1
High Spectral Efficiency 400G Transmission
Xiang Zhou
AT&T Labs-Research, Middletown, NJ 07748
Invited
Abstract — This paper presents an overview on the generation
and transmission of 450Gb/s wavelength-division multiplexed
(WDM) channels over the standard 50GHz ITU-T grid at a net
spectral efficiency (SE) of 8.4b/s/Hz. It is shown that the use of
near ideal Nyquist pulse shaping, spectrally-efficiecnt high-order
modulation format, distributed Raman amplficiation, distributed
compensation of ROADM filtering effects, coherent equalization
and high-coding gain forward error correction (FEC) code may
enable us to operate future 400G systems over the standard
50GHz-grid optical network.
Index Terms—Spectral efficiency, modulation format,
coherent, digital, QAM, Optical transmission, fiber,
capacity,ROADM, optical filtering, pulse shaping
I. INTRODUCTION
Lowering optical transport costs traditionally has been
achieved by increasing per-channel data rates and spectral
efficiency (SE). The commercial availability of 100Gb/s WDM
technology in 2010 now is being followed by research on
per-channel bit rates beyond 100Gb/s with spectral efficiencies
greater than 2bit/s/Hz to explore possibilities for the
next-generation transport standard, which appears likely to be at
400Gb/s [1]. From the historical point of view, increasing the
transport interface rate by a proportional increase of the spectral
efficiency minimizes the cost per transmitted bit. Following this
trend, a spectral efficiency of 8b/s/Hz might be needed for
future 400Gb/s systems (see Fig. 1). Such a SE essentially
enables operation of future 400Gb/s systems over existing
optical networks with 50GHz WDM channel spacing and
therefore is very attractive from the carrier’s perspective.
Due to several fundamental limitations, transmitting
400Gb/s per-channel signals on the 50GHz WDM grid is very
challenging. First, according to the information theory of
Shannon, 11.76dB higher signal-to-noise ratio (SNR) is
required for an 8b/s/Hz 400Gb/s system as compared to current
2b/s/Hz 100Gb/s systems, even without considering the fiber
nonlinearity. Second, the fiber nonlinearity limits the allowable
launch power and, therefore, fundamentally limits the
achievable signal SNR. Furthermore, a higher-SE modulation
format is less tolerant toward fiber nonlinearity due to the
reduced Euclidean distance. Third, non-ideal passband shapes
from the optical network components, such as the widely used
reconfigurable add/drop multiplexer (ROADM), will cause
significant channel narrowing. Finally, a high-SE modulation
format is less tolerant toward the laser phase noise, which may
introduce extra penalty.
Spectrally-efficient high-order coherent modulation
formats, coherent detection, transmitter-side and receiver-side
digital signal processing, distributed Raman amplification,
high-coding-gain FEC code, and possibly new low-loss and
low-nonlinearity fibers all have been considered as potential
enabling technologies for the next generation of high-speed
transport systems. These technologies are being experimentally
explored in the research community. For examples,
single-channel bit-rates beyond 100-Gb/s have been
demonstrated by using both single-carrier (up to 448-Gb/s) [2]
and multi-carrier based high-order coherent modulation formats
(up to 448-Gb/s [3,4] and 1.2-Tb/s [5]). For WDM
transmission, 50GHz-spaced, 10x224Gb/s over 12x100km at a
SE of 4b/s/Hz [6], 10x456Gb/s in 70GHz channel spacing over
8x100km at SE=6.1bit/s/Hz [7], and 100GHz-spaced,
3x485Gb/s over 48x100km at SE=4b/s/Hz [8] have been
demonstrated by using PDM-16QAM based modulation, digital
coherent detection, distributed Raman amplification as well as
new ultra-large area fiber (ULAF).
Projected
SE (b/s/Hz)

8
7
6
5
4
3
2
1
0
Deployed
In 2011-2012
0
100
200
300
400
Transport interface rate (Gb/s)
Fig. 1. Projected demand for spectral efficiency (SE) for the next
generation transport standard
Recently, 400G transmission over the standard 50GHz
WDM grid has been demonstrated by using PDM-32QAM
based modulation [9-11]. In [10], 8×450Gb/s WDM signals
over 400km of ULAF fiber and passing through one
50GHz-grid WSS-based reconfigurable add/drop multiplexer
(ROADM) has been demonstrated at a net spectral efficiency
(SE) of 8b/s/Hz. This is the first experimental demonstration for
400G WDM system over the standard 50GHz-grid otpical
network. In [11], the transmission reach has been extended to
800km by introducing a broadband optical spectral shaping
technique for ROADM filtering effects compensation. This is
1
2
the longest transmission diatance demonstrated for WDM SE
beyond 4b/s/Hz. The key enabliong technologies as well as the
experimental results will be reviewed in the following sections.
