1 CHAPTER 1 INTRODUCTION 1.1 MOTIVATION With the ever-expanding growth of Internet traffic, we are witnessing a new era in telecommunications. This era was ushered when data traffic began to exceed the voice traffic as a result of Internet explosion. The advances in Dense Wavelength Division Multiplexing (DWDM) technology enabled the optical networks to squeeze more data through fiber-optic cables. The bottlenecks faced by the next generation DWDM fiber optic system has been discussed in Dhodhi et al (2001). The increase in the field intensity due to more number of channels (or wavelengths) in the fiber causes a fiber to behave as a nonlinear fiber and bottlenecks (limitations) like Stimulated Raman Scattering (SRS) and Four Wave Mixing (FWM) arise. These bottlenecks will limit the maximum optical power that could be dumped into an optical fiber, limit the number of wavelengths that could be dense wavelength division multiplexed, reduces optical amplifier spacing and increases the intensity of the interfering terms coinciding with the original optical channels. The decrease of the capacity of a communication channel beyond certain input optical power (at which non-linearity is coming into picture) is illustrated in Mitra and Stark (2001). Hence in order to increase the capacity of a fiber optic communication system to meet the growing traffic demand, the limiting factors are to be analysed and improvement methods are necessary.
2 1.2 OBJECTIVES OF THE RESEARCH The objective of research is to remove the bottlenecks (SRS and FWM) in DWDM systems. a) To analyse the Stimulated Raman Scattering effect and explore the method of reducing the optical power tilt in the DWDM optical spectrum. b) To analyse the Four Wave Mixing and explore the method of reducing the interfering terms that coincides with the original wavelengths in the optical spectrum. Any three wavelengths from the input wavelengths mix together to produce the fourth wave (interfering),when it has a value equal to or very nearer to the original wavelengths. A detailed literature review on different bottlenecks has been done and presented in this thesis. A study on optical power tilt for the three cases of Unequal channel spacing in the DWDM Spectrum has been performed. 1.3 LITERATURE REVIEW 1.3.1 Bottlenecks in DWDM The unwanted system limitations arising out of the increased number of wavelengths and increased optical power of the individual channels are described in many articles. Procedures are presented for estimating the critical powers for nonlinear optical processes in single-mode fiber in Rogers (1980). The effect of GVD on Stimulated Raman Crosstalk is illustrated in the Literature
3 Cotter (1984). General expressions in the early years are derived to estimate transmitter power limitations due to Stimulated Raman Scattering in WDM system by Chraplyvy (1984). The effect of variation of the maximum input power with the number of channels due to FWM is dealt in the Literature Waarts and Braun (1986). The optical system power limitations on lightwave communications imposed by optical fiber nonlinearities and the effect of bit pattern in the adjacent channels on the power transfer from lower wavelength channels to higher wavelength channels are dealt by Chraplyvy (1990). The mechanism of Stimulated Brillouin Scattering and its degradations are described in Daniel (1993). A specific fiber dispersion management allowing the simultaneous suppression of linear dispersion penalties and degradations arising from fiber nonlinearities is proposed by Christian Kurtzke (1993). For every channel, there is a maximum allowable input power. Xiupu et al (1994) showed that if the applied input optical power is exceeding this limit, then SRS cross talk arises. The exploitation of optical wavelength division multiplexing and the SRS interaction between subcarriers in Subcarrier Multiplexing (SCM) video distribution system is envisioned in the Literature Wang et al (1995). Suitable algorithm has been developed by Fabrizio et al (1995) for the optimum unequal channel spacing to reduce the FWM effect. An exact analytical solution for the evolution of Stimulated Raman Crosstalk in a single mode WDM fiber system is presented by Christodoulides and Jander (1996). The techniques of WDM and OTDM for increasing the capacity as well as switching and routing is explained well in the Literature David (1997). The effect of group velocity dispersion on stimulated Raman crosstalk in intensity modulated multichannel transmission systems is investigated in Jinsong (1998).
