High Capacity Optical Access Networks

High Capacity Optical Access Networks
Ali@ua.pt
Ali Shahpari
Department of Electronics, Telecommunications and Informatics, University of
Aveiro and Instituto de Telecomunicações, Campus Universitário de Santiago,
3810-193 Aveiro, Portugal.

Outline
• Motivation
• Access Capacity Motivation
• Coherent UDWDM PONs
• Terabit+ PON
• High-Capacity PON
• Crosstalk Mitigation
• Real-time Coherent PON
• Field-Trial
• Flexible Optical Access Networks
• Photonic Integrated Circuits in PON
• Starting field NGPON2
• Packaging and Test
• Conclusions
18-19 May, CpqD, Campinas - SP

Access Capacity Motivation
Mobile
Backhaul/
Front-haul
Higher Data
Rates per
Device or Ap.
Increasing #
Device per
Subscriber
Business
Subscribers
Growing
Residential
Subscriber
Growing
C. Knittle, “IEEE 100 Gb/s EPON” OFC 2016.
Source : Cisco VNI

Context
Video enabler solution
- Best wavelength band, with small
constrains Challenge: start
fiber deployment/
adoption

Context
GPON solution
- Target low cost /reasonable bandwidth to
compete with copper Challenge: get the
volume to lower
prices

Context
Trying to get further bandwidth with the
same principles of GPON (US in low
dispersion)
Challenge: good
cheap lasers at 10G

Context
Increasing substantially the bandwidth and
adding flexibility
Challenge: Good
slightly tunable
lasers and receivers
@10G/2.5

Context
A. Shahpari et al, “Multiple System Configuration for Next Generation Optical Access Networks
with Real-Time Nyquist UDWDM-PON”, ECOC2015, P7.18
Adding the extra flexibility and global
control.
Challenge: Good
tunable lasers and
receivers

Context
ITU-T recommendation G.989.2 (draft), April 2014.
threat
opportunity
Spectrum in optical access after NG-PON2
Future optical access networks will target:
 Higher data rate per user
 Spectral efficiency
 High number of user per ODN
 Extended reach
 Flexible network

Current technologies
Essential (now) Optional (near future)
Low Cost Wide tunability
Slight tunability High ODN loss tolerance
Tight control of wavelength >10Gbit/s rate
10Gbit/s rate High spectral density
GPON
XGPON
NGPON2
DWDM+Coherent+
Advanced modulation formats
(Core+metro)

Coherent Brings Inherently
• PON architecture compatibility
• Compatibility with higher order modulation formats
• Filtering
• Tight control
• Depends on LO
• Tunability
• Depends on the LO
• High ODN is due to the coherent inherent gain
• Compatibility with digital signal processing
Essential (now) Optional (near future)
Low Cost Wide tunability
Slight tunability High ODN loss tolerance
Tight control of wavelength >10Gbit/s rate
10Gbit/s rate High spectral density

Can we do this in PON architecture and at potential
low cost?
• Several groups believed this and started looking for
solutions
• Nokia Siemens Networks
• 2009
• Following
• Instituto de Telecomunicações, PT
• Barcelona, SP
• Pisa, IT
• Torino, IT
• UCL, UK
• …
• (only some example recent research is listed here and is for
indication only, clearly not complete)
Multichannel
generation in shared
resources with low
cost laser
Ultra high bandwidth
and ultimate flexibility
Simplified coherent
detection with direct
modulationCombined
coding/modulation
and reception
schemes
Carrier reuse and self
homodyne schemes

Where could we enhance performance of coUDWDM
• General DSP (single polarization & intradyne detection)
• Hybrid bidirectional coherent PON system with video overlay.
2sps
2sps
1sps
1sps
1sps
Ix
Qx
1sps
2sps
Ix
Qx
Analogsystem
2sps
NyquistShaped
Pre-Emphasis
Normalization
ClockRecovery
FrequencyRecovery
PhaseRecovery
SymbolDecoding
SymbolCoding
Tx DSP Rx DSP
101100... 101100...

Nyquist UDWDM-PON
OLT ONU
• Digital Frequency up/down-shifting:
• Reduces the impact of Raleigh Back-
Scattering;
• Reduces the impact of dynamic Stimulated
Raman Scattering.
• Carrier reuse:
• Simplifies ONU;
• Simplify dynamic mechanisms to track
wavelength.