The remainder of this paper is organized as follows. Section
II devote to the description of the used 450Gb/s
PDM-Nyquist-32QAM transmitter. Section III presents the
coherent receiver and the digital signal processing (DSP)
algorithms. Two WDM transmission experiments (including the
measured back-to-back results) are presented in section IV.
Discussions and summaries are given in section V.
-20
D/A
converters
(a)
-40
-60
-50
-50
-25
-25
00
25
25
50
50
PC
ILF
9.2G
clock
ILF
>33dB
Original 50GHzspaced signal
MZM2
-20
-40
-50
-50
Q
(b)
Pre- equalized
9Gbaud digital
Nyquist 32QAM
D/A
converters
-60
-80
POL
MUX
IQ
MOD 2
PBS
PC
EDFA
OC
25/50G PC
ILF
12.5/25G
VOA
Power (dBm)
MZM1
IQ
MOD 1
I
EDFA
Laser
Q
I
Freq. offset (GHz)
Pre- equalized
9Gbaud digital
Nyquist 32QAM
-25
-25
0
0
25
25
50
50
Freq. offset (GHz)
Fig. 2. 450Gb/s PDM-Nyquist-32QAM transmitter. VOA: variable attenuator,
PC: polarization controller, PBS:polarization beam splitter.
Time (Symbol period)
Time (Symbol period)
(a)
(b)
Fig. 3. (a) Interpolated impulse response of the Nyquist filter and (b) the
resulting eye diagram of the generated baseband 32QAM signal.
9GHz
9GHz
25
Amplitude (dB)
Amplitude (dB)
20
15
10
-1
25
20
15
10
5
5
-0.5
0
0.5
Frequency (unit: baud rate)
1
-1
-0.5
0
0.5
1
Frequency (unit: baud rate)
(a)
(b)
Fig. 4. Measured relative amplitude spectra of 9Gbaud Nyquist 32QAM
baseband electrical drive signals (after D/A converter) without pre-equalization
(a) and with pre-equalization (b).
Power (dBm)
To overcome the limitation of available digital-to-analog (D/A)
converter bandwidth, a frequency-locked five-subcarrier
generation method has been utilized to create the 450Gb/s
per-channel signal [10,11]. Fig. 2 shows the demonstrated
450Gb/s PDM-Nyquist-32QAM transmitter. The output from a
continuous-wave (CW) laser with linewidth  100kHz is split
by a 3dB optical coupler (OC). One output is sent to a
Mach-Zehnder modulator (MZM-1) driven with a 9.2GHz
clock to generate two 18.4GHz-spaced subcarriers per channel
(i.e. the two first-order signal components) offset by ±9.2GHz
from the original wavelengths. After an erbium-doped fiber
amplifier (EDFA) and a 12.5/25GHz interleaver filter (ILF), the
original wavelengths and second-order harmonics are
suppressed by more than 40dB relative to the first-order
components (see inset (a) in Fig. 2). The signal is then equally
split between two outputs of a polarization beam splitter (PBS)
prepared via a polarization controller (PC). The two subcarriers
on one PBS output are sent to an IQ modulator (IQ MOD1),
driven with a pre-equalized 9Gbaud Nyquist 32-QAM signal
with 215-1 pseudorandom pattern length. The Nyquist pulse
shaping has roll-off factor of 0.01, and the digital Nyquist filter
has a tap length of 64. Fig. 3 shows the Nyquist filter impulse
response used in this experiment and the resulting eye diagram
of the generated 32QAM baseband signal in one quadrature.
Frequency-domain based pre-equalization [12] is used to
compensate for the band-limiting effects of the D/A converters,
which have 3-dB bandwidths < 5GHz at 10bit resolution and a
24GSa/s sample rate. The measured relative amplitude spectra
of the generated 9Gbaud Nyquist 32QAM baseband electrical
drive signals (after D/A converters) with and without
Amplitude (V)
9.2G
clock
0
>40dB
Power (dBm)
II. 450GB/S PDM-NYQUIST-32QAM TRANSMITTER
pre-equalization are shown in Fig. 4. One can see that the
filtering effects caused by the D/A converters are effectively
compensated by using frequency-domain based digital
pre-equalization.