4 The influence of laser in optical fiber transmission system is dealt by Tiwari et al (1999). The effect of dispersion compensation on system performance with the nonlinearities was dealt by Pipipetskii et al (1999). The elimination of Stimulated Raman Scattering using spectral inversion techniques is demonstrated in Alexandra et al (1999). The impact of stimulated Raman scattering on the power distribution of a 32-channel multiplex after 100-km transmission over various fiber types are measured by Sebastien et al (1999). Anujkumar (1999) discussed about the anticipated problems and probable solutions in WDM on existing optical fiber. The Stimulated Raman crosstalk variance in WDM systems is derived analytically in a closed form formula for all systems with different walkoff length in Keang (2000). The simple expressions for transmission limitation of WDM transmission systems with dispersion compensated links in the presence of fiber nonlinearities are derived by Ivan (2001). The information capacity and the information density of communication systems using optical amplifiers in the direct intensity modulation case is studied by Mecozzi and Shaif (2001). The trends in Optical Networks like paving the way for increased capacity and more intelligence in optical networks are examined in the Literature Erman (2001). The dispersion free fiber is not suitable for DWDM fiber optic communication system as proved by Jau Tang (2001). The FWM noise power for channel spacings like 1 nm, 2 nm etc., have been done. However, to meet today s traffic demand, the channel separation of 0.5 nm or even less is common in the DWDM signal. Evgenii and Mitra (2002) have given correction to the channel capacity due to the nonlinearity of the fiber. In nonlinear channels, an increase of the signal power would not necessarily improve the system capacity as demonstrated by Turitsyn et al (2003). In this thesis, investigations are being done with 0.5 nm and less than 0.5 nm channel spacing. The program has been done in Matlab. To see the real perspective of an optical communication
5 system, it has been simulated in Optsim (an optical communication simulation software) so that the actual DWDM spectrum at the output of the fiber is visualized. The explosive growth of Internet is told in a nutshell in his Guest Editorial by Cheung (2003). The emerging broadband access networks in Korea is dealt in depth indicating the data traffic demand by Yong and Dongmyun (2003). The supercontinuum generation and multiwavelength processing is given in the Literature Thomas (2003). The novel analytical model to assess the signal quality in nonlinear dense wavelength division multiplexed transmission is presented in Hadrien (2003). The Raman gain coefficient in G.652 (standard single mode fiber) based on the power transfer between a high intensity pump signal and on a counter-propagated broadband probe signal technique is measured in Paulo (2003). Singh and Hudiara (2004) has given a piece wise linear solution for nonlinear SRS effect in DWDM fiber optic communication systems. The optical power transfer among the different wavelength channels, while DWDM optical signal is propagated through an optical fiber is also clearly explained. The factors affecting the physical layer of the optical network are dealt by Singh et al (2004). The SRS effect is evaluated in a piece wise linear manner. Here the equal channel spacing is set for the DWDM signals. We have proposed an unequal channel spacing with three cases like ascending order, descending order and random order. Optical nonlinearities like SRS and FWM present in the single mode fiber have been thoroughly reviewed with recent examples by Toulouse (2005). The use of prechirped pulses to reduce FWM is investigated in Neokosmidis (2005). A comparative study of the Shannon Channel capacity is presented for a dispersion-free fiber, a fiber with constant dispersion and a fiber with variable dispersion in the Literature Jau Tang (2006). The Stimulated Raman Scattering phenomenon is explained
6 with a neat illustration in Agrawal (2006). The applications of super DWDM technologies to terrestrial Terabit transmission systems are described by Hiro et al (2006). The Interchannel separation in WDM Transmission systems in the presence of fiber nonlinearities is optimized by Gurmeet and Singh (2007). The variation of the power depletion with the number of channels for different peak input power per channel is described in Ataur (2007). The capability of an optical communication in the next decade and Long-term outlook is well analysed in the Literature Masahiko (2007). An investigation of the power penalties imposed by four-wave mixing on G.652 (Single Mode Fiber-SMF), G.653 (Dispersion Shifted Fiber-DSF) and G.655 (Non-Zero Dispersion-Shifted Fiber-NZDSF) is presented by Paula et al (2008). The Four Wave Mixing (FWM) effect will also affect the wavelength routing and assignment, as investigated by Tan et al (2008). The effect of FWM on transmitted optical power is discussed in Gurmeet et al (2009). The deployment of bandwidth hungry applications such as High Definition Television (HDTV) in access networks is well explained in the Literature Zhang (2009). The effect of dispersion on the FWM power for various channel spacing is illustrated in Amarpal (2009). The FWM power with pump light parameters in the three fibers SMF, DSF, NZDSF are compared in Li et al (2009). The simulation of 32- channel point to point DWDM system and its analysis are carried in the Literature Yan et al (2009). Analysis has been carried out to evaluate the SNR in the presence of Amplifier Spontaneous Emission noise by Patterh et al (2010). Capacity per unit bandwidth (i.e spectral efficiency) for the fiber which encounters nonlinear phenomenon is clearly predicted by Essiambre et al (2010). The impact of fiber nonlinearities in optical DWDM
