Chapter-2 LITRATURE REVIEW 2.1 INTRODUCTION

Transcription

1 Chapter-2 LITRATURE REVIEW 2.1 INTRODUCTION The use of DWDM technology in communication system remodelled the field of communication. This is due to the fact that the usable transmission bandwidth of the optical fiber is so tremendous that it is has the capability of transmitting multi channels over large span. The excessive usage of Internet traffic encouraged network operators to use even higher transmission ability in their terrestrial fiber backbone networks. The escalation of the users bandwidth demand has caused the fast and reliable provisioning of the optical paths to be the main objective of the optical communications service providers. Super- Dense Wavelength Division Multiplexing (SDWDM) technology emerged to help the service providers to handle the users increasing traffic. DWDM increases the carrying capacity of the physical fiber by assigning incoming optical signals to the particular frequencies of light (wavelengths, or lambdas) within a certain frequency. When the channel spacing between the channels is 0.2 nm, then it is called SDWDM. In DWDM, since each input signal is carried independently of one another, therefore, each channel has its own dedicated bandwidth. Thus, all signals reaches at the same time inspite of being broken up and carried in time slots. This chapter presents literature survey on DWDM communication system s performance and applications. The first part, i.e. Part A, deals with the modulation format used in the DWDM communication systems. Part B is based on ring topology of DWDM Networks; Part C is based on WDM and DWDM communication systems. Part D is based on the dispersion compensation techniques and transient effects in EDFA. 26

2 2.2 PART-A In very high capacity optical networks like DWDM, it is necessary to utilize the available bandwidth as efficiently as possible. One of the approaches for efficient utilization of bandwidth is a selection of modulation formats that require less bandwidth. The selected modulation format must have good performance in terms of interference and it is also required that the selected format should also be resistant to non-linearity of fiber and noise. The implementation effort at high speed must comply with economic constraints also. M. Jaworski et al. [22] presented the survey of different modulation formats. The author presented survey on Modulation formats used in trials in terabit DWDM Systems. Kerr Non-linearity and chromatic dispersion of the fiber are essential deteriorating issues of DWDM Systems. The author did simulations of 4 x 40 Gb/s DWDM System over 4 x 80 km distance of single mode fiber followed by Dispersion compensation. Then simulations were done on various modulation formats. Fiber parameters like attenuation, dispersion, dispersion slope and non-linear co-efficient were used in simulations. Several modulation formats were conferred offering significant advantages for transmission in order to overcome fiber-induced degradation. Simulations of 320 km transmission of DWDM signal shows that Carried Suppressed Return to Zero (CSRZ) and Duobinary Phase Modulation (DB-PSK) were inferior to Differential Quadrature Phase Shift Key (DQSK) modulation with respect to the signal degradation due to Kerr non-linearities. A. Sangeetha et al. [23] presented a series of simulations to correlate Non Return to Zero (NRZ), Return to Zero (RZ) and Chirped Return to Zero (CRZ) systems. The author studied the dynamics of CRZ systems. It was observed that CRZ modulation format had many benefits over NRZ modulation format. The performance of WDM systems was not satisfactory at 10 Gb/s over a span of 5000 km or more. At distance more than 5000 km with reasonable power margins, the author used CRZ modulation format and also examined the evolution of individual CRZ pulses. It was observed that the evolution of pulses was 27

3 dominated by chromatic dispersion and was less affected by the non-linearity at optimal power levels. Authors concluded that the BER in CRZ was found to be low when compared to NRZ and RZ. It was found that the use of symmetric dispersion compensation was more important than asymmetric compensation to minimize the effect of non-linearity. A. Sheetal et al. [24] presented the simulation analysis of 40 Gb/s long haul DWDM System with enormous capacity up to 1.28 Tb/s. The simulations were done for Carrier- Suppressed Return-to-Zero (CSRZ), Duobinary Return to-zero (DRZ) and Modified Duobinary Return-to-Zero (MDRZ) modulation formats. A modified modulation format MDRZ was presented in the paper. A maximum transmission distance of 1450 km for 32 channels DWDM system was obtained with MDRZ modulation format using symmetric compensation. V. Bobrovs et al. [25] showed the influence of nonlinear optical effects on the NRZ and RZ modulation signals. The authors demonstrated the effect of non-linear effects to the Non-Return-to Zero (NRZ) and Return-to-Zero (RZ) modulation formats in WDM systems. Simulation results showed Nonlinear Effects (NOE) reliance on dispersion and input power levels. NRZ has merit of ease of generation and less signal bandwidth as compared to RZ format which has advantages of less inter-symbol interference. NRZ systems have nearly zero dispersion value. Newly produced harmonic peak power is twotimes lowered in RZ format systems in comparison to NRZ. J. Dowine et al. [26] investigated experimentally the effectiveness of Maximum Likelihood Sequence Estimation Electronic Dispersion Compensation (MLSE-EDC) for signals with Self Phase Modulation (SPM) and at various dispersion levels. The authors studied Non- Return-to-Zero (NRZ) and Duobinary signal modulated at 10.7 Gb/s and found that the betterment realized with an MLSE-EDC receiver strongly depends on the level of residual chromatic dispersion with the two chosen modulation formats. 28