A second Mach-Zehnder modulator (MZM-2) driven with a
9.2GHz clock is placed at the second PBS output to generate
first-order signal components at 0GHz and 18.4GHz offsets
from the original wavelength. After MZM-2, the signals pass
through two 25/50GHz interleavers to suppress the 0GHz signal
components and the unwanted harmonics (see inset (b) in Fig.
2). The second ILF re-inserts the original CW signal (from the
second 3dB OC output), resulting in three 18.4GHz-spaced
subcarriers from the original wavelength. These three
sub-carriers pass through an IQ modulator (IQ MOD2), driven
by a second pre-equalized 9Gbaud Nyquist 32-QAM signal
having 215-1 pseudorandom pattern length and originating from
a second D/A converter. Then the sets of two and three 45Gb/s
sub-carriers are passively combined and polarization
multiplexed with 20ns relative delay, resulting in a 450Gb/s
signal that occupies a spectral width of 45.8GHz, sufficiently
confined to be placed on the 50GHz ITU grid (see Fig. 5).
45.8GHz
BW=0.01nm
Freq. offset (GHz)
Fig. 5. Measured optical spectrum of the generated 450Gb/s PDM-Nyquist
32QAM signal consisting of five subcarriers.
2
3
COHERENT RECEIVER HARDWARE AND ALGORITHM
A DSP-enabled coherent receiver is used for the detection and
demodulation of the received PDM-Nyquist-32QAM signal.
the polarization- and phase-diverse coherent receiver front-end
consists of a polarization-diverse 90-degree hybrid, a tunable
external cavity laser (ECL) of ~100kHz linewidth serving as the
local oscillator (LO), and four balanced photo-detectors. An
optical tunable filter (OTF) with ~50GHz -3dB bandwidth is
used to select the desired channel for detection. The subcarriers
are selected by tuning the LO to within 200MHz of their center
frequencies. A four-channel real-time sampling scope with
50GSa/s sample rate and 16GHz analog bandwidth performs the
sampling and digitization (ADC) function, followed by
post-transmission DSP of the captured data on a desktop
computer.
The flow chart for the offline receiver DSP is shown in Fig.
6. After digital compensation of optical front-end errors
(sampling skews and hybrid phase errors) and anti-aliasing
filtering, the 50 GSa/s signal is first down-sampled to a sample
rate equal to 2symbol rate. After that, the bulk chromatic
dispersion (CD) is compensated using a fixed T/2-spaced finite
impulse response (FIR) filter with 72 complex-valued taps. (In a
practical system, frequency-domain based equalization is more
efficient than time-domain based equalization.) Next, we
performed simultaneous polarization recovery and residual CD
compensation with four complex-valued, 21-tap, T/2-spaced
adaptive FIR filters, optimized by a two-stage equalization
strategy: The classic constant-modulus algorithm is used in the
first stage to achieve pre-convergence, and after the
preconvergence
is
reached,
a
decision-directed
least-mean-square algorithm is then used for the steady-state
optimization.
Error
calculation
IX
j
hX
X
Qx
Front-end
correction
2baud rate
Re-sample
j
X
Carrier
recovery
Decoding,
decision &
BER count
Y
Carrier
recovery
Decoding,
decision &
BER count
hxy
CD
Comp.
IY
hxx
hY
hyx
Y
hyy
Error
calculation
CMA: pre-convergence
DD-LMS: steady-state optimization
Carrier frequency and phase recovery are implemented after
the initial equalization stage. The frequency offset between the
LO and the signal is estimated by using a constellation-assisted
two-stage blind frequency search method [12]: the frequency
offset is first scanned at a step size of 10MHz and then at a step
size of 1MHz, and the optimal frequency offset is the one that
gives the minimum mean-square error. For each trial frequency,
the carrier phase is first recovered (with best effort) by using a
newly proposed hybrid blind phase search (BPS) and maximum
likelihood (ML) phase estimation method [13], and decisions
made following this phase estimation are then used as reference
signals for mean-square error calculation. Note that this
WDM EXPERIMENTS
Two 450Gb/s per-channel WDM transmission experiments
using PDM-Nyquist-32QAM has been demonstrated [10,11]. In
the first WDM transmission experiment, no optical pulse
shaping is used to compensate for the filtering effects from the
50GHz-grid ROADM. In the second WDM experiment, a
liquid-crystal-on-silicon (LCoS)-based flexible bandwidth
WSS has been used as a broadband optical pulse shaper to
mitigate the ROADM filtering effects.