Bidirectional UDWDM PON
1545.15 1545.2 1545.25 1545.3 1545.35 1545.4 1545.45
-70
-65
-60
-55
-50
-45
-40
-35
-30
Wavelength [nm]
OpticalPower[dBm]
DS Channels
US Channels
• DS and US Spectra

Terabit+ PON
OSA
40 km SSMF
Coherent Rx
Digital
Real-Time
Oscilloscope
4x90º
Hyb.
ECL
BD
BD
PLΔf/2
MZM VOA
W
D
M
100 GHz
1549.98 1550.08 1550.18 1550.28
70
65
60
55
50
45
40
35
30
Wavelength [nm]
OpticalPower[dBm]
Bi directional 1.25 Gbaud UDWDM at 2.5 GHz
Upstream
Downstream
UDWDM over DWDM
IQM
Δf
MZM
ECL
AWG
π/22π100 kHz
16x(NchxRS Gbaud@Δf GHz) @ 100 GHz
WSDFB
DFB
.
.
.
I Q
RS Gbaud
W
D
M
100 GHz
PC
• Bi-directional Nyquist Shaped 16QAM UDWDM over DWDM (100 GHz)

EVM: 10 Gb/s per channel
• Sensitivity-32 dBm (single), -27 dBm (UDWDM)
A. Shahpari et al, “Terabit+ (192 x 10 Gb/s) Nyquist shaped UDWDM coherent PON with upstream and downstream over a 12.8 nm
band” OFC’13.

Self-Homodyne PON
ONU
OLT
ECL
MZM
4 GHz
PTVM
IQSignal
PT
PBS
AWG
CORX
SignalLO
BPFFP
CORX
Signal
PT
PBS
IQAWG
30 km
SSMF
ECL – External cavity laser, 100 kHz
AWG – Arbitrary waveform generator
MZM – Mach-Zehnder modulator
PTVM – Pilot tone vector modulator1
BPF – Band-pass filter
FP – Fabry-Perot filter
Downstream signal
Upstream signal
DS (Pol. X+ Pol. Y)
and US Spectra
• Single laser for US and DS
• Comb generator – 12 tones@4 GHz
• Prototype PTVM to generate signal +
pol. mux. pilot tone
• 2.5 Gbaud 16QAM with raised cosine
pulse shape and 0.05 roll-off, upshifted
(DS) or downshifted (US) 1.5 GHz
• Signal and PT separation by PBS at
ONU
• 2-stage PT filtering at ONU (BPF+FP)
A. Shahpari et al, “Fully coherent self- homodyne bi-directional enhanced performance PON” OFC’14.

20 Tbit/s PON
ONU
OLT
ECL
DFB
DFB
×83
MZM
4 GHz
PTVM
IQSignal
PT
PBS
AWG
CORX
SignalLO
BPFFP
CORX
Signal
PT
PBS
IQAWG
30 km
SSMF
• 1008 channels UDWDM PON
• Burst-mode operation
• 33 nm spectral occupation
burst-mode signal
0.82ms
R. Luis and A. Shahpari el at, “Ultra high capacity self-homodyne PON With simplified ONU and burst-mode upstream” PTL’14.

Field-Trial Network
 First real-time experimental demonstration of digital Nyquist
coherent UDWDM-PON using FPGA-based 8-bit DSP:
 OLT transmitter (tuneable) in real-time;
 ONU receiver (intradyne) in real-time.
• FPGA Virtex-6 (ML605 board)
• Hardware parallelization of 16 (156.25 MHz)
• 2.5 Gsa/s DACs
• 14-bit 4DSP FMC230 @ 1.4 GHz
 FPGA Virtex-7 (VC707 board)
• Hardware parallelization of 16 (156.25 MHz)
 2.5 Gsa/s ADCs
• 8-bit 4DSP FMC125 @ 2 GHz
R. M. Ferreira et al, “Field-trial of a real-time bidirectional UDWDM-PON coexisting with GPON, RF video
overlay and NG-PON2 systems,” ECOC’15.

• IQM driven by the 2.5 Gb/s DQPSK real-time electrical
Tx1
• 4×16 UDWDM channels emulated by a comb
generator (2.5 GHz of channel spacing) and 4 ECLs
(50 GHz of frequency spacing – λ2 at ~1545 nm)
• 8.4 km of field deployed fiber
• 50 km of SSMF in lab
• Coherent detection (4×90o optical hybrid) with a free-
running ECL local oscillator
• Electrical 1 GHz low-pass filter
• Real-time DSP applied in the Rx1
Field-Trial Setup – DS UDWDM
EDFA
OLT UDWDM
1 GHz
ECL λ2
ONUs UDWDM
BD
BD
Rx1
Tx1
4.2 km
(field)
4.2 km
(field)
10 km
(lab)
40 km
(lab)
UDWDM UDWDM
UDWDM
B
A A AA
4×90o
Hybrid
IQM
...
1:42:4
ECL λ1
ECL λ2
ECL λ3
ECL λ4
...
WSComb
λ2-2.5GHz λ2
. . .
λ2+2.5GHz
. . .
16 channels

