G WDM deployment unrepeatered submarine cable - Virgin Media (UK – Ireland: 238 km/58 dB)- Commercial operation

100G deployment on unregenerated terrestrial cable- Verizon network between Frankfurt and Paris (900 km )- 50 GHz channel spacing with existing 10Gs on link with ROADMs- Commercial operation

N x 40G WDM upgrade on repeatered submarine cable- 50 GHz spacing augmenting existing 10G wavelengths- Commercial operation- Certified N x 100G as future upgrade w/o reengineering on ~2,300 km

100G WDM upgrade trial on repeatered submarine cable- 50 GHz spacing augmenting existing 40G wavelengths- Link distance: ~2,000 km


A Look to the Future – Maximizing Spectral Utility

10Gps@100GHz

10.7Gps@50GHz

46Gps@50GHz

112 Gps@50GHz

224 Gps@50GHz

448 Gps @80GHz

1000 Gps @170GHz 25

20

15

10

5

0

5

4

3

2

1

0

6

1995 2000 2005 2010 Future…

Bits

per

sec

onds

per

her

tz

Tbit/

s in

C-b

and

Source: “100G and Beyond with Digital Coherent Signal Processing”, K. Roberts et.al., IEEE Communications Magazine, July 2010


Submarine Deployment TrendsUpgrades take center Stage

Lower marginal cost of bandwidth

Lower hardware and deployment cost

Lower operational cost

Maximize utilization of cable assets

1998 2000 2002 2004 2006 2008 2010 2012

5

10

15

20

25

Source:Telegeography

Number of cables entering service 1998 – 2012Capacity upgrades diminish new cable deployments

Source: Global Capacity Demand Drivers & Network Deployment Trends, A. Mauldin (Telegeography), SubOptic, May 2010

Cable RFS Year Capacity Increase

Atlantic Crossing-1 1998 700%

Australia-Japan Cable 2001 297%

Americas-II 2000 863%

Pacific Crossing-1 1999 150%

Southern Cross 2000 733%


Meshing the Submarine cables

Lower marginal cost of bandwidthMaximize utilization of cable bandwidth(shared protection across cables)

Lower maintenance cost and cost of downtime

Enhance value of servicesCompete on Quality of Services (QoS)

Expand the Classes of Service (CoS)

Higher-value application-specific offerings (e.g. international data center connectivity with latency SLAs)


PWR & CTRL

HybridSwitchCore

DWDMSLTE

DWDMTLTE Mesh

CoreSwitch

PWR & CTRL

MeshCore

Switch

DWDMS/LT

DWDMSLTE

DWDMTLTE

CrossConnect

MeshCore

Switch

PWR & CTRL

Landing Station PoP/CO

Landing Station and/or PoP/CO

The Global NetworkArchitecture evolution

Traditional ArchitectureDiscrete network components

Integrated ArchitectureSubmarine/terrestrial WDM integrationSimplified managementSpace and cost reduction

Global Mesh ArchitectureUniversal ULH Packet/TDM-OTSEnd-to-end provisioning & managementGlobal mesh protection & restorationDisassociation of wet plant

Submarine adopts 40G and 100G - North American … · © Ciena Confidential and Proprietary...

Documents

Transcript of Submarine adopts 40G and 100G - North American … · © Ciena Confidential and Proprietary...