100g transport systems

8/20/2019 100G Transport Systems

1/12


2/12

20 China Communications • April 2013

used direct-detection on/off keying (OOK) to

transmit the first single-carrier, electronically

multiplexed 100G signals [3]. In order to

compress the fairly wide-band spectrum of

these optical OOK signals, both in-phase and

quadrature components of the optical fieldwere subsequently exploited by implementing

100G Differential Quadrature Phase Shift

Keying (DQPSK) with delay demodulation.

In 2007, this technology culminated in the

industry’s first field trial of live video trans-

mission using single-carrier 100G on a

504-km network route between Tampa and

Miami, Florida, United States as part of Ver-

izon’s® in-service production network [4].

Realizing that one further multiplexing step

would be necessary to enable operation on a

50-GHz Wavelength-Division Multiplexing

(WDM) grid, the above efforts were followed

by introducing Polarization Division Multi-

plexing (PDM) at around 25 Gbaud, which in

a commercial context can be most efficiently

done through coherent detection [5-7] and at

100G was first demonstrated in research in

2008 [8].

Soon after, in November 2009, Alcatel-

Lucent performed the first 100G field trial based on state-of-the-art Polarization-Division

Multiplexed Quadrature Phase Shift Keying

(PDM-QPSK) with coherent detection, trans-

mitting with commercial margins a 112-Gb/s

channel together with 40-Gb/s and 10-Gb/s

channels on a 1088-km link between Madrid

and Merida via Sevilla, in Telefónica’s net-

work. Less than a year later, in June 2010,

Fig.1 The key benefits of the PSE chipset

Alcatel-Lucent launched its first-to-market

single-carrier 100G PDM-QPSK coherent

solution, leveraging highly integrated sili-

con-based ultra-fast monolithic signal digiti-

zation and Digital Signal Processing (DSP),

unique algorithms, and high coding-gain For-ward Error Correction (FEC) techniques.

Availability of 40-Gb/s Polarization-Division

Multiplexed Binary Phase Shift Keyed

(PDM-BPSK) single-carrier interfaces fol-

lowed shortly, completing the portfolio of co-

herent solutions. All above solutions were

in-house developments, from the choice of

best-in-class DSP algorithms provided by Bell

Labs to the design of ultra-high speed digital

circuits that make coherent detection a com-

mercial reality. The resulting 100G trans-

ponder comprised a board that featured a high

level of integration, both electronic and

photonic, and a best in class density. Owing to

its early adoption of single-carrier 100G tech-

nologies also allowed gathering significant

experience in the fabrication, testing and de-

ployment of such transponders across various

deployed optical networks.

In Dec 2011, an enhanced 100G solution

was introduced to improve the 100G trans-

ponder’s reach from 1 500 km to 2 000 km. In

March 2012, the reach was further extended to

3 000 km using the Photonic Service Engine

(PSE), which scales to enable future 400G

per-channel interface rates. Key features of

this PSE monolithic chip include (see also in

Figure 1):

1) High resolution 4 channels ADC-DSP

and support 400G dual-carrier solution

2) DSP algorithms for robust plug and play

fiber operation3) Advanced optical monitoring and FEC

4) Ultra-fast signal adaptation/synchroniza-

tion

5) Advanced design for ultra-long-haul sys-

tems without optical compensation

6) 70M+ gates

Many other 100G coherent DSP solutions

have been announced by various vendors.

Concrete deployments from system integrators

are documented for dual-carrier 100G solu-

100G coherent techn-

ology is introduced by

the explosively band-

width demand, and

PDM-QPSK is the only

mature approach in

the industry. With

WDM technology, the

capacity of each fiber

can be up to 10Tb/s.

With Bell-labs’ contin-

uous innovation, Alca-

tel Lucent is the first

vendor to introduce

single-carrier 100G

solution and leading

the commercial evolu-

tion to 400G. Thispaper introduces the

key benefits of Alcatel-

Lucent 100G and

bench mark testing in

China market segment,

and also discusses the

enabling technologies

for Terabits/s era.


3/12


4/12


5/12

China Communications • April 2013 23

Fig.4 OSNR measurement in ROADM-based WDM systems

the crosstalk.