Field-Trial Setup
EDFA
BD
BD
OLT UDWDM
EDFA
1 GHz
ECL λ2
ONUs UDWDM
BD
BD
ECL λ2+1.25G
Tx2
Rx1
Tx1
Rx2
4.2 km
(field)
4.2 km
(field)
10 km
(lab)
40 km
(lab)
UDWDM UDWDM
UDWDM
NGPON2
GPON
Video
IQM
B
A A AA
ONU
GPON /
RF Video Overlay
OLT
XFPch1
GPON
XFPch2
XFPch3
W
M
XFPch4
ONU
NGPON2
RF Video Overlay
Head End
ECL λ2+3.75G
ECL λ2-1.25G
4×90o
Hybrid
IQM
CEx
...
1:42:4
ECL λ1
ECL λ2
ECL λ3
ECL λ4
...
GPON /
NGPON2
WSComb
ECL λ2+1.25G
4×90o
Hybrid
DS spectrum
NG-PON2
RF Video
Overlay
GPON

• DS sensitivity: ~ -44.5 dBm
 Similar results have been observed for the remaining three sets of 16
UDWDM channels (λ1, λ3 and λ4)
• US sensitivity: ~ -43.5 dBm
 Penalty of ~1 dB due to DS back reflections
Receiver Sensitivity (DS and US) – Field Deployed Fiber
At the BER limit, a stable performance
was observed for several hours!
Transmitted power per
channel: -4 dBm
DS channel 1
DS channel 8
DS channel 16
A. Shahpari et al, “Real-time bidirectional coherent Nyquist UDWDM-PON coexisting with multiple deployed
systems in field-trial (Post-Deadline),” JLT’16.

Real-Time Flexible Heterogeneous UDWDM
System for Coherent PON
• Receiver sensitivity for single- and multi-channel for QPSK, 8PSK and 8QAM
R. M. Ferreira et al, “Real-Time Flexible Heterogeneous UDWDM System for Coherent PON,” submitted to ECOC’16.

Heterogeneous UDWDM Configurations
• Required received power vs. transmitted power per channel to
keep the BER at 3.8×10-3.
R. M. Ferreira et al, “Real-Time Flexible Heterogeneous UDWDM System for Coherent PON,” submitted to ECOC’16.

• ECL sensitivity (single channel): -49.5 dBm
• ECL sensitivity (UDWDM): -47.5 dBm
• DFB sensitivity (single channel): -46.5 dBm
• DFB sensitivity (UDWDM): -44 dBm
Receiver Sensitivity for ECL vs. DFB
8 of 11
 16 UDWDM channels
 2.5 GHz channel spacing
 ECL: 100 KHz-linewidth
 DFB: 20 MHz-linewidth
A. Shahpari et al, “Coherent Access (Invited paper),” to be submitted JLT’16.

• ECL sensitivity (single channel): -49.5 dBm
• ECL sensitivity (UDWDM): -47.5 dBm
• DFB sensitivity (single channel): -46.5 dBm
• DFB sensitivity (UDWDM): -44 dBm
Frequency Shift for ECL vs. DFB
8 of 11
 16 UDWDM channels
 2.5 GHz channel spacing
 ECL: 100 KHz-linewidth
 DFB: 20 MHz-linewidth
ECLDFB
A. Shahpari et al, “Coherent Access (Invited paper),” to be submitted JLT’16.

Flexible Optical Metro Networks Based on Adaptive
Stokes Space Polarization Demultiplexing
• CMA effectively be employed in DSP: but data format dependent operation
and being liable to suffer from singularity issues when operating in blind mode.
• Stokes space PolDemux: enabling higher transparency, robustness, immunity
to phase noise and singularity issues as well as allowing a faster convergence.
S. Ziaie et al, “On the Benefits of Flexible Optical Metro Networks Based on Adaptive Stokes Space Polarization
Demultiplexing,” submitted to ECOC’16.