4 S. Alam et al. [27] suggested a signal part processing approach and ODTR data signals were transmitted through single made transmission systems. The authors proposed methods to evaluate the true average signal loss in fiber optical communications. The single ended measurement had advantages in time, result reliability and processing efforts. The simulation results shows improvement in the system BER after optimization of the detected signals at central wavelengths in the region of 1310 nm. BER was reduced by using the RZ signal generator through electro-absorption modulation technique. S. Bang et al. [28] compared the performance of Q-factor in CPM effects and the FWM power for 32 channel optical soliton transmission and NRZ. The solitons are the pulses of light, even when transmitted over long distances, propagated undistortedly and maintained their format. It was observed that when EDFA gain value was increased, the Q-factor became better and FWM power increases in optical solitons. The Q-factor performance associated with CPM effect was better using a low pass filter in transmission systems. The simulation results shows a the better performance of soliton format as compared to the NRZ format for filter lengths over 1000 km. R. Mu et al. [29] presented comparisons of three different modulation formats, NRZ, NR and CRZ. The authors studied WDM systems with signal-channel data transmission rates of 10 Gb/s. It was found that to obtain a transmission distance greater than 5000 km, the performance of CRZ modulation was good. The authors also studied the evolution of individual of CRZ pulses. It was observed that the evolution of CRZ pulses was dominated by the chromatic dispersion and was also less influenced by the non-linearity at optimal power levels. It was observed by the authors that symmetric dispersion compensation was significant to use as an alternative of the asymmetric compensation to reduce the effects of non-linearity. It was found that the inter channel non-linearities lead to spreading of eye diagrams, but they did not change the CRZ pulse dynamics. 29

5 Anamika et al. [30] used statistical methods to evaluate XPM induced cross talk for N- channel, M-spam WDM system. The authors did a comparative study between DPSK and OOK modulation format for different data rates and duty cycle of the pulse shape. It was observed that XPM crosstalk decreases for both modulations formats when the duty cycle of the pulse was decreased. It was also found that all the pulse shapes suffered a 30% decrease in cross talk compared to NRZ pulse as 1 bit occupy a fraction of the time slot. DPSK modulation format suffers quite less crosstalk because in DPSK modulation format only differential phase of the signal gets changed due to XPM induced phase shift. It was found that RZ-DPSK signal with 33.3% duty cycle suffered lowest crosstalk in comparison to NRZ-DPSK signal, NRZ-OOK signal, RZ-DPSK and RZ- OOK signal with 50% duty. S. Bang et al. [31] discussed the effects of XPM and FWM in DWDM systems. XPM and FWM effects in conjunction with various signal formats in 16 channels DWDM systems was evaluated by the authors. The authors demonstrated the Q factor performance resulting from XPM and FWM effects for the signal formats like NRZ, RZ, DPSK and the RZ - Soliton. It was observed that degradation due to XPM effects can be reduced by increasing the channel separation. It was found in simulations that by using LPF in 16-channel DWDM systems, the Q-factor performance in case of XPM can be improved. It was also found that RZ-Soliton was most efficient among all data formats used for 16 channel DWDM systems. 2.3 PART-B R. Randhawa et al. [32] investigated and compared the various network topologies. Analysis is done with the signal, as it passes through each node in each of the network topology. It was found that in case of ring topology, there was no detectable signal degradation in the ring network. An increase in quality factor was found in a ring network. In the case of bus topology, it was observed that with the increased number of nodes, the 30

6 quality of signal decreases and the power penalty increases. For the star topology, it was observed that the received power values of each node at a same distance from the hub are same and performance was also same. The performance of tree topology was almost indistinguishable to the performance of the ring topology. S. Singh et al. [33] investigated the performance of tree, star and bus network topologies in the presence of optimized SOAs for 10 Gb/s channel. For investigation, OptSim simulation software was used. Factors like optical couplers, number of users supported and Semiconductor Optical Amplifiers (SOAs) were used to correlate the performance of the three proposed optical network topologies. It was concluded by the author that the performance of proposed optical network topologies were degraded when the input power is more than the saturation power of the semiconductor optical amplifier. In all the topologies used for investigation, SOAs were required to support maximum number of users. In tree network topology, it was observed that minimum SOAs and optical couplers supported maximum number of users. But in star network topology, due to splice and insertion loss, the number of users supported were increased. Maximum 27 number of users were supported in bus network topology. Hence, it was concluded in the paper that a low cost solution for users connecting over a large span was optimized SOAs which was implemented in tree network topology. G. Chawdhary et al. [34] presented two different algorithms SGMOS and SGRS in the research paper. The authors firstly described the issue of survivable multicast traffic grooming for WDM bi-directed ring networks. In the study, the authors assumed the signal link failure model. The authors presented SGMOS and SGRS algorithms to groom multicast sessions with the aim to efficiently handle available bandwidth, reduce network resources and protect multicast sessions from the single link failures. It was observed in the simulations that the proposed algorithms consume much less resources like backup links, wavelength and ports. 31