Polarization-Dependent Loss (PDL) can

lead to a significant OSNR measurement error

since the noise with the same polarization as

the signal can have a different power compared to the noise in the orthogonal polariza-

tion.

For a polarization multiplexed signal, there

is a separate signal on each of the two or-

thogonal polarizations so it is not possible to

extinguish the signal using a polarization

beam splitter. Hence, it is not possible to use

this method of OSNR measurement for these

signals.

For 100G PDM-QPSK signals, which in-

herently contain signal in both polarizations,the polarization extinction method fails to

measure the OSNR accurately. The industry

sometimes uses temporary laser shutdown to

measure OSNR, but this is an off-line method,

and has poor repeatability and low accuracy,

which cannot be used for in-service mainte-

nance.

So far Alcatel-Lucent has developed the

first commercial solution of in-band OSNR

measurement applicable to all data-rates of

relevance (10G, 40G, 100G, 400G, etc.) based

on its Wavelength Tracker technology and

advanced DSP algorithms. Figure 5 shows the

working diagram of the novel OSNR meas-

urement technique: A narrow-band tunableoptical filter dynamically selects a WDM

channel to be measured by a specific algo-

rithm. The output of the tunable optical filter

contains the signal and ASE noise within the

band. As the optical signal carries wavelength

tracker and ASE information, we can accu-

rately evaluate the in-band ASE noise and

signal by using DSP. Through lab trials and

field trials with two China operators, China

Mobile Communications Corporation (CMCC)

and China Unicom (CU), OSNR measure-ments have been shown to work for 10G, 40G,

and 100G signals with high precision

(±1.5 dB). Table I summarizes the results of

an OSNR online measurement campaign.

To the best of our knowledge, the above

described technique is the only commercially

implemented method to measure OSNR online

for dual-polarization signals. It has the capa-

bility to detect at all points within the network

in the pure optical domain without any (Optical-


6/12


Fig.5 OSNR online measurement working diagram

Table I Online OSNR measurement results

ALU 100G PDM-QPSK OSNR measurement demo results

OSA instrument MS9710C_off-line ALU WT solution_on-line OSNR Delta

Test A

Signal Power:-20.7 dBm

ASE noise power:-36.3 dBmOSNR=15.6 dB

OSNR=16.1 dB 0.5 dB

Test B


ASE noise power:-36.3 dBmOSNR=18.1 dB

OSNR=18.8 dB 0.7 dB

Test C


ASE noise power:-36.3 dBm

OSNR=19.4 dB

OSNR=20.3 dB 0.9 dB

Table II Alcatel-Lucent’s 100G test results inCMCC lab trial and field trial

Key factor Test Result Standard

B2B OSNR(dB) 13.3(HD) 15.0(HD)

SPAN 16(HD) 12(HD)

OSNR margin(dB) 5.7~6.6 5

Q margin(dB) 4.04~5.09 3

Protection


7/12


Fig.6 Alcatel-Lucent’s 100G test configuration in CMCC field trial

covers equipment to network, from manage-

ment to multi-vendor interconnection.

The test configuration was large-scale, us-

ing 40 single-carrier wavelengths at 100G, and

transmitting them over up to 1 600 km of fiber

for lab trials and over more than 1 000 km in

field trials. Figure 6 shows the field trial con-

figuration.

During these 100G BMT activities in 2012,

Alcatel-Lucent’s 100G WDM/OTN system

showed superior transmission performance,

and is capable of supporting ODU0/1/2/2e/3/4

switching with less than 50ms protection

switching. The 100G solution is mature, which

is also proven by the massive global deploy-

ment; and it is time to launch large-volume

commercial deployments in China.

