Optical line technologies for rates
above 100G
Pete Anslow
28 May 2010
Line vs Client
The requirements for optical line technology are different from those on
the client interface.
• Long distance transport made economic by sharing of optical line
infrastructure (amplifiers, ROADMs etc.) across many channels.
• Spectral efficiency (bit/s / Hz) and reach before needing O-E-O
regeneration are key parameters.
Client
Line
1000 km
10 km
2
Three ways of capacity evolution
QPSK
increase
Constellational multiplicity
Amplitude
Phase
Polarization
• Constellational multiplicity
• Subcarrier multiplicity
• Symbol rate
Symbol rate
f
freq
3
r
i e ty
r
ar lici
c
b p
su ulti
m
Constellations
Modulation formats on each optical carrier

(N)RZ ASK
Modulation
Format
4
4
Bits/Baud
Baud Rate
(GBaud)
for
450Gbit/s
Baud Rate
(GBaud)
for
1120Gbit/s
(N)RZ IM
1
450
1120
(D)BPSK
1
450
1120
(D)QPSK
2
225
560
DP-QPSK
4
112.5
280
DP-8QAM
6
75
186.7
DP-8PSK
6
75
186.7
DP-16QAM
8
56.25
140
(D)BPSK
DP-(D)BPSK
(D)QPSK
DP-(D)QPSK
X-Pol
Y-Pol
DP-8QAM
DP-8PSK
DP-16QAM
560Gbit/s CoFDM DP-16QAM experiment
5
560 Gb/s CoFDM DP-16QAM
5
Advanced
Modulation
Formats
0
Relative Amplitude [dB]
-5
-10
Relative Amplitude [dB]
0
-5
Channel
Spacing
100 GHz
-10
-15
-20
-25
-15
-20
-30
191.5
192
192.5
-25
-30
-35
193.55
193.6
193.65
193.7
193.75
Fre que ncy [THz]
193.8
193.85
193.9
Test Channel
Transmitter
WDM
Transmitter
34 x DP-QPSK 140 Gb/s
5
Link
193
193.5
194
Fre que ncy [THz]
194.5
Demux
Filter
195
195.5
Coherent
Receiver
0
Relative Amplitude [dB]
-5
-10
-15
-20
5 spans x 80km/span NDSF
-25
-30
-35
191.5
5
192
192.5
193
193.5
194
Fre que ncy [THz]
194.5
195
195.5
Data Capture &
Post- Processing
560 Gb/s
5
0
-5
-15
-20
-25
-30
193.3
Y-Pol. Out
193.6
193.7
193.8
Fre que ncy [THz]
2
2
2
2
0
Imaginary
4
Imaginary
4
0
0
-2
-2
-2
-4
-4
-4
-4
0
Real
2
4
-4
-2
[Charles Laperle Oct 2009]
0
Real
194
194.1
0
-2
-2
193.9
Y-Pol. Out
X-Pol. Out
4
Imaginary
Imaginary
193.5
4
-4
6
193.4
5 x 80 km = 400 km NDSF
Back-to-Back
X-Pol. Out
100 Gb/s
-10
560 Gb/s
Relative Amplitude [dB]
•Dual Carrier DP-16-QAM
•100 GHz WSS
•35 Gbaud Nyquist Generation
•SiGe BiCMOS DACs
•215-1 pattern
2
4
-4
-2
0
Real
2
4
-4
-2
BER = 0.019
0
Real
2
4
Recently reported results
7
Format
Spectral
efficiency
bit/s / Hz
Reach
km
Published
224
PDM 16-QAM
4
1,200
OFC 2010 PDPB8
448
CO-OFDM 16-QAM
5.2
2000
OFC 2010 PDPC2
560
CoFDM DP-16QAM
5
400
OFC 2010 Workshop
640
RZ Conjugation
0.25
100
OFC 2010 PDPC6
1.08
CO-OFDM
3.3
600
OFC 2009 PDPC1
1.21
PDM-OFDM QPSK
3.3
400
OFC 2009 PDPC2
Rate
Gbit/s
Spectral efficiency evolution
Recently deployed optical line technology has achieved considerable
improvement in spectral efficiency with each successive generation:
10 Gbit/s channels on 50 GHz grid – 0.2 bit/s / Hz
40 Gbit/s channels on 50 GHz grid – 0.8 bit/s / Hz
100 Gbit/s channels on 50 GHz grid – 2.0 bit/s / Hz
Recent WDM experiments
Reach (km)
10000
• 450 Gbit/s on 100 GHz grid – 4 bit/s / Hz
probably achievable with acceptable reach.
1000
• 1Tbit/s channels probably won’t see
improved spectral efficiency over this
unless reach is significantly compromised
100
0
8
2
4
6
8
Spectral efficiency (bit/s / Hz)
Summary
To achieve 400G or 1T channels for long distance transport with
acceptable reach and spectral efficiency, advanced modulation formats
are required that involve combinations of:
• complex constellations
• high symbol rate
• multiple sub-carriers
Experimental results that demonstrate feasibility of 400G and 1T channels
is emerging.
1 T channels will require multi-carrier technology and may not show
significantly better spectral efficiency than 400G channels depending on
reach. If this proves to be the case then the benefit of going to 1T is less
compelling.
9
Thank You
10
© Ciena Confidential and Proprietary