7 V. Kumar et al. [35] studied the consequences of inter channel cross talk at ADM (Add- Drop Multiplexer) of ring network in sweeping bandwidth of filter. Analysis had been done in a ring network that contains four nodes communications signals over two channels at nm at the bit rate of 10 Gbps, ADMs at each node was modeled by using WDM add and drop components. This distance between the nodes was taken 12.5/25 km. To compensate for fiber loss, an ideal amplifier was added just before the node. It was observed that with each sweep iteration, WDM add and drop multiplexer s bandwidth increased from 10 to 40 GHz in steps, thereby, varying the performance of the ring network. The minimum BER was further minimized and the Q-factor, eye opening were increased. Y. Qu et al. [36] analyzed the traffic caring capacity of the DWDM networks and proposed means to enhance the traffic caring capacity of DWDM network using Wave bonding and Perennials treatment to different class of traffic. To elastrator the proposed technique, the authors considered linear bus topology with 3 nodes and ring topology with 5 nodes. It was shown by the authors that proposed technique can solve the fairness issues related with single and multi-hop traffic that reduced the overall traffic carrying capacity of the network. 2.4 PART-C K. Singh et al. [37] analyzed three WDM systems. Analyses were done on four channels CSRZ systems. All the four WDM systems were designed for 10 Gbps. The length of the fiber waved from 40 km to 100 km for simulations. Analyses were done on the dispersion values of the three WDM systems with four channels each. It was found that CSRZ signal was less sensitive to fiber non-linear effects and was immune against transmission impairments. The RZ systems had less dispersion tolerance and spectral efficiency. The NRZ systems had improved dispersion tolerance, but it was affected by ISI so it was found that NRZ systems were not suitable for higher bit rates and long transmission distances. 32

8 J. Li et al. [38] presented a less complicated coherent receiver to refine spectral efficiency in WDM. The two tasks performed at receiver side are Partial-response equalization and Maximum-likelihood sequence detector. The channel with a known response is shaped into an intermediate state using Partial-response equalization and the without estimating channel, it is recovered by Maximum-likelihood sequence detector. By using pre-filtering, the severe inter symbol interference is divided proportionally between the partial-response equalization and MLSD. The feasibility of the receiver-side partial-response shaping relaxes the efforts and requirements on the transmitter side pre-filtering, which allow the mature WDM components to implement the pre-filtering. The authors also introduced the receiver design rules in the presence of revere ISI. The benefits from the proposed solution are then analytically and experimentally demonstrated. H. Tan et al. [39] purposed a reconfigurable all optical format changing from Optical Time Division Multiplexing to Wavelength Division Multiplexing with flexibility in terms of wavelength allocation and channel spacing. The proposed method comprises of Raman amplification based upon a parametric sampling gate and multi wavelength pulse compressor. By using the amplification/compression process simultaneously with parametric sampling gate, the OTDM to WDM conversion is done with an increased degree of flexibility in terms of channel spacing, wavelength allocation and pulse width tenability. The use of Raman amplifier provided energy for parametric sampling without additional EDFAs. The advantages of the proposed reconfigurable all optical conversions were configurability scalability and energy efficiency. A. Aloisio et al. [40] presented a DWDM network data transmission system that had relevance in the underwater neutrino telescope. DWDM network constitute fibers, an optical amplifier and passive optical fibers. The authors measured performance of DWDM optical network using the metrics of BER and OSNR. The authors used three DWDM channels on the same medium. The results showed a high optical power budget around 38 33

9 dbm; as the BER of is the unfavorable case of values of optical power at receiver close to the Loss of Signal (LOS) state. Because of the Raman amplifier, the OSNR did not drop at the signals travel along the optical path. The fiber in the system also increased the system performance due to decreased amount of optical noise fed to the receiver. N. Dahdah et al. [41] evaluated the performance of OPA. As Fiber Optical Parametric Amplifiers (OPA) offer the amplification over larger band widths, the authors evaluated OPAs first as a transmitter amplifier, then as an intermediate amplifier within a transmission line, to ensure that the nonlinear nature of the OPA noise does not introduce any strong limitations. The authors evaluated the performance of a short fiber OPA with Gbps RZ-DPSK DWDM channels, with 2880 km total transmission distance. In the paper, it was demonstrated that a fiber OPA at either the transmitter or an intermediate node induces no further limitation on the maximum transmission distance of Gbps RZ-DPSK DWDM channels. The performance of OPA was not changed even on using perpetrated signals. Y. Malhotra et al. [42] presented a Hybrid Raman-Erbium Doped Fiber amplifier in the research paper. Hybrid Raman-Erbium Doped Fiber amplifier was employed for increased distance in an Ultra-Dense Wavelength Division Multiplexing (U-DWDM) system of 96 channels transmitted at ultra-narrow channel spacing of 0.1 nm. The authors investigated long haul communication in the presence of both hybrid and conventional amplifiers and compared the performance of the hybrid and conventional amplifiers. It was reported that 60 km of SSMF forms an optimal span distance of 2070 km achieved by Raman-EDFA hybrid amplifier as compared to EDFA which achieved a maximum distance of 1794 km that was much less than Hybrid amplifier. The quality factor of db was obtained with Raman-EDFA with a high distance. Using the results, authors recommended Raman- EDFA amplifier for long haul communication with high capabilities. 34