V. R ESEARCH BEYOND 100G

WDM transmission systems with per-channel

data rates of up to 1 Tb/s are being actively

researched worldwide to meet the ever in-

creasing capacity demands [10-11]. In order to

achieve net information rates in the 400-Gb/s

to 1-Tb/s range, novel signal synthesis and

detection approaches are being pursued. Su-

perchannel transmission [12-13] has recently

attracted much attention in this context. Using

this approach, multiple optical carriers are

modulated individually at moderate symbol

rates, and are then combined to form a super-

channel that delivers the desired high data

rates at high spectral efficiency. This scheme

exploits the benefits of mature electronic and

optoelectronic components at moderate speeds

and uses optical parallelization in the fre-

quency domain to achieve high aggregate data

rates beyond the limits of the electronics and

optoelectronics. Large-scale photonic integra-

tion is essential to reduce the cost per bit for

superchannel implementations that require a

large number of transmitter and receiver fron-

tends. Aggregate information rates per channel

at or beyond 1 Tb/s have been demonstrated

using different implementation choices [13-


8/12


16]. With the advances in high-speed elec-

tronics and optoelectronics, Bell Labs has re-

cently increased the coherent modulation and

detection speed that can be achieved by a sin-

gle pair of transmitter and receiver frontends

to 80 Gbaud for 16-QAM [16] and 107 Gbaudfor QPSK [17]. Based on this 16-QAM inter-

face, a dual-carrier 1-Tb/s superchannel was

formed with the use of only two pairs of

transmitter/receiver frontends [16].

Figure 7 summarizes measured optical

spectra and signal constellations of recent

Tb/s-class transmission demonstrations by

Bell Labs [13-16]. Chandrasekhar et al. dem-

onstrated the transmission of a 1.2-Tb/s super-

Fig.7 Recent demonstrations of Tb/s-class transmiss-

ion by Bell Labs researchers using (a) 24-carrier

PDM-QPSK [13]; (b) 8-carrier PDM-OFDM-

16QAM [14]; (c) 4-carrier PDM-16QAM [15]; and

(d) 2-carrier PDM -16QAM [16]

channel, consisting of 24 PDM-QPSK signals,

seamlessly multiplexed at the Orthogonal

Frequency-Division Multiplexing (OFDM)

condition, over 7 200 km of ultra-large area

fiber (ULAF) [13]. Liu et al. demonstrated

the transmission of a 1.5-Tb/s superchannel,consisting of 8 PDM-OFDM-16QAM signals,

over 5 600 km of ULAF, achieving a record

Spectral-Efficiency-Distance-Product (SEDP)

of over 32 000 km⋅ b/s/Hz for Tb/s-class su-

perchannel transmission with >5 b/s/Hz net

spectral efficiency [14]. Renaudier et al. dem-

onstrated high-SE long-haul transmission of

22-Tb/s using 1-Tb/s superchannels over

2 400 km [15]. Each of the 1-Tb/s superchan-

nels consisted of four 40-Gbaud PDM-

16QAM signals, multiplexed under the quasi-

Nyquist-WDM condition [18-19]. Finally,

Raybon et al. showed dual-carrier Terabit

transmission over 3 200 km of fiber [16]. This

Tb/s transceiver implementation requires the

minimum optical hardware, as compared to

other implementations demonstrated so far.

Figure 8 shows the schematic of the trans-

mitter setup used for the 1.5-Tb/s superchan-

nel generation reported in Ref. [14]. Eight

32.8-GHz spaced external cavity lasers wereseparated into odd and even signal groups,

which were modulated by two I/Q modulators.

The two drive signals for the first I/Q modu-

lator were provided by two 50-GSamples/s

DACs. The two drive signals for the second

modulator were the complementary outputs

from the same two DACs. The inputs to the

DACs were provided by a Field-Programma-

ble Gate Array (FPGA) based real-time logic

circuit with stored OFDM-16QAM wave-

forms.

The drive waveforms were pre-equalized to

compensate for frequency roll-off of the DAC.

The guard band between two adjacent

30-Gbaud signals was 2.8 GHz, which limits

the coherent crosstalk between the signals and

readily allows for efficient sub-band signal

processing. A Low-Density Parity Check

(LDPC) code with a rate of 0.864 was imple-

mented for the inner SD-FEC, and a typical

7%-overhead HD-FEC code was assumed as


9/12


Fig.8 Schematic of the transmistter setup for generating a 1.5-Tb/s superchannel (upper ) and the measured

spectra of the generated superchannel (lower )

the outer code [14]. The overall FEC overhead

was 23.46%. The net superchannel data rate

was 1.51 Tb/s and the channel spectral effi-

ciency was 5.75 b/s/Hz. This channel used anoptical spectral bandwidth of 262.5 GHz,

which is compatible with the new ITU stan-

dard on flexible grid WDM with 12.5-GHz

granularity (ITU-T G.694.1).