Flexible Heterogeneous UDWDM DP-QPSK and DP-
16QAM using Stokes PolDemux
-48 -46 -44 -42 -40 -38 -36 -34 -32 -30 -28 -26
10
-5
10
-4
10
-3
10
-2
10
-1
Power, dBm
BER
BER=3.8x10-3
BER=1.5x10-2
BTB
80km
Field+ 40 km
80km-FSO
S. Ziaie et al, “On the Benefits of Flexible Optical Metro Networks Based on Adaptive Stokes Space Polarization
Demultiplexing,” submitted to ECOC’16.
• Sensitivity Evaluation for 12.5 Gbps DP-QPSK and 25 Gbps DP-16QAM

In the Integration World
Electronics
Vacuum
tubes
Transistor ICs
Optics
Free space components
Single
components
packaged
PICs

Why PIC?
• Lasers, modulators, receptors and amplifiers ... In a single
chip
• Simplifies the design of the optical systems
• Decreases the complexity of the electronic control
• SWaP: Size, Weight and Power are lower
• All-in-one package

PICadvanced – Photonic Integration Target
Advanced
photonics
design
Chip design in several technologies
Specific ONU/ONT advanced design
Advanced
packaging
Flip Chip approach
Simplified chip alignment
BOSA compatibility
Services
Advanced RF-to-chip board
design and implementation
• PICadvanced
• Aims at bridging Photonic Integration and applications.
• Starting field NGPON2
• Accelerating advanced solutions in optics and photonics to
market
• Starting by advanced BOSA’s and PIC’s
*Confidential - for Use within the persons engaged in the NDA with PICadvanced

Electronics Assembly (partner)
Boards and Components assembled in 2014
Number Boards
assembled: 3,6 Millions
Number of Components
assembled: 72 Millions
(max 300k comp/hour)
Number of empleyees:
220
SMT THT Tests PACKINGSTORAGE

Discrete Components
• compliant with the next generation (NGPON2)
standards
• One of the first in the market
• Highly competitive with a tunable receiver (key
component)
• Solutions for all scenarios:
• Class N1 or N2, Types A or B
• 2.5Gbit/s or 10Gbit/s Tx
• High power
• 10Gbit/s or 2.5Gbit/s Rx
• High Sensitivity
• Small Form Factor for OLT already available in XFP
• BOSA On Board Solutions

Our future Looks Like this
• 20-30% more space for Electronics
• Increased complexity devices
• Much simpler electrical connections to optics
PIC

Experience in Chip
SMART Photonics
 OLT/ONU (4λ Tx + 4λ Rx)
»Already packaged
and tested
»Doesn’t fulfill NG-
PON2
wavelengths and
power
specifications
 ONU and test components
»Already packaged
and tested
»Doesn’t fulfill NG-
PON2
wavelengths and
power
specifications
FhG-HHI
 Band splitter and receiver
Band splitter

Full Process Control
Design
Production
Packaging

AWG band splitting + Balanced PIN
AWG
Balanced PIN
PIN Photodiode
MMI 2x2
IN1
OUT1
OUT2

Input and Outputs

Conclusion
WAN
Mobile Backhaul
Optical Access
QPSK
……
16QAM
QPSK
8QAM
8QPSK
8QAM
QPSK
8QAM
Data Center
• Flexible, spectral efficient
and high performance
technologies, real-time
demo, network
optimization
• In Instituto de Telecomunicações (IT): we are developing new
coherent system technologies for metro and access networks.

Conclusion
WAN
Mobile Backhaul
Optical Access
QPSK
……
16QAM
QPSK
8QAM
8QPSK
8QAM
QPSK
8QAM
Data Center
• Low cost and high
performance Tunable
OLT/ONU modules
• In PICadvanced: low cost, size and power consumption SFP+
modules.

Conclusion
WAN
Mobile Backhaul
Optical Access
QPSK
……
16QAM
QPSK
8QAM
8QPSK
8QAM
QPSK
8QAM
Data Center
• Low cost
Tunable
XFP/SFP+
modules
• In PICadvanced: low cost, size and power consumption
XFP/SFP+ modules.

Conclusion
WAN
Mobile Backhaul
Optical Access
QPSK
……
16QAM
QPSK
8QAM
8QPSK
8QAM
QPSK
8QAM
Data Center
• PIC will bring the required speed, tunability and sensitivity for
future coherent PON system with lower SWaP and cost.

Acknowledgements
• Funding
• FutPON
• PICZR
• COPRESS
Thank you!
PICadvanced.com

High Capacity Optical Access Networks

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Ähnlich wie High Capacity Optical Access Networks

Ähnlich wie High Capacity Optical Access Networks (20)

Mehr von CPqD

Mehr von CPqD (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

High Capacity Optical Access Networks