10 I. Trifonovs et al. [43] studied EDFA amplification and dispersion compensation technique and compared it with standard bidirectional DWDM system. In DWDM system, 72.8 km long optical fiber was used to combine and transmit 4 10 Gb/s channels. OptSim 5.0 simulator was used for optimization and simulation of the system. For each channel s output, eye diagram was captured and observed to assess the performance of the proposed design. In comparison to dispersion compensation and EDFA amplification, BER increases nearly to the critical value, also eye opening got worse, but that value was suitable for normal operation of the transmission system. G. Kaur et al. [44] did a simulation of the DWDM optical transmission system to find the optimized value of amplifier separation in a long haul system in the presence of fiber nonlinearities. With the decrease in the amplifier spacing, the SRS and FWM effects increased while Amplifier Spontaneous Emission (ASE) noise reduces. The authors' only objective was not to optimize the amplifier spacing with the help of simulations, but to maximize the efficiency of the system too. To find the optimized spacing between the amplifiers, BER was used to evaluate. An optimized value of BER was found with an amplifier spacing of 50 km. The authors also concluded that with the increase in the number of channels in the system without changing the channel spacing, BER increases. S. Jansen et al. [45] discussed several transmission experiments and aspects of Optical Phase Conjugation (OPC). For OPC, Difference Frequency Generation (DFG) in a Periodically Poled Lithium Niobate (PPLN) waveguide was used. PPLN waveguide offers high conversion efficiency and does not distort the signal during the conversion. To overcome the polarization dependence polarization diversity, structure was employed using OPC. The compensation of chromatic dispersion for 40 Gb/s WDM transmission was presented. For the experimentation, Gb/s channels were transmitted over 800 km of SSMF with EDFA. Then the transparent transmission of 40 Gb/s NRZ format over 800 km of SSMF was discussed. The compensation of interchannel nonlinear impairments is 35

11 discussed in a 4 40 Gb/s WDM ultra long haul transmission line. By adding an OPC to a transmission line with DCF for dispersion compensation, a 50% increase in transmission reach was obtained due to interchannel non-linearity compensation. Over 40 db improvement in Q-factor was reported due to the compensation of nonlinear phase noise impairments. Then transmission of 22 x 20 Gb/s DQPSK WDM was presented in the paper. For the experimentation, OPC was used to compensate for chromatic dispersion and nonlinear impairments for a transmission distance of km. On comparison of OPC aided transmission with conventional optimized DCF, 44% increase in transmission distance was found in the case of OPC aided transmission. K. Kitayama et al. [46] proposed a novel design of DWDM allocation for millimeter-wave fiber radio systems, which adopts the optical Single Side Band (SSB) filtering technique for optical frequency interleaved signals at the receiver side by using a square response of Fiber Bragg Grating (FBG) filter. The authors studied the effect of FBG in DWDM fiberradio systems. They observed that signal fading occurs to a 60 GHz band optical DSB signal near band edge of FBG that results in WDM channel internal of GHz. But in this case, SSB filtering was almost free from the dispersion effect of the filter at the band edge and it required no spectral tailoring to the optical DSB signal at the transmitter side. The authors demonstrated a drastic narrowing of internal to 83.6 GHz by using the ideal square response FBG and the error free DWDM transmission two 60 GHz band, Mb/s DPSK fiber radio signals over 25 km long SMF with the minimal channel internal. Finally, it was shown in the paper that by sectioning the zone 1000 Base Stations (BSs) under the coverage of signal Central Office (CO), proposed fiber-radio was feasible in the micro or Pico cellular DWDM millimeter wave radio system. S. Susskind et al. [47] studied the transmission of a 40 channels DWDM systems at 10 Gb/s. The authors analyzed the DWDM systems without employing optical amplification in any part of the system. Channels are distributed among wavelengths of nm and 36

12 nm. The authors preferred Dispersion Compensation Fiber (DCF) in the link to allow higher reaching of channels and enhancing the system performance. Low input is used to reduce non- linear effects. A transmission distance of 116 km was achieved using 6 km of DCF with dispersion of 250 ps/nm.km. M. Nakazawa et al. [48] presented the technologies, Dispersion Managed (DM) solutions and waveform reshaping in the paper. For DM soliton transmission, the author described 1 Tb/s WDM soliton transmission over 1500 km. The dispersion managed solitons have large power margin and dispersion tolerance. The author described the 1.28 Tb/s linear transmission over 70 km for a single channel for which soliton effects were used to generate a stable femtosecond train. For OTDM terabit transmission, the fourth order dispersion was compensated by prechirping scheme of wave shaping. The author also described the progress in Time Division Multiplexing and Wavelength Division Multiplexing. S. Ming et al. [49] presented the research paper on nonlinear effect and impact of Polarization Mode Dispersion (PMD) on DWDM. Transmission equation of DWDM system consists of PMD, SPM, XPM and FWM were derived. Based on the equations, numerical simulation of an 8 x 40 Gb/s DWDM system has been carried out. On the basis of simulation results, it was found that at low signal power, PMD had more effect as compared to the nonlinear effects. As the signal power becomes large, nonlinear effects become dominant. N. Ahmed et al. [50] proposed technique to implement electronically post compensate fiber non-linearity in long haul, WDM transmission systems with bit rate above 40 Gb/s. The authors validated the technique by performing the simulations of a 42.7 Gb/s RZ and CSRZ modulated WDM system. The author analyzed the proposed technique for return to zero and carrier suppressed return to zero modulation format using two different dispersion maps. In the first dispersion map, each fiber distance consisted of 80 km of Dispersion Shifted Fiber (DSF) followed by standard single mode fiber. In the second dispersion map, 37