The superchannel was launched into a

re-circulating loop consisting of four 100-km

ULAF spans with 19.7-ps/nm/km dispersion.

Each span was amplified by a hybrid Ra-

man/Erbium-Doped Fiber Amplifier (EDFA).

The average span loss was ~19 dB, and the

Raman gain per span was ~8 dB. After trans-mission, each of the signals was sent to a

digital coherent receiver with offline DSP. The

sampling rate of the ADCs was 50 GS/s, the

same as that of the DAC. The offline DSP

includes a dispersion-optimized 1-step-per-

span Non-Linearity Compensation (NLC) in

which a pre-dispersion compensation of 1 780

ps/nm was applied so that the NLC was ap-

plied to the beginning portion of each of span

where the signal power was high. Deci-

sion-directed phase compensation, initiated by

training symbols, was used to realize pilot-free

phase estimation. The SD-FEC decoding was

performed using the same methodology reported in [20]. Figure 9 shows the measured

BER of all eight 30-Gbaud PDM-OFDM-

16QAM signals that made up the 1.5-Tb/s

superchannel after 5 600-km transmission at

Fig.9 Measured BER of all the signals inside a 1.5-Tb/s superchannel after

5 600-km transmission at the optimal signal launch power [14]


10/12


the optimal signal launch power. The output

BER values from the SD-FEC are all below

the outer HD-FEC threshold. The superchan-

nel spectrum after 5 600 km transmission is

also shown. This demonstration indicates the

promising potential of Tb/s-class superchanneltransmission in future long-haul optical trans-

port systems. As IEEE is defining 400GbE as the next bit

rate for high speed Ethernet interface, there is

a call for interesting session for 400GE in the

coming IEEE meeting in March. Before this

meeting, most of the members vote for 400GE

to be the candidate. Together with the 400GbE

interface definition, and the 400GE transmis-

sion solution will be defined by ITU-T corre-

spondingly. Actually all vendors are working

actively for beyond 100G solution. For exam-

ple, France Telecom-Orange and Alcatel-

Lucent announced the deployment of the

world’s first optical link offering a capacity of

400 Gigabits per second (Gb/s) per wave-

length in a live network recently. In China,

from 2010, a few beyond 100G research pro-

jects are supported by national projects such

as 973, 863 and NSFC (National Natural Sci-

ence Foundation of China). CCSA (ChinaCommunications Standard Association) also

starts the research project about beyond 100G

technology. Local vendors also announced

several lab experiment result for beyond 100G

publicly.

VI. SUMMARY

As the demand for network traffic grows con-

tinuously and exponentially, scaling line

speeds from 10 Gb/s and 40 Gb/s to 100 Gb/sand beyond is becoming necessary, while

transmission performance should not be sub-

stantially degraded and the overall economics

must prove in.

We described Alcatel-Lucent’s single-car-

rier 100G transmission system utilizing an

advanced electro-optics engine that leverages

the concepts of optimized signal modulation

coupled with coherent detection as one of the

leading 100G solutions in the industry. The

underlying ASICs comprise ultra-fast Ana-

log-to-Digital Converters (ADCs) and Digi-

tal-to-Analog Converters (DACs) and feature

advanced DSP, monolithically integrated in

CMOS in a very compact, space and power

efficient design. The outstanding transmission performance and the total cost of ownership

benefits provided by this innovative elec-

tro-optics engine made 100-Gb/s single-carrier

transmission commercially viable, as evi-

denced by its abundant deployment in many

different networks today, spanning both re-

gional and long haul reaches over several dif-

ferent fiber types, and operating with 10G,

40G and 100G mixed signals. Regarding the

evolution to 400G, Alcatel-Lucent has recently

field-deployed the first such high-speed inter-

face, and Bell Labs is intensely working on

ways to reach Terabit/s interface rates in a

commercially attractive way.

References

[1] TKACH R W. Scaling Optical Communications

for the Next Decade and Beyond[J]. Bell Labs

Technical Journal, 2010, 14(4): 3–9.

