ICTON 2007

ICTON 2007

171

Tu.D1.5

Tu.Dl .5

Ultra High-Speed Optical Transmission based on LDPC-Coded Modulation and Coherent Detection for Employment in All-Optical Network Scenario
Ivan B. Djordjevic, Milorad Cvijetic*, Lei Xu**, Ting Wang** University ofArizona, Department of Electrical and Computer Engineering, Tucson, AZ, USA *NEC Corporation ofAmerica, Herndon, VA, USA * *NEC Labs ofAmerica, Princeton, NJ, USA Tel. (520) 626 5119, Fax. (520) 626 3144, e-mail. ivan@,ece. arizona. edu ABSTRACT The coded modulation schemes enabling 100+ Gb/s transmission by using commercially available components operating at 40 Giga symbols/s have been proposed and discussed in this paper. We considered bit-interleaved coded modulation (BICM) in combination with the low-density parity-check (LDPC) codes that are used as component codes, and in combination with coherent detection scheme. It was shown that coherent detection scheme outperforms the direct detection one and provides additional margin needed for application in an alloptical network scenario. Keywords: Bit-interleaved coded modulation (BICM), low-density parity-check (LDPC) codes, direct detection, coherent detection, optical communications. 1. INTRODUCTION Unprecedented rate of Internet traffic growth will need deployment of optical transmission systems with bit rates higher than rate of currently available 40 Gb/s systems, and this was a subject of several papers published recently [1-7]. However, at those data rates, such as 100 Gb/s, signal quality is significantly degraded mainly due to impact of polarization mode dispersion (PMD), and intra-channel nonlinear effects. Although, the main components related to electrically time-division multiplexed (ETDM) transceivers operating at -100 Gb/s are becoming commercially available, they are still expensive. On the other hand, there is a possibility to use commercially available components operating at lower speed as an alternative approach to enable transmission at 100 Gb/s and beyond. In this paper we present and analyze 100+ Gb/s transmission scenario by using commercially available components operating at lower speeds. The main element in our approach includes bit-interleaved coded modulation (BICM) scheme (presented in Section 2) combined with low-density parity check (LDPC) codes as component codes [1]. In our approach, modulation, coding, and multiplexing are performed in a unified fashion, which means that transmission, and all signal processing related to detection and decoding are effectively done at much lower symbol rates (e.g. 40 Giga symbols/s). In such a way, dealing with the nonlinear effects and polarization-mode dispersion (PMD) is more manageable, while the aggregate data rate is maintained at or above 100 Gb/s. Above architecture is combined with coherent detection to gain an additional power margin to be used against impacts of networking effects in an all-optical network scenario. 2. BIT-INTERLEAVED LDPC-CODED MODULATION (BI-LDPC-CM) Herewith we analyze the transmitter architecture employing BI-LDPC-CM scheme in combination with the coherent detection scheme. The analysis related to direct detection has been also performed as a reference case (please see Fig. la,b,c). The source bit streams coming from m information sources (e.g., carrying 40Gb/s traffic) are encoded by using identical (n, k) LDPC codes of code rate r = k I n (k-the number of information bits, n-the codeword length). The LDPC decoder outputs are written to the mxn block-interleaver row-wise. The mapper accepts m bits, c= (CI,C2,. ,cm), at time instance i from the (m x n) interleaver column-wise and determines the corresponding M-ary (M= 2m) constellation point si= (Ii, Qi) = Isilexp(j O) (please see Fig. la). In coherent detection case the data phasor Oie {0 ,2i M ..,2i(M-l) M} is sent at each ith transmission interval. On 0 the other side, in direct detection, which serves here as a reference case, the differential encoding is required so that the data phasor Q,= ,-,1A0, where AO, E {0,2icM,..,2n(M-1)M}, is sent instead at each ith transmission interval. The receiver input electrical field at time instance i for an optical M-ary differential phase-shift keying (DPSK) receiver configuration from Fig. lb is denoted by Ei=JEiJexp(joy). The outputs of I- and Q-branches (upper and lower-branches in Fig. lb) are proportional to Re{EiE*i-13 and Im{EiE*i-13, respectively. The corresponding coherent detector receiver architecture is shown in Fig. Ic, where S. = |S| efiSi (.f i =w t + Oi + s-PN ) is

coherent receiver input electrical field at time instance i, while L =

ILI ei'L

( oL = WLt + oL PN) is the local laser

1-4244-1249-8/07/$25.00 (C2007 IEEE

Tu.D1 .5

172

ICTON 2007

electrical field. For homodyne detection the frequency of the local laser (c) is the same as that of the incoming optical signal (c), so the balanced outputs of I- and Q-channel branches (upper- and lower-branches of Fig. 1 (c)) can be written as
v(t)=R|Sk

VQ(t)

ILIcos((o,osP, -fLP) (i-o)T + R Sk L sin ((q+yPN-L PN) (oi-1) os

. <t<iT

(1)