A. 8450-Gb/s over 400km without optical shaping
The experimental setup for the WDM transmission of
8450Gb/s PDM-Nyquist-32QAM over 400km is shown in Fig.
7. The eight 450Gb/s C-band channels are based on odd (192.30
to 192.60THz) and even (192.35 to 192.65THz) sets of
multiplexed, 100GHz-spaced external cavity lasers (ECL),
which are combined using a 3dB optical coupler (OC) and
modulated in the 450Gb/s PDM-Nyquist-32QAM transmitter
(see Fig. 1). The measured optical spectrum of a single 450Gb/s,
32-QAM channel and the eight-channel WDM spectrum prior
to transmission are shown in Fig. 5 and Fig. 8, respectively.
ECLs
1
3
5
100km ULA fiber
Odd
7
2
4
6
8
4
‘ROADM’
OC
PDM-Nyquist
32QAM TX
Delay EDFA
Coherent
RX
50GHz
WSS
Even 50/100G
11dBm
WDM
Fig. 6 Post-transmission off-line digital signal processing flow chart
IV.
WDM
QY
frequency recovery method is applicable for any modulation
format and can reliably recover carrier frequency by using only
tens of symbols (64 are used in this experiment). The carrier
phase is estimated by using a two-stage method: the carrier
phase recovered from the previous symbol is used as an initial
test phase angle. The “decided” signal made following this
initial test phase angle is then used as a reference signal for a
more accurate ML-based phase recovery through a
feed-forward configuration [13]. For the first block of data, the
initial phase angle is obtained by using the BPS method [14]. To
reduce the probability of cycle slipping (no differential
coding/encoding is used in this experiment), a sliding-window
based symbol-by-symbol phase estimation is employed. For
bit-error-ratio (BER) calculation, errors are counted over more
than 1.2×106 bits of information.
Raman
pumps
Fig. 7. Experimental setup for 8450-Gb/s over 400km transmission
-15
Power (dBm)
III.
-25
-35
BW=0. 1nm
-45
-55
1555.5
1556.5
1557.5
1558.5
1559.5
Wavelength (nm)
Fig. 8. Measured optical spectrum of the generated 8450-Gb/s WDM signals.
3
4
Bit error ratio
10-1
standard single-mode fiber jumpers and including a
1450/1550nm WDM coupler for the counter-propagating
Raman pumps at the span outputs, the total span losses are 19.2,
19.6, 19.2, and 18.9dB. Hybrid Raman-EDFAs are used, with
an on-off Raman gain of 11dB per span from ~1450nm pumps.
Due to the gain-flattened EDFAs and the narrow ~3nm total
bandwidth of the eight 450Gb/s channels, the spectrum had
flatness to within 1dB after 400km, for a large range of span
input powers.
Theory
BB, single subcarrier
BB, single channel
BB, WDM
10-2
10-3
10-4
26
28
30
32
34
36
38
40
Fig. 9 shows the measured back-to-back BER performance
for three different conditions: a single subcarrier operating at 9
Gbaud, a single 450Gb/s channel composed of 5 subcarriers,
and one of the center 450Gb/s channels of 8450Gb/s WDM
channels spaced by 50GHz. The displayed OSNR for the single
subcarrier in this figure is a scaled result obtained by
multiplying the actual OSNR of the single subcarrier signal by a
factor of 5. No ROADM filtering was used in these
back-to-back measurements. The recovered Nyquist-32QAM
constellation diagram at an OSNR of 38.9dB for a single
450Gb/s channel also is shown in this figure. For comparison, a
theoretical curve is included in Fig. 9. There is ~6-dB
implementation penalty at BER 2×10-3. Thanks to the use of
digital Nyquist pulse shaping, the OSNR penalty (at 2×10-3
BER) from inter-channel WDM crosstalk is very small even
without narrow optical filtering, because the 450Gb/s signal is
well confined to within 45.8GHz bandwidth. The OSNR
penalty from inter-subcarrier crosstalk is less than 1 dB. A
portion of the inter-subcarrier crosstalk originates from the
out-of-band aliased spectral components from the electrical
drive signals.
