Optimizing Long Haul and Subsea

Connections
Thursday, February 18, 2021
Agenda
• Introduction
• Simon Stanley, Analyst at Large, Heavy Reading
• Service Provider Keynote
• Steve Grubb, Global Optical Architecture, Facebook
• Sponsor Presentation
• Geoff Bennett – Director, Solutions and Technology, Infinera
• Panel Discussion
• Geoff Bennett – Director, Solutions and Technology, Infinera
• Matt Adams, Senior Product Manager, Viavi Solutions
• Sushin Suresan, Senior Product Manager, Cisco
• Steve Grubb, Global Optical Architecture, Facebook
• Q&A
Optical Networking 400G+ Revenues
Growing Network Bandwidth
Channel Capacity
Symbol Rate Bits per Symbol
Bits per Error
96 Gbaud
Modulation 1200G
Symbol Margin
64QAM 6 0.1
800G
16QAM 4 0.23
64 Gbaud QPSK 2 0.71 400G
BPSK 1 1
OOK 1 0.5 200G O-band E-band S C L

32 Gbaud 100G 1300 1400 1500 1600

Wavelength (nm)
40G
10G
Spectrum
Single Fiber C-band C+L S+C+L O+E+S+C+L

Multple • Network Bandwidth

Fibers
• Channel capacity
• Number of channels per fiber
Multi-core
Fibers • Number of fibers
Space
Source: Earlswood Marketing
The Shannon Limit
• Channel capacity
64-QAM

Achievable Rate (bits/symbol)

• The theoretical maximum data 6
rate over a communications
medium in the presence of noise
• Developed by Claude Shannon 16-QAM
4
and outlined in a paper written in
1948
QPSK
• Established the theory of Forward 2
Error Correction (FEC) to correct
for errors due to noise
• The number of redundant
symbols needed to enable the 0 25
correction of errors depends on
SNR (dB)
Theoretical Transmission Rates with AGWN
the channel conditions
Coherent Optical Developments
• 10 years of development and • 600G Coherent - 2017
deployments • 64 Gbaud, 64-QAM
• QPSK ➔ 64QAM • 38.4 Tbit/s using extended C-band
• 32 Gbaud ➔ >90 Gbaud • 64 channels, 75 GHz
• 100G ➔ 800G per optical channel • 76.8 Tbit/s using C+L bands
• 100G Coherent - 2010 • 800G Coherent - 2019/2020
• 32 Gbaud, QPSK • First solutions for DCI
• 8 Tbit/s using C-band DWDM • >90 Gbaud, 64-QAM
• 4THz - 80 channels, 50 GHz • 400G Pluggables - 2020/21
• 200G Coherent – 2013/14 • Based on 7nm DSPs
• 400G Coherent - 2015 • CFP2-DCO
• Flex-rate 100/200/300/400G • 400ZR and ZR+
• QPSK/8QAM/16QAM • QSFP-DD and OSFP
Open Optical Networking and IPoDWDM
Open Optical Networking SDN-based
Network Management

Switch/ Switch/
Router Transponder A Transponder C Router
Open 80-6000km Open
Switch/ Mux/ Line Line Mux/ Switch/
Transponder B System Amp Transponder A
Router Amp System Router
(OLS) (OLS)
Transponder C Transponder B
Switch/ Switch/
Router Router

IPoDWDM with 400ZR Pluggable Modules

Switch/ SDN-based Switch/

Router Network Management Router
QSFP-DD/OSFP
400ZR

Switch/ OLS OLS Switch/

Router Router
40-120km
Switch/
(more with ZR+) Switch/
Router Router
Keynote:
Steve Grubb
Global Optical Architecture, Facebook
Keynote:
Geoff Bennett
Director, Solutions and Technology, Infinera
Optimizing Longhaul
and Subsea Capacity
Geoff Bennett
Director, Solutions and Technology, Infinera