< t < iT

where R is photodiode responsivity, while (S,PN and (oL,PN represent the laser phase noise of transmitting and receiving (local) laser, respectively. These two noise sources are commonly modeled as Wiener-Levy process [8], which is a zero-mean Gaussian process with variance 2n1(Avs+AvL) tl, where Avs and AVL are the laser linewidths of transmitting and receiving laser, respectively. The transmitted signal constellation point si Isil exp(j O) can be perfectly recovered in accordance with Eq. (1) only in the absence of laser phase noise.
Source channels

1
m

LDPC eIncder LDPC enco6der

m -<

MZM

to fiber

(a)
T

so,Re{E
>

E_ i}

from fiber

,~

LDP Decde

from fiberLDCD (b)

Ei=lEle

*lR

L Si

Lcos(S

PL)

L e'fL

VmS1*

S1 L.n(S

QL)

(c)

Fig. 1 Bit-interleaved LDPC-coded modulation scheme. (a) transmitter architecture, (b) direct detection as a reference case, (c) coherent detection receiver architecture. Ts1IR5, R, is the symbol rate.
The outputs at I- and Q-branches (in either coherent or direct detection case), are sampled at the symbol rate, while the symbol log-likelihood ratios (LLRs) are calculated in a posteriori probability (APP) demapper block as follows (2) Fromloc(al rs), where P(s r) is determined by using Bayes' rule

P(S r)-= EP(r)sp)P(psS)

(3)

Please notice that se(Ii, Q) (as mentioned earlier) is the transmitted signal constellation point at time instance i, while r(ru , rQ), r1 vi(t iTs), and rQb VQ(te iTs) are the samples of I- and Q-detection branches from Fig. lb,c. P(rls) from eq. (3) is estimated by evaluation of histograms, employing sufficiently long training sequence. With F(s) we denoted the a priori probability of symbol s, while s0 is a referent symbol. We need to mention that normalization in (2) is introduced to equalize denominator from (3). The bit LLRs cj ( = 1, 2,..., m) are determined from symbol LLRs of eq. (2) as

ICTON 2007

The APP demapper extrinsic LLRs (the difference of demapper bit LLRs and LDPC decoder LLRs from previous step) for LDPC decoders become

With LD,e(C) we denoted LDPC decoder extrinsic LLRs which is initially set to zero value. The LDPC decoder is implemented by employing the sum-product algorithm [9]. The LDPC decoders extrinsic LLRs (the difference between LDPC decoder output and input LLRs), LD,e, are forwarded to the APP demapper as a priori bit LLRs SO (LMa)o that the symbol a priori LLRs are calculated as

By substituting eq. (6) into eq. (3) and then eq. (2), we are able to calculate the symbol LLRs for the subsequent iteration. The iteration between the APP demapper and LDPC decoder is performed until the maximum number of iterations is reached, or the valid code-words are obtained.
1o-1 J
2

102 10-- '

-~0 1
10*= 10-7]

10878-

10-94

Optical SNR, OSNR [dB 0.1 nm] (a)

>4-tg',*^~ 8-PSK:
173

Tu.Dl.5

L( 1) I LQ )=logt

] Slog =1 exp[pZ(s)4(exp[A(s)]4)
0

LMe (Cj )=L (C8)-LD, (cj

(5)

'a (s) =logP(s)

m-l

j=O

(1-Cj)LDe (Cj)

(6)

8-DPSK: Uncoded
LDPC(4320,3242), 40 Giga symbols/s (Gray mapping) --*\> , LDPC(4320,3242)-coded OOK (120 Gbs)
* ~ +<r <FBI-LDPC-CM (3bits/symbol), 10hiter

0.

.^

10o-

|I-\ 8S_

l 7

BI-LDPC-CM(3bits/symb LDPC(4320,3242), 40 Giga

Uncoded

symbols/s:

iter

-6

--* Gray mapping QPSK (Gray mapping):

LDPC(4320,3242)-coded, 50 Giga symbols/s

10

12

14

16

18

20

*10

8-DPSK:

12
Io-'
10 4 ~~~~~ ~ ~ ~ ~ 0 *~ @,,/

BI-LDPC-CM,

/|

<

BI-LDPC-CM, 5th iter.: ~~>Natural, Av= 1 MfI{z)

10Natural - 8-PSK: -A- Uncoded, Av= 1 MI{z /- -Uncoded, Av= 00 kHz

-E-Gray

5th iter.:

i o05

Natural, Av 100 kHz)

10-6

10-7-

12

16

Number of spans, N

20

24

28

32

(b)
Fig. 2 BER performance ofBI-LDPC-CM/coherent detection scheme. (a) on an A WGN channel modelfor 8-PSK, and (b) for dispersion map from [3].