13 each spam consisted of 100 km of transmission SSMF followed by 20 km of temped DCF. The proposed technique effectively improves the overall system margin by more than 1.0 db in both RZ and CSRZ WDM transmission systems. M. Pfennighbauer et al. [51] investigated the influence of filter bandwidth and flank steepness of both multiplexing and de-multiplexing filters in DWDM system in the presence of coherent wavelength division multiplexing crosstalk. Using the simulation, author determined the optimum values of bandwidth and shapes of MUX and DEMUX filters in DWDM system. The numerical result for NRZ, RZ and CSRZ were presented in the paper. CSRZ coding showed the best performance followed by RZ and NRZ. Using simulation, it was concluded that steeper filters give better Optical Signal to Noise Ratio (OSNR) values. However, steeper filters reduced tolerance to variation in filter bandwidths. M. Jagayerifer et al. [52] investigated DWDM communication system based on the Fiber Optical Parametric Amplifier (FOPA) as inline amplifier and compared it with EDFA. Parametric amplification was based on an effective four wave mixing process that occur in highly nonlinear fiber. The author derived an analytical expression for the performance of a long haul DWDM communication system using the fiber optic parametric amplifier. An additive noise model was presented for the performance evaluation of long haul WDM transmission links employing FOPAs as inline amplifier. The dominant cause of performance degradation at high power level is the nonlinear effects of transmission fiber comparatively than the nonlinear cross talk due to FOPA. However, if the number of DWDM channels is large, the nonlinear cross talk due to FOPA was reduced by increasing the pump power. It was concluded in the paper that a broadband long haul DWDM transmission system based on inline FOPA is feasible. 38

14 2.5 PART-D At high data rates, Inter-Symbol Interference (ISI) occurs after the photo receiver due to optical dispersion which further results in broadening of the optical pulse and can appreciably limit the reach of installed fibers. To compensate such effects, Electronic Dispersion Compensators (EDCs) can be introduced next to photo receiver and equalizer circuits that have been used in copper-based channels. To compensate fiber dispersion in optical communications, EDCs were introduced. J. Malhotra et al. [53] employed EDC (Electronic Dispersion Compensation) based on minimum mean square error optimization to improve the performance of 32 channels, long haul DWDM optical soliton link. Though DCF (Dispersion Compensated Fiber) was a useful method to negate the accumulated dispersion but it lacks adaptability. Whereas EDC uses the part decisions to correct the error at the decision threshold and thus helps the equalizer in compensating the ISI. The authors used Symmetrical optical dispersion compensation to negate the accumulated dispersion in the fiber span using DCF. Because of the use of EDC, the eye opening improved thereby reducing the ISI. The reduction in BER in the order of 10-7 and a corresponding enhancement of quality factor of 2 db was achieved. The mid span Optical Dispersion Compensation (ODC) mitigates the FWM power by introducing a destructive interface between the first and the second halves of fiber optic link. Q. Yu et al. [54] discussed the potential and future trends of electrical signal processing methods that can be used to mitigate ISI effects within the range of optical applications. One of the techniques used to mitigate the problem of ISI discussed in the paper was adaptive electronic equalization. The authors provided extensive literature on electronic equalization and other techniques to compensate ISI. The author described Linear Equalizer (LE), DFE (Decision Feedback Equalizer), MLSE (Maximum Likelihood Sequence Estimation), MAP (Maximum-A-Posterior) based equalization and adaptation 39