10-1
Sub1
Sub3
Sub5
X
10-3
X
7
8
12
9
10
11
Total launch power (dBm)
Transmission (dB)
-40
-20
0
20
Frequency Offset (GHz)
Y
Y
13
Fig. 11. BER performance of the five subcarriers of the center DWDM channel
at 192.50THz after 4100km transmission for a range of total launch power
into the fiber spans.
10-2
FEC threshold
-5
10-3
-15
-25
After 400km
-35
BW=0.1nm
-45
1555.5 1556.5 1557.5 1558.5 1559.5
10-4
1555.5
0
-10
-20
-30
-40
Sub2
Sub4
Average
10-2
Power (dBm)
24
OSNR at 0.1 nm noise bandwidth
Fig. 9. Measured back-to-back performance under three different conditions, as
explained in the text.
Bit error ratio
22
Bit error ratio
20
1556.5
1557.5
1558.5
1559.5
Wavelength (nm)
Fig. 12. The average BER of the five subcarriers of all eight channels after
400km transmission with optimum launch power. The inset shows the optical
spectrum after 400km.
40
Fig. 10. Measured power transmission of the 50GHz-grid WSS used to emulate
the ROADM.
For the WDM transmission the eight 450Gb/s signals pass
through a 18, 50GHz-spaced WSS based on liquid-crystal
technology to emulate the filtering from a ROADM. Odd and
even channels are sent to separate WSS output ports, for
maximum filtering, and a relative delay of 175 symbols
decorrelates the odd and even channels before they are
recombined using a 3-dB optical coupler. Filtering from the
WSS passband is significant, as the -3dB and -6dB bandwidths
are 42.2 and 46.6GHz, respectively, as shown in Fig. 10.
Following the ROADM, the transmission line consists of four
100-km spans of ULAF with, at 1550nm, average Aeff of 135
um2, average attenuation of 0.179 dB/km, and average
dispersion of 20.2 ps/nm/km. After splicing the span inputs to
The results for the 8450Gb/s transmission experiment are
shown in Fig. 11 and Fig. 12. The BERs of the five sub-carriers
of the center channel at 192.50THz were measured as the total
launch power to the spans was varied. As shown in Fig. 11,
based on the average BER of the five sub-carriers, the optimum
total launch power per span is 11dBm, corresponding to an
average of 2dBm per 450Gb/s channel and -5dBm per
sub-carrier. At this launch power, the OSNR was 34dB/0.1nm
after the 400km transmission. It has been assumed that digital
signals from all five sub-carriers would co-exist on one silicon
chip, and as such, the FEC would be encoded on a per-channel
basis, not on a per-subcarrier basis. Thus, the net BER of the
450Gb/s channel is the average BER of the subcarriers [15]. Fig.
12 shows the performance of each of the eight 450Gb/s DWDM
channels at the optimum 11dBm total launch power. The
average BER of the five subcarriers of all eight channels is
better than 3.810-3, which is lower than the BER threshold for a
7% continuously interleaved BCH code (4.510-3 [16]). The
4
5
inset in Fig. 12 shows the optical spectrum after 400km, where
spectral filtering by the WSS is clearly evident.
ECLs
1
ROADM
WDM
Optical
spectral
shaping
PDM-Nyquist
32QAM TX
50GHz
WSS
-40
Launch to the fiber
100 km ULA fiber
switch
switch
Raman
pumps
3dB
Coupler
9dBm
1556.8
1557.3
1557.8
1558.3
1558.8
(c)
Fig. 14. Measurements of the optical spectra for a single channel with (thick)
and without (thin) optical spectral shaping (a) before the ROADM and (b) after
the ROADM, and (c) spectra of the five DWDM signals with optical spectral
shaping before and after the ROADM (launch to the fiber).
The measured optical spectra of a single 450Gb/s, 32-QAM
channel before and after the 50GHz ROADM are shown in Fig.
14a and b, respectively. The thin lines show the spectra without
optical spectral shaping. It is clear that when the signal (without
optical spectral shaping) passes through the 50GHz ROADM,
significant filtering occurs. The power loss due to this filtering
effect can be largely pre-compensated by using broadband
optical spectral shaping, as shown by the thick lines in Fig.