[2] KOROTKY S K. Traffic Trends: Drivers and

Measures of Cost-Effective and Energy-Eff-

icient Technologies and Architectures forBackbone Optical Networks[C]// Proceedings

of Optical Fiber Communication Conference

2012: March 4-8, 2012. Los Angeles, CA, USA,

2012: OM2G.1.

[3] WINZER P J, RAYBON G, DUELK M. 107-Gb/s

Optical ETDM Transmitter for 100G Ethernet

Transport[C]// Proceedings of 31st European

Conference on Optical Communication: Sep-

tember 25-29, 2005. Glasgow, Scotland, 2005,

6: 1-2.

[4] WINZER P J, RAYBON G, SONG Haoyu, et al .

100-Gb/s DQPSK Transmission: from Labora-

tory Experiments to Field Trials[J]. Journal of

Lightwave Technology, 2008, 26 (20):

3388-3402.

[5] DERR F. Coherent Optical QPSK Intradyne

System: Concept and Digital Receiver Realiza-

tion[J]. Journal of Lightwave Technology, 1992,

10(9): 1290-1296.

[6] TAYLOR M G. Coherent Detection Method

Using DSP for Demodulation of Signal and

Subsequent Equalization of Propagation Im-

pairments[J]. IEEE Photonics Technology Let-

ters, 2004, 16(2): 674-676.


11/12


[7] LEVEN A, KANEDA N, KLEIN A J, et al . Re-

al-Time Implementation of 4.4 Gbit/s QPSK

Intradyne Receiver Using Field Programmable

Gate Array[J]. Electronics Letters, 2006, 42(24):

1421-1422.

[8] FLUDGER C R S, DUTHEL T, VAN DEN BORNE,

D, et al . Coherent Equalization and POLMUX-

RZ-DQPSK for Robust 100-GE Transmission[J].

Journal of Lightwave Technology, 2008, 26(1):

64-72.

[9] LEE J H, JUNG D K, KIM C H, et al . OSNR

Monitoring Technique Using Polarization-

Nulling Method[J]. IEEE Photonics Technology

Letters, 2011, 13(1): 88-90.

[10] WINZER P J. High-Spectral-Efficiency Optical

Modulation Formats[J]. Journal of Lightwave

Technology, 2012, 30(24): 3824-3835.

[11] WINZER P J. Beyond 100G Ethernet[J]. IEEE

Communications Magazine, 2010, 48(7):26-30.

[12] CHANDRASEKHAR S, LIU Xiang. OFDM Based

Superchannel Transmission Technology[J]. Jo-

urnal of Lightwave Technology, 2012, 30(24):

3816-3823.

[13] CHANDRASEKHAR S, LIU Xiang, ZHU Benyuan.

Transmission of a 1.2-Tb/s 24-Carrier No-

Guard-Interval Coherent OFDM Superchannel

over 7200-km of Ultra-Large-Area Fiber[C]//

Proceedings of the 35th European Conference

on Optical Communication: September 20-24,

2009. Vienna, Austria, 2009-Suppl.: 1-2.

[14] LIU Xiang, CHANDRASEKHAR S, WINZER P J,

et al . 1.5-Tb/s Guard-Banded Superchannel

Transmission over 56x 100-km (5 600-km)

ULAF Using 30-Gbaud Pilot-Free OFDM-

16QAM Signals with 5.75-b/s/Hz Net Spectral

Efficiency[C]// Proceedings of the 38th Euro-

pean Conference and Exhibition on Optical

Communication (ECOC’12): September 16-20,

2012. Amsterdam, the Netherlands, 2012:

Th.3.C.5.

[15] RENAUDIER J, BERTRAN-PARDO O, MARDO-

YAN H, et al . Spectrally Efficient Long-Haul

Transmission of 22-Tb/s Using 40-GbaudPDM-16QAM with Coherent Detection[C]//

Proceedings of Optical Fiber Communication

Conference and Exposition and the National

Fiber Optic Engineers Conference (OFC/NF-

OEC): March 4-8. 2012. Los Angeles, CA, USA,

2012: OW4C.2: 1-3.

[16] RAYBON G, RANDEL S, ADAMIECKI A, et al .