15 equalization techniques. The author also described the EDC for applications of 10 Gb/s in Local Area Network (LAN) and Metro Area Network (MAN). Long haul and ultra-long haul networks Synchronous Optical Network (SONET-EDC) applications were also described for 4 Gb/s and 8 Gb/s in Storage Area Networks (SANS). S. Chandrasekhar et al. [55] compared the performance of a Maximum Likelihood Sequence Estimation (MLSE) receiver with the standard threshold receiver in a dispersionmanaged eight span transmission at 10.7 Gb/s in the presence of fiber non-linearities. It was concluded that the intrinsic property of MLSE receiver can comfort chromatic dispersion that was significantly impaired when the received channel is affected by SPM and XPM. So, the MLSE receiver is beneficial to use in linear DWDM transmission systems because the performance of MLSE is limited in non-linear system due to SPM. Z. Liu et al. [56] investigated the performance of Manchester coded 10 Gb/s optical signal employing the Feed Forward Equalizer (FFE) and Decision Feedback Equalizer (DFE). The capability of chromatic dispersion compensation using FFE-DFE was evaluated under Balanced Detection (BD) and Single Ended Detection (SD) for majority of the Chromatic Dispersion (CD) values. The simulation results indicated an extraordinarily increase in OSNR despite of employed FFE-DFE type and sampling scheme utilized, in the presence of residual dispersion beyond 53 ps/nm. It was also found that in order to improve the receiver sensitivity; the effectiveness of BD was observed less for four samples per bit based equalizer than two samples per bit based equalizer. In order to achieve better Chromatic Dispersion tolerance, two DFE taps were required. It was also observed that the numbers of FFE taps decreased remarkably, when two samples were taken at maximum eye opening points in the first and second half symbol for equalization at the cost of required OSNR. 40

16 S. Chandramouli et al. [57] presented electronic dispersion compensator using linear and nonlinear equalization techniques. The author demonstrated the circuit architecture to perform decision-feedback equalization at multi Giga-byte per second data rate. An Analog approach for the compensation technique was used. The Analog Decision Feedback Equalizer (ADFE) concerted of one-tap non-linear analog feedback filter and four-tap linear analog feed-forward filter and the circuit was designed using 0.18 µm CMOS technology. It was further explained that fully analog EDC with uncloaked decisionfeedback equalization was used to compensate dispersion in 10 Gb/s optical fiber. The EDC was used to compensate chromatic dispersion in SMF and model dispersion in MMF. Using these technique, transmission distances of SMF and MMF were increased to 300 m and 120 km respectively at 10 Gb/s. P. Singh et al. [58] presented the comparison of various dispersion compensation techniques on the base of Q factor and BER. The authors discussed the OPC, FBG, DCF, CDM and AIO (All-In-One) as dispersion compensating techniques. In comparison, it was found that AIO desperation compensation is the best technique for WDM systems, having a bit rate of 40 Gbps as the technique which provides better results in term of Q factor and BER. FBG has worst Q factor at 200 km for 16 channel WDM system. Each channel was transmitted at bit rate of 40 Gb/s. DCF was the next technique to compare after FBG. As FBG provided Q factor of db, db and 6.0 db for channels 1, 8 and 16 at 150 km but channel 16 was completely degraded at 200 km, the OPC gave the worst performance at 150 km and Q factor of 9.0 db for channels 1and 16 and Q factor of 6.0 db for channels 1 and 16 at 200 km. N. Kahlon et al. [59] presented the various dispersion compensation techniques like DCF, EDC, FBG and digital filters. The digital filters were preferred over FBG, as FBG had a very complex architecture. Digital filters when used along with digital processing can compensate chromatic dispersion. All pares lossless optical filters are used for dispersion 41

17 compensation. Lossless all pares filter can approximate any desired phase response although maintaining a constant unity amplitude response. The author also discussed other filters like Band Pass filter, Super Gaussian filters, Butterworth filters and Fabry Perot filters. FIR filters can also be used to compensate polarization mode dispersion and chromatic dispersion. LMS (Least Mean Square) filter is most advanced to the chromatic dispersion and carrier phase noise. The author also discussed Digital Signal Processing (DSP) to compensate dispersion. The author discussed various transforms, equation and multiplexing techniques that can be used to compensate the dispersion. When channels were dropped or added because of system reconfiguration in DWDM systems, power increases or decreases of surviving channels. Dropping channels produces non-linear effects like Brillioun scattering since the threshold for non-linear effects is surpassed by the surviving channels power. Surviving channels powers decreased on adding further channels. Hence, power excursion needs to be limited in order to maintain the quality of the surviving channel. Himshikha et al. [60] analyzed the methods of minimizing transients in a gain of EDFA. Fabry Perot and Ring Laser were the methods presented for limiting the power excursion and to improve the quality factor of surviving channels. The two methods were implemented for adding and dropping of channels. The author used OptSim tool for simulation. For both the techniques, four channels were used. It was concluded that Ring laser reduction method was proved inferior to Fabry perot method. B. Dumas et al. [61] simulated WDM system and analyzed its transient response by dropping and adding of channels. The authors analyzed four configurations. For two configurations, EDFA was used followed by Distributed Fiber Raman Amplifier (DFRA). For the other two configurations, DFRA was used followed by EDFA. For simulations, 20 WDM channels was considered having each channel spacing of 1545 nm-1565 nm. DFRA Hybrid amplifier s transient response was observed faster and larger when 17 channels 42