14a-b. Using this spectral shaping, the measured optical spectra
of the five 450Gb/s DWDM signals prior to and after the
ROADM are shown in Fig. 14c. The filtering effects of all the
five channels have been largely pre-compensated.
10-1
EDFA
1556.3
Wavelength (nm)
Even 50/100G
WDM
4
Odd
BW=0.1nm
-55
1555.8
ILF
OC
2
Delay
EDFA
5
-25
Launch to the ROADM
Bit error ratio
3
Power (dBm)
B. 5450-Gb/s over 800km with optical shaping
The experimental setup for the WDM transmission of
5450Gb/s PDM-Nyquist--32QAM over 800km is shown in
Fig. 13. For this experiment, an LCoS-based dynamic,
broadband optical spectral shaper (with 1GHz resolution)
followed by a booster EDFA were inserted before the
50GHz-grid ROADM. This optical spectral shaper is used to
pre-compensate the ROADM filtering. Because such
pre-compensation will result in enhanced inter-channel WDM
crosstalk (due to the limited channel isolation of the 50GHz
WSS), a 50/100G interleaver is used inside the ROADM
emulator to combine the odd and even channels in order to
further suppress the inter-channel WDM crosstalk. The
re-circulating loop contains the same four 100-km spans of
ULA fiber with 11dB on-off Raman gain, as previously
described for the 400km experiment. After two circulations
(800km), the spectrum of the five 450Gb/s channels was flat to
within 1 dB, for total span input powers ranging from 6 to
12dBm. The optimal total launch power at the span inputs was 9
dBm.
-10
Theory
Back-to-back
After ROADM, with optical shaping
After ROADM, no optical shaping
10-2
10-3
4
Coherent
RX
10-4
20
20
Fig. 13. Set-up for 8450-Gb/s transmission over 800km. OTF: optical tunable
filter. ILF: interleaver filter. OC: optical coupler; PC: polarization controller
-15
-15
Power (dBm)
Power (dBm)
-25
45.8GHz
-35
-45
BW=0.01nm
-55
-35
BW=0.01 nm
-45
No optical shaping
With optical shaping
-65
1557
1557.2
1557.4
Wavelength (nm)
(a)
1557.6
45.8GHz
-25
No optical shaping
With optical shaping
-55
1557
1557.2
1557.4
Wavelength (nm)
(b)
1557.6
24
32
2828
24
3636
32
OSNR at 0.1 nm noise bandwidth
4040
Fig. 15. OSNR sensitivity for the generated 450Gb/s signal
Figure 15 shows the measured OSNR sensitivity for the
450Gb/s PDM-Nyquist-32QAM signal for three different
scenarios: a single channel without optical filtering from the
ROADM (i.e. the back-to-back case, no optical shaping
applied), with filtering from the ROADM while the optical
spectral shaping is applied, and finally with filtering from the
ROADM but without optical spectral shaping. For comparison,
a theoretical OSNR sensitivity curve is also displayed in this
figure. Compared to the previous experiment, the back-to-back
sensitivity is improved by about 2 dB at 2×10 -3 BER, achieved
mainly by improved bias optimization for the two IQ
modulators. In the presence of ROADM filtering, optical
spectral shaping improves the OSNR sensitivity by more than 2
dB at 2×10-3 BER.
5
6
Power (dBm)
-18
-28
After 800km
BW=0.1nm
-38
-48
1555.8
1556.8
1557.8
1558.8
Wavelength (nm)
Fig. 16. Measured optical spectrum after 800 km transmission
FEC Threshold
Bit error ratio
510-3
410-3
310-3
610-3
210-3
310-3
6
6
110-3
1556
8
11
9
7
10
Total Launch power (dBm)
8
10
1558
1557
Wavelength (nm)
12
12
1559
Fig. 17. Measured BER of the five 450Gb/s DWDM channels after 800km
transmission, where the inset displays the measured BER for the center DWDM
channel versus total launch power for all five 450Gb/s channels.
The transmission results are shown in Figs. 16 and 17, where
the measured optical spectrum after 800km transmission is
shown in Fig. 16. The measured OSNR is 31dB/0.1nm. The
measured BER of all five WDM channels at the optimal total
launch power of 9 dBm (2dBm/ch) is presented in Fig. 17,
where the inset shows the measured BER for the center channel
located at 192.5THz as the total launch power into the spans is
varied. The BER of all five channels is better than 3.810-3, and
the worst subcarrier BER is 4.310-3, both of which are lower
than the BER threshold for a 7% continuously interleaved BCH
code (4.510-3 [16]).