The results of simulations for 30 iterations in the sum-product algorithm and for an additive Gaussian noise (AWGN) channel model and 8-DPSK/8-PSK schemes are shown in Fig. 2a. The information symbol rate is set

Tu.Dl .5

174

ICTON 2007

to 40 Gsymbols/s, while 8-DPSK/8-PSK is employed in simulations, which produces an aggregate bit rate of 120 Gb/s. Two different mappers are considered: Gray, and natural mapping. The expected coding gain for 8-DPSK at BER of 10-12, obtained by extrapolation, is about 12.8dB. The line rate of considered scheme is 53.4 Gsymbols/s. The coherent detection scheme provides an additional improvement of about 2.5 dB. The results of Monte Carlo simulations for the dispersion map same as in [3] are shown in Fig. 2b. The dispersion map under consideration has N spans of total length L = 120 km, where each span consists of 2L/3 km of D+ fiber followed by L13 km of D fiber. The pre-compensation of -1600 ps/nm is also employed. The simulations were carried out with the average transmitted power per symbol of 0 dBm, the central wavelength is set to 1552.524 nm, while 8-DPSK/8-PSK with RZ pulses of duty cycle 3300 are considered. The propagation of a signal is modelled by the nonlinear Schrodinger equation. The effects of self-phase modulation, nonlinear phase-noise, intrachannel cross-phase modulation, intrachannel four-wave mixing, stimulated Raman scattering, chromatic dispersion, laser phase noise, ASE noise and intersymbol interference are all taken into account. While, by using BI-LDPC-CM and direct detection in a point-to-point transmission scenario it was possible to achieve the transmission distance of 2760 km at 120 Gb/s aggregate rate with LDPC codes with BER threshold of 10-2, the coherent detection scheme is able to extend the transmission distance by about 600 km. It was shown that laser phase noise did not have a major impact on transmission performance. It was found that the coherent detection outperforms corresponding direct detection scheme by achieving an additional margin of 2.5 dB. This is very significant result from optical networking perspective if we consider transmission in an all-optical environment where a number of reconfigurable optical add-drop muxes (ROADM) and/or wavelength crossconnects (WXC) may be employed at the wavelength path. We assume that an additional margin created by employment of coherent detection scheme is enough to keep the same distance as one attributed to direct detection scheme in point-to point transmission scenario, but this time with inclusion of all optical networking constraints, i.e. penalties due to deployment of ROADMs and WXC. 3. CONCLUSIONS Ultra high-speed transmission architecture, which employs a bandwidth-efficient LDPC-coded modulation, BILDPC-CM, and coherent detection scheme, has been proposed. In this architecture the aggregate bit rate above 100 Gb/s is maintained, while modulation, coding, signal processing, and transmission are done at 40 Gb/s, which makes the implementation easier, and the impact of nonlinearities and PMD is less important. Moreover, once the ETDM technology at 100 Gb/s becomes mature enough, the schemes considered in this paper can be used to achieve transmission at much higher rates that 100 Gb/s. It was found that the coherent detection scheme outperforms the direct detection one by enabling an additional 2.5 dB margin to be used either to extend the transmission distance by more than 20%, or to compensate for penalties due to networking effects in an all optical network deployment scenario. ACKNOWLEDGEMENTS The authors would like to thank B. Vasic from University of Arizona for his involvement in earlier work on high spectrally efficient schemes for high-speed optical transmission REFERENCES [1] I.B. Djordjevic, et al.: Low-density parity-check codes for 40 Gb/s optical transmission systems, IEEE J. Sel. Top. Quantum Electron., vol. 12, no. 4, pp. 555-562, July/Aug. 2006. [2] I.B. Djordjevic, and B. Vasic: 100 Gb/s transmission using orthogonal frequency-division multiplexing," IEEE Photon. Technol. Lett., vol. 18, no. 15, pp. 1576-1578, Aug. 2006. [3] I.B. Djordjevic, M. Cvijetic, L. Xu, and T. Wang: Proposal for beyond 100 Gb/s optical transmission based on bit-interleaved LDPC-coded modulation, IEEE Photon. Technol. Lett., accepted for publication. [4] I.B. Djordjevic, and B. Vasic: Multilevel coding in M-ary DPSK/differential QAM high-speed optical transmission with direct detection, J. Lightw. Technol., vol. 24, pp. 420-428, Jan. 2006. [5] A. Zapata, et al.: Next-generation 100-gigabit metro Ethernet (100 GbME) using multiwavelength optical rings, J. Lightw. Technol., vol. 22, pp. 2420-2434, Nov. 2004. [6] G. Raybon, P.J. Winzer, and C.R. Doerr: 10x107-Gb/s electronically multiplexed and optically equalized NRZ transmission over 400 km, in Proc. OFC Postdeadline Papers, Paper no. PDP32, 2006. [7] M. Daikoku, et al.: 100 Gb/s DQPSK transmission experiment without OTDM for 1 OOG Ethernet transport, in Proc. OFC Postdeadline Papers, Paper no. PDP36, 2006. [8] M. Cvijetic, Coherent and Nonlinear Lightwave Communications, Artech House, Boston 1996. [9] H. Xiao-Yu, et al.: Efficient implementations of the sum-product algorithm for decoding of LDPC codes, in Proc. IEEE Globecom, vol. 2, Nov. 2001, pp. 1036-1036E.