18 were added/dropped out of 20 channels km long link was used to analyze the transient response for cascading of HAs, whose losses were compensated by 12 HAs using four hybrid configurations. The transient effects were accumulated on cascading HAs. With HA cascades having longer response time, initial disturbances were observed in the transient response followed by steady conditions. The steady state region and initial disturbances were increased with HA cascades. Till third stage of amplification, power excursions were determined by amplifier s relative location. It was more about DFRA + EDFA configuration. In the forward DFRA+EDFA, power excursion value for steady state was observed to be 7.98 db, also, overshoots at 7.75 db were detected in fourth amplification stage. In backward DFRA + EDFA configuration, power excursion value for steady state was observed to be 7.83 db, also, overshoots from 6.41 db to 8.45 db were detected between second and sixth amplification stage. V. Radisavlyevic et al. [62] presented a novel technique for transients caused by signal add/drops in optical communication networks with EDFA. The technique utilized the electronic automatic gain control that combined both feedback and feed forward control of EDFA. The feedback control is a non-linear integral control and feed forward control in steady state gain scheduling technique. The simulation results indicated that the proposed controller completely eliminate steady state photon power transients caused by signal add/dropped with steady state being reached within few milliseconds. O. Tonguz et al. [63] presented an approximate analysis that could be used to predict the performance of power and Signal to Noise Ratio (SNR) equalization for DWDM systems with Erbium-Doped Fiber Amplifier (EDFA) cascades. The author presented the expressions for relating the maximum number of amplifiers, EDFA gain imbalance, bit rate, transmitter power, receiver dynamic range and number of channels. The number of amplifiers supported by power equalization is generally limited by received SNR and provides a 2-5 fold increase compared to an unequal Zed system employing APD receivers. 43

19 In comparison, the number of amplifiers that can be supported with SNR equalization is limited by a residual power imbalance when employing APD receivers and by the residual power imbalances or transmitter power when employing PIN/FET receivers. SNR equalization provided two fold increase a number of amplifiers compared to the unequal zed systems. It was observed that the relative performance of SNR equalization was improved as the number of channels, bit rate and receiver dynamic range increases and was deteriorated as EDFA gain imbalance increases. X. Tang et al. [64] examined the fast transient effect of the cascaded erbium-doped amplifiers in wavelength routed optical networks. The author studied BER and power transient of surviving channels in a wavelength routed optical ring network. The author chose double-fiber unidirectional self-healing ring architecture consisted of three OADMS. Two wavelength nm and nm were applied in network for operation and analysis. It was observed that the transient rate increased fast along with the number of traversing nodes. For 20 khz high adding/dropping frequency case, the power penalty was relatively small and increased along with traversing optical nodes. But for 0.2 khz and 2 khz lower modulated frequency case, the received signal degraded severely. The adding/dropping of frequencies lead to degradations due to transient effect in cascaded EDFAs. T. Sakamoto et al. [65] presented hybrid amplifier that consists of Thulium-Doped Amplifiers (TDFA) and Erbium-Doped Fiber Amplifiers. When the TDFAs and EDFAs are cascaded in series, the total gain of amplifiers becomes a product of gains or losses of two amplifiers. A high gain was achieved in the wavelength range from 1462 nm to 1560 nm by using hybrid amplifiers. Low NF below 7 db was obtained from 1460 nm to 1560 nm by cascading TDFA followed by EDFA and from 1480 nm to 1560 nm by cascading EDFA followed by TDFA. The authors also optimized the gain relations between TDFA and EDFA in terms of minimizing the GV (Gain Variations) in the hybrid amplifier when 44

20 the target wavelength region was from 1460 nm to 1537 nm. The GVR was minimized to less than 0.4 by optimizing the gain ratio. The amplifiers achieved the averaged gain of 20 db, gain excursion of 2 db, total output power of 14.5 dbm and NF of less than 7dB in the wavelength range from nm. W. Naiet et al. [66] proposed a long-haul CATV system designing scheme with EDFA and Raman Fiber Amplifier (RFA) cascaded in transmission link. In the proposed scheme, EDFA was deployed as forward pumping in order to utilize its largest gain and the quality of ensuring high Carrier to Noise Ratio (CNR). RFA was used as backward pumping so as to utilize its distributed structure to get lower non liners indices, including Composite Second Order (CSO) and Composite Triple Beat (CTB) distortion. For experimentation, the authors used G.652 fiber and only one channel of CATV signal was tested. Experiment design in the paper showed that the proposed scheme ensures the clarity of TV image and efficiently control non-linear effects so as to avoid TV channel overlapping. J. Delavaux et al. [67] presented problems related to the architecture of multi-stage EDFAs and advantages of EDFA in optical fiber transmission system. The authors described the features of optical amplifier like high gain, low noise, fiber compatibility, linearity, high optical power, wavelength transparency and wide bandwidth. Generic formation was introduced by the author to characterize the complex multi-stage amplifier. Generic formation was also used to compare and classify different amplifier topologies. Performance optimization technique like gain flattening and navel amplifier configurations were also discussed. The authors also reviewed the uses of multi stage amplifier in optical repeater, pre amplifier and power booster applications. S. Singh et al. [68] proposed gain flatten approach for transmission of sixteen channels at 40 Gb/s WDM system. The authors increased the transmission distance and capacity of an optical system using gain flattens approach of EDFAs at 40 Gb/s. For the arrangement of 60 km SMF, one or more DCF and an EDFA were added at the end. In the paper, authors 45