V.
CONCLUSIONS AND DISCUSSIONS
In conclusion, this paper presents a detailed review on two
recently demonstrated high-SE (8.4b/s/Hz) 400G transmission
experiments. These two experiments demsontrate, for the first
time, that 400Gb/s per channel WDM signals can be transmitted
over the conventional 50GHz ITU-T grid with a transmission
reach up to 800km (eight 100km spans) and more over, passing
through one 50GHz grid ROADM. These results are
accomplished by the use of a spectrally-efficient high-order
modulation format, Nyquist-shaped PDM-32QAM, as well as
both pre- and post-transmission digital equalization.
Low-nonlinearity fiber and low-noise Raman amplification also
have been used to address the reduced nonlinear and noise
tolerance of the high-order modulation format.
To overcome the bandwidth limitation of the available D/A
converters, a frequency-locked five subcarrier generation
method with high sideband suppression has used to create the
450Gb/s PDM-Nyquist-32QAM signals. The use of five
frequency-locked subcarriers and near ideal Nyquist pulse
shaping essentially allows the 450Gb/s spectrum being well
confined within the 50GHz channel spacing (occuping a signal
spectral width about 45.8GHz). Using multiple subcarriers
(with Nyquist pulse shaping) within each channel could enable
all-optical sub-wavelength grooming, which may be useful for
future 400Gb/s and beyond ultra-high-speed systems.
Another key enabling tehcology is LCoS based broadband
optical spectral shaping. It is shown that this spectral shaping
technique can be used to mitigate the narrow ROADM filtering
effects. Because an optical spectral shaper could be designed
within each ROADM and essentially enbles distributed
compensation of ROADM filtering effects. Distributed
ROADM filtering compensation method has some advantages
compared to transmitter-side pre-equalization or receiver-side
post-equalization
because,
unlike
transmitter-side
pre-equalization, distributed compensation does not require
more launch power into the fiber, and thus it does not enhance
the fiber nonlinear effects. Distributed compensation also does
not enhance the noise components, as occurs with receiver-side
post-equalization.
The
disadvantage
of
distributed
compensation is the need for extra optical amplification to
compensate for the loss caused by the optical spectral shaper.
REFERENCES
[1] M. Camera, B. E. Olsson, and G. Bruno, “Beyond 100Gbit/s: System
implications towards 400G and 1T,” ECOC 2010 Symposium toward 1Tb/s,
September 2010, Torino, Italy, paper Th10G1.
[2] P. J. Winzer,A. H. Gnauck, S. Chandrasekhar, S. Draving, J. Evangelista,
and B. Zhu, “Generation and 1,200-km transmission of 448-Gb/s ETDM
56-Gbaud PDM 16-QAM using a single I/Q modulator,” ECOC 2010,
Torino, Italy, September 2010, PDP 2.2.
[3] J Yu, X Zhou, MF Huang, D Qian, PN Ji, “400Gb/s (4× 100Gb/s)
orthogonal PDM-RZ-QPSK DWDM Signal Transmission over 1040km
SMF-28,” Optics Express, Vol. 17, Issue 20, pp. 17928-17933 (2009)
[4] X. Liu, S. Chandrasekhar, B. Zhu, A.H. Gnauck, and D. W. Peckham,
“Transmission of a 448-Gb/s reduced-guard-interval CO-OFDM signal with
a 60-GHz optical bandwidth over 2000 km of ULAF and five 80-GHz-Grid
ROADMs,” in Proc. OFC-NFOEC 2010, San Diego, CA, March 2010, paper
PDPC2.
[5] S. Chandrasekhar, X. Liu, B. Zhu, and D. W. Peckham, “Transmission of a
1.2-Tb/s 24-carrier no-guard-interval coherent OFDM superchannel over
7200-km of ultra-large-area fiber,” ECOC 2009, Viena, Austria, September
2009, PDP 2.6.
[6] A. H. Gnauck, P. J. Winzer, S. Chandrasekhar, X. Liu, B. Zhu, and D. W.