7 transmission systems at different data rates has been calculated by Gurmeet et al (2010). The necessity of new modulation formats other than the traditional NRZ modulation for the 40 Gb/s system is analysed by Yin et al (2010). 1.3.2 Improvement Methods to reduce Bottlenecks The variation of Spectral efficiency or the capacity per unit bandwidth with respect to the input power per channel for the purely linear noise, purely nonlinear noise and both linear and nonlinear noise are reported in Jason (2001). The increase of the number of wavelengths in WDM system also leads to a substantial increase of the field intensity in the fiber and the corrections to the channel capacity is derived in Jau Tang (2002). As indicated by Turitsyn et al (2003), in nonlinear channels, an increase of the signal power would not necessarily improve the system capacity. Information theoretic limits to spectral efficiency in Dense Wavelength Division Multiplexed (DWDM) transmission systems are reviewed considering various modulation techniques in Joseph (2004). The performance of four different modulation formats in ultrahigh spectral efficient WDM systems are analyzed in the Literature Bosco et al (2004). The nonlinearities are observed as perturbation of a linear case or as the multiplicative noise and the achievable information rates for high-speed optical transmission (40 Gb/s and above) are calculated in Ivan et al (2005). The decrease of channel capacity per unit bandwidth (Spectral Efficiency) due to the increase in the input optical power beyond certain limit in the fiber channel is well explained in the literature Rene et al (2008). Singh and Hudiara (2004) have given a Piece wise linear solution for nonlinear SRS effect in DWDM fiber optic communication systems. The SRS effect is evaluated in a piece wise linear manner. Here the equal
8 channel spacing is set for the DWDM signals. The unequal channel spacing is not analysed here. The effect of FWM on transmitted optical power is discussed in Kaur and Singh (2009). The FWM noise power for channel separations like 1 nm, 2 nm etc., have been done. However, to meet today s traffic demand, the channel separation of 0.5 nm or even less is common in the DWDM signal. The analysis is being done with 0.5 nm channel spacing. The program has been done in Matlab. From the literature study, it is observed that FWM could be reduced by unequal channel spacing. However, the effect of unequal channel spacing on SRS has not been done. This analysis has been made and presented in this thesis. Also the wavelength spacing is made 0.5 nm and less than 0.5 nm. To see the real perspective of an optical communication system, it has been simulated in Optsim (an optical communication simulation software) so that the actual DWDM spectrum at the output of the fiber is visualized. Also the calculation of exhaustive number of interfering terms for equal and Unequal channel spacing for different bandwidth expansion factors is being done in this thesis. An unequal channel spacing with three cases like ascending order, descending order and random order have been proposed for the evaluation of an optical power tilt in the DWDM spectrum.
9 1.4 OVERVIEW OF THE THESIS This thesis analyses the bottlenecks for various settings in the fiber optic link in optsim software. The rest of the chapters are organized in the following manner. Chapter 2 is an introductory section which presents the fundamental concepts of DWDM. It describes the existing DWDM standards and various bottlenecks. The decrease of channel capacity for the increase in the input optical power beyond the threshold power is explained. The root cause of the Stimulated Raman Scattering is presented in Chapter 3. Starting from the spontaneous Raman scattering, how an incoming optical signal interacts with the molecules of silicon, which may lead to an optical power tilt in the DWDM spectrum is explained. Chapter 4 describes the concepts of Four Wave Mixing in the DWDM system. As the number of wavelengths are increased day by day in order to cater the traffic demands, FWM interfering components and their strength are increasing. The methods of mitigating this effect is discussed. Chapter 5 descibes the variation of Spectral Efficiency or the Capacity per unit bandwidth of an optical fiber with dispersion and input power. Chapter 6 presents the conclusion of this research work and possibilities for future enhancement.