21 presented three different approaches of gain flattened techniques. The author simulated sixteen channels NRZ-DPSK signal transmission distance over 490 km at 40 Gb/s and the transmission is achieved by gain flattening approach of EDFA. Z. Khaki et al. [69] presented techniques of optimizing an EDFA based WDM system. The proposed method offered better results in terms of gain flattening. In the proposed method, pump power and fiber length was controlled to optimize EDFA for better gain flattening. The optimized value of pump power was the 23 mw and optimized length of fiber was 5 m. The EDFA gave gain flatters of 24±0.299 db from 1546 nm to 1558 nm bandwidth. The simulation results showed output power of db and average noise figure of 6 db. WDM system based on the EDFA had good performance of BER to The authors also compared the optimized technique with an all optical feedback loop inline erbium doped fiber amplifier. All optical feedback loop in inline erbium doped fiber offered a long transmission distance with less gain deviation. A. Willner et al. [70] analyzed the transmission of WDM channels through a cascade of EDFA in long distance links and ring based network with the EDFA non-uniform gain. For long distance system, optimal operating conditions were found for realizing a high SNR for 20 WDM channels faced 0.5 nm apparent. The authors also incorporated optical filters in EDFA cascade. By incorporating optical filters, the authors found optimal cases for passively equalizing more WDM channels maintaining a high SNR. The performance of potential mega meter is realized by using a notch filtered having 3 db bandwidth with wavelength of 1560 nm which was placed after every 20 equalized SNR. S. Singh et al. [71] investigated the post, pre and symmetrical power compensation by using hybrid RAMAN-EDFA. The investigations on the HOA were done by placing it in different positions on the fiber link. By increasing the input power in the three compensation methods, evaluation was done on the basis of BER, eye closure penalty and 46

22 output received power. It was found that RAMAN-EDFA, the post power compensation method performed well as this method gave least BER (10-40 ) and high out power. 2.6 MOTIVATION DWDM Optical networks have revolutionized the data transmission because of low loss, high speed, better bandwidth and high capacity. So a lot of research work is going on in this field since last few years. According to literature survey, it is found that most of the research done in this field is by using less number of nodes, least number of channels, and point to point networks however not much work has been done on ring networks. In this research, we have used ring network with 45 channels having 0.4 nm of channel spacing with bit rate of 10 Gbps. Different modulation formats have been compared on the basis of eye diagram and BER. According to ITU-T standards, Cross talks effect is examined on channel spacing of above 25 GHz yet no evaluation has been done for channel spacing less than 25 GHz. Cross talk effect decreases with the increase in channel spacing. Dispersion in the fiber also happens due to fiber material, non-linearity and the distance travelled by the signal inside the fiber. There is a need to reduce this dispersion by using some techniques. The one technique which is used is Electronic Dispersion Compensation. This technique is used to compensate dispersion due to single mode fiber. The EDC was analyzed for defining the optimum results. EDFA is the most commonly used electronic amplifiers due to its low pump power and high gain. Signal is amplified using optical amplifiers and recovered from fiber losses, thus making optical amplifiers as the essential components of the optical network. Therefore, EDFA behavior in a DWDM ring network needs to be investigated. Power transient effect was also analyzed during this research. 47

23 2.7 OBJECTIVES OF THESIS WORK Implementation of Multichannel, Multi-Gigabits/Sec speed DWDM ring. Optimization of DWDM Ring using various Modulation Formats. To enhance the distance of the optimized DWDM ring using electronic equalization technique. To observe the effect of power transients in cascaded EDFAs in the proposed DWDM ring. 2.8 ORGANISATION OF THESIS This thesis includes seven chapters. An outline of each chapter is given below: The first chapter gives an introduction to optical fiber, WDM, DWDM, TDM, Ring network, OADM and various modulation formats. Chapter 2 is dedicated to the literature survey related to WDM, DWDM, Ring network, Dispersion compensation techniques and hybrid optical amplifiers. Chapter 3 is based on the investigation of 45 channels OADM DWDM Ring system on the basis of cross talk and channel spacing. Chapter 4 includes a performance analysis of the OADM DWDM ring for different modulation formats. Chapter 5 presents an electronic equalization technique used in the optimized OADM DWDM Ring to enhance the distance. 48

24 Chapter 6 presents a short comparison of hybrid optical amplifiers and the effect of power transients in cascaded EDFAs in OADM DWDM Ring. Chapter 7 concludes the work carried out in this thesis with the future perspective of the OADM DWDM Ring network METHODOLOGY As per the given objectives, Literature review has been carried out to study the performance optimization of Multichannel, Multi-Gigabits per second DWDM Ring using optical software. Further, the DWDM Ring system will be analyzed for Crosstalk and channel spacing. The designed DWDM Ring system will be analyzed by using various modulation techniques. The performance analysis to enhance the distance of the optimized DWDM Ring using electronic equalization technique. A study of short comparison for different HOA and to observe the effect of power transients in cascaded EDFAs in proposed DWDM Ring. Technical programme /conduct of Research /Major Equipment /Lab work Facilities needed Various facilities required for the proposed work are as follows: 1. OPTSIM 5.3 Simulator. 49

25 2. Various books from library. 3. Referred journals and publications. 4. Internet facility. 50