Peckham, “10 × 224-Gb/s WDM transmission of 28-Gbaud PDM 16-QAM
on a 50-GHz grid over 1,200 km of fiber,” ,” in Proc. OFC-NFOEC 2010,
San Diego, CA , March 2010, paper PDPB8.
[7] Y.-K. Huang, E. Ip, M.-F. Huang, B. Zhu, P. N. Ji, Y. Shao, D. W.
Peckham, R. Lingle., Jr, Y. Aono, T. Tajima, and T. Wang, “10×456-Gb/s
DP-16QAM Transmission over 8×100 km of ULAF using Coherent
Detection with a 30-GHz Analog-to-Digital Converter,” OECC 2010, City?,
Japan, July 2010, paper PDP3.
[8] X. Liu, S. Chandrasekhar1, P. J. Winzer1, B. Zhu, D. W. Peckham, S.
Draving, J. Evangelista, N. Hoffman, C. J. Youn4, Y. H. Kwon, and E. S.
Nam, “3× 485-Gb/s WDM Transmission over 4800 km of ULAF and 12×
100-GHz WSSs Using CO-OFDM and Single Coherent Detection with
80-GS/s ADCs” OFC 2011, paper JThA37.
[9] H. Takahashi, K. Takeshima, I. Morita, and H. Tanaka, “400-Gbit/s Optical
OFDM Transmission over 80 km in 50-GHz Frequency Grid,” Tu.3.C.1,
ECOC 2010,
[10] X. Zhou, L. E. Nelson, P. Magill, R. Issac, B. Zhu, D. W. Peckham, P.
Borel, and K. Carlson, “ 8 450-Gb/s, 50-GHz-spaced, PDM-32QAM
6
7
transmission over 400km and one 50GHz-grid ROADM,” in Proc. OFCNFOEC 2011, Los Angeles, CA, March 2011, paper PDPB3.
[11] X. Zhou, L. E. Nelson, P. Magill, R. Issac, B. Zhu, D. W> Peckham, P.
Borel, and K. Carlson, “ 800km transmission of 5
450-Gb/s
PDM-32QAM on the 50GHz grid using electrical and optical spectral
shaping ,” accepted for presentation at ECOC 2011, Geneva, Switzerland,
September 2011, paper We.8.B.2.
[12] X. Zhou, J. Yu, M. F. Huang, Y. Shao, T. Wang, L. Nelson, P. D. Magill1,
M. Birk, P. I. Borel, D. W. Peckham and R. Lingle, "64-Tb/s, 8 b/s/Hz,
PDM-36QAM transmission over 320 km using both pre- and
post-transmission digital signal processing," J. Lightwave Technol., Vol. 29,
No. 4, pp. 571-577, February 15, 2011,
[13] X. Zhou, "An improved feed-forward carrier recovery algorithm for
coherent receiver with M-QAM modulation format," IEEE Photonics
Technol. Lett., Vol. 22, No. ??, pp. 1051-1053, 2010.
[14] T. Pfau, S. Hoffmann, and R. Noé, “Hardware-efficient coherent digital
receiver concept with feed-forward carrier recovery for M-QAM
constellations,” J. Lightwave. Technol., vol. 27, no. 8, pp. 989-999, April 15,
2009.
[15] M. Tomizawa and Y. Yamabayashi, “Parallel FEC code in high-speed
optical transmission systems,” Electronics Lett., Vol. 35, No. 16, pp.
1367-1368, 1999.
[16] F. Chang, K. Onohara and T. Mizuochi, “Forward error correction for
100G Transport Netwroks, ” IEEE Communications Magazine, Vol. 48, No.
3, pp. S48-S55, 2010
Xiang Zhou received his Ph. D degree in electrical
engineering from Beijing University of Posts &
Telecommunications in 1999. From 1999 to 2001, he
was with Nangang Technological University, Singapore,
as a Research Fellow, doing research on optical CDMA
and wide-band Raman amplification. He has been a
Senior Member Technical Staff in AT&T Labs-Research
since October 2001, working on various aspects of
long-haul optical transmission and photonic networking
technologies,
including
Raman
amplification,
polarization-related impairments, optical power transient control, advanced
modulation formats and digital signal processing at bit rate 100Gb/s and
beyond. He has authorized/co-authorized more than 100 peer-reviewed
journals and conferences publication, and holds 28 USA patents. He currently
serves as an associate editor of Optics Express. He is member of OSA, and a
senior member of IEEE.
7