February 2021
Theoretical capacity limit: Shannon Equation
There’s a non-linear limit to signal power
The “bandwidth term” The “SNR term”

Fiber capacity C = B log2 ( 1 +

S
N
)
How much of the fiber
spectrum can we use? Better fiber: Today  Large Area, Low Loss

Expand the C-Band 4 THz → 4.8 THz → 6 THz

Better line systems – e.g. Raman vs EDFA
C+L Band “Double” capacity at LH reach

“Exotic” fiber FUTURE: Hollow core, ZBLAN

Better transponders:
More fibers SDM in Submarine Networks 1st→2nd→3rd→4th→5th Generation
12
Transponder Evolution Challenge

The “bandwidth term” The “SNR term”

Fiber capacity C = B log2 ( 1 +

S
N
) The optical performance benefit
from better SNR suffers from
diminishing returns
Bandwidth term

Capacity
Should the focus be on the
Bandwidth Term?
Log term

Capability
13
 We also need to consider the scale of upgrades
Expand the C-Band 4 THz → 4.8 THz → 6 THz
Better fiber – e.g. Large Area, Low Loss
C+L Band “Double” capacity at LH reach

“Exotic” fiber FUTURE: Hollow core, ZBLAN

Better line systems – e.g. Raman vs EDFA
More fibers SDM in Submarine Networks

These options involve new fiber deployments or line system upgrades

But new transponders can be Better transponders:

1st→2nd→3rd→4th→5th Generation
deployed quickly and easily
14
 Infinera
ICE6 ICE6 Performance
Program Update

METRO/LONG-HAUL SUBMARINE
NA Operator 800G over 950KM MAREA 28T 60% Hardware
G.652 SMF-28 FIBER 6,644 KM 650Gb/s Reduction

800G over 730KM

Pacific 5.4T 50% Capacity
700G over 1,460KM 200Gb/s Improvement
G.652 SMF-28 FIBER 9,125 KM Dispersion Managed FIBER

800G over 667KM

600G over 2,283KM
G.655 LEAF FIBER

ADDITIONAL PERFORMANCE ENHANCEMENTS HAVE

OCCURRED SINCE THESE TRIALS ON PATH TO PRODUCTION

This 5th Generation transponder is exceeding

performance expectations at all data rates
15
 Thank you!

gbennett@infinera.com
Panel Session
Panelists

Simon Stanley Steve Grubb

Heavy Reading Facebook
Analyst at Large Global Optical Architecture

Geoff Bennett Matt Adams Sushin Suresan

Infinera VIAVI Solutions Cisco
Director, Solutions and Senior Product Manager Senior Product Manager
Technology
Panel Topics
• Driving capacity and spectral efficiency at the edge of the
Shannon Limit
• Adoption of C+L band, multiband transmission, spatial
multiplexing
• Open optical networking for long haul and subsea applications
• The role of pluggable modules in long haul networks
• Q&A
Driving capacity and spectral efficiency
at the edge of the Shannon Limit
• What is the practical channel capacity today for terrestrial and
subsea?

• What are the next steps in increasing spectral efficiency and

channel capacity?

• What are the challenges in deploying and testing networks that

are near channel capacity?
Adoption of C+L band, multiband
transmission, spatial multiplexing
• How widespread is the adoption of C+L band in deployments
today?

• Do you see a big role for multi-band transmission and if so, in

what timescale?

• Which approaches to spatial multiplexing will have the biggest

impact?
Open optical networking for long
haul and subsea applications
• How important is open optical networking for terrestrial and
subsea applications?

• What are the key challenges in deploying and testing open optical
networking for terrestrial and subsea applications?
The role of pluggable modules in
long haul networks
• What do you see as the role of pluggable modules in long haul
networks

• How will the architectures of metro and long haul networks

change to take advantage of 400ZR and ZR+?
Q&A
Please submit your questions for the panel
15 min. Break
Session 2 will begin at 11:45 a.m. ET