1-Tb/s Dual-Carrier 80-Gbaud PDM-16QAM

WDM Transmission at 5.2 b/s/Hz over

3 200km[C]// Proceedings of IEEE Photonics

Conference (IPC): September 23-27, 2012. San

Francisco, CA, USA, 2012: 1-2.

[17] RAYBON G, ADAMIECKI A L, WINZER P J, et al .

All-ETDM 107-Gbaud (214-Gb/s) Sin-

gle-Polarization QPSK Transmitter and Co-

herent Receiver[C]// Proceedings of European

Conference and Exhibition on Optical Com-

munication (ECEOC’12): September 16-20,

2012. Amsterdam, the Netherlands, 2012:

We.3.A.2.

[18] GAVIOLI G, TORRENGO E, BOSCO G, et al . In-

vestigation of the Impact of Ultra-Narrow

Carrier Spacing on the Transmission of a

10-Carrier 1 Tb/s Superchannel[C]// Pro-

ceedings of Optical Fiber Communication

Conference and Exposition and the National

Fiber Optic Engineers Conference (OFC/N-

FOEC): March 21-25, 2010. San Diego, CA,

USA, 2010: OThD3.

[19] BOSCO G, CARENA A, CURRI V,et al

. Per-formance Limits of Nyquist-WDM and

CO-OFDM in High-Speed PM-QPSK Sys-

tems[J]. IEEE Photonics Technology Letters,

2010, 22(15): 1129-1131.

[20] SCHMALEN L, BUCHALI F, LEVEN A, et al . A

Generic Tool for Assessing the Soft-FEC Per-

formance in Optical Transmission Experi-

ments[J]. IEEE Photonics Technology Letters,

2011, 24(1): 40-42.

Biographies

ZHANG Xiaohong, is currently a Senior Product

Manager for WDM product in Optics China. He

works for the OTN product management and fo-

cuses on product strategy definition and product

evolution.

YI Xiaobo, is a Senior System Consultant Engineer

and DMTS in Optics China. He works for the system

and technology team and focuses on high speed

optical transmission, packet transmission network

and related transmission architecture design for wire

line and wireless network backhauling.

LIU Xiang, is a Distinguished Member of TechnicalStaff at Bell Labs, Alcatel-Lucent, USA. He received his

Ph.D. degree in applied physics from Cornell Univer-

sity, USA. Since joining Bell Labs in 2000, LIU has been

primarily working on high-speed optical communica-

tion technologies including advanced modulation

formats, coherent detection schemes and fiber non-

linear impairment mitigation. Recently he was

recognized as a core member of the “100Gb/s Co-

herent (Long Haul–High Capacity WDM Interface)

Team” and was awarded the 2010 Bell Labs Presi-

dent’s Award. Dr. LIU has authored/coauthored more


12/12


than 250 journals and conference papers. He holds

over 45 US patents. Dr. LIU is a Fellow of the OSA and

an Associated Editor of Optics Express. He has served

in technical committees of various conferences such

as OFC, ACP, FiO, OSA and IEEE Summer Topical

Meetings.

Peter J. Winzer, received his Ph.D. in electrical engi-

neering from the Vienna University of Technology,

Austria in 1998. Supported by the European Space

Agency, he investigated space-borne Doppler lidar

and laser communications using high-sensitivity

digital modulation and detection. In 2000, he joined

Bell Labs, USA, focusing on many aspects of fiber

optic networks, including Raman amplification, opti-

cal modulation formats, advanced optical receiver

concepts, digital signal processing and coding, as

well as on robust network architectures for dynamic

data services. He demonstrated several high-speed

and high-capacity optical transmission records from

10 to 100 Gb/s and beyond, including the first 100G

and the first 400G electronically multiplexed optical

transmission systems and the first field trial of live100G video traffic over an existing carrier network.

He has widely published and patented and is actively

involved in technical and organizational tasks with

the IEEE Photonics Society and the OSA. He was

promoted Distinguished Member of Technical Staff at

Bell Labs in 2007 and since 2010 he has headed the

Optical Transmission Systems and Networks Research

Department. He is a Fellow of the IEEE and a Fellow of

the Optical Society of America (OSA).

100g transport systems

Documents