Tunable fiber Bragg gratings (TFBGs) are investigated for single-channel and for dual-channels drop applications. The TFBGs are capable of dropping single DWDM channel or two consecutive DWDM channels in an 8-nm bandwidth in the C-band while allowing the rest of the C-band channels to passthrough. Extensive stability, reproducibility and system characterization in addition to numerical simulations were conducted to investigate the feasibility of using TFBG for ROADM applications in metro-edge optical communication networks.

An overlaid technique utilizing DWDM channels for the capacity upgrade of already installed full-mesh CWDM AWG-STAR systems has been proposed and investigated experimentally. In the experiment, SOA transceivers were examined for simple and cost-effectiveness point of view, and it revealed the potential of over 10 km transmission. In the proposed configuration, relatively high total insertion losses resulted from additional AWGs for DWDM channels caused fewer margin for the loss budget. Our results suggested that SOAT with higher gain and saturation power, or receivers with much higher sensitivity should be important role for this configuration in order to achieve much longer transmission distance.

The recent perspectives of evolution of the access part driven by the Fibre-To-The-Home technology will create new needs at the metro area in the next decade. This paper describes different alternatives for the metro area already proposed or investigated in the past, and points some interesting research directions in this metro area.

This paper presents the Photonic Integrated Extended Metro and Access Network (PIEMAN), an access network which integrates the optical access and metro parts of a network, thereby simplyfying the network and reducing cost. Key features are symmetrical downstream and upstream data rates of 10 Gbit/s, an all optical reach of 100 km, 32 wavelengths and a 512 fold split per wavelength.

We propose a cost effective optical packet switching system using shared wavelength converter pool. We also present its cost model and show the overall system cost for the target performance. The benefits of the proposed system is compared with the conventional approaches in terms of blocking performance and overall system cost.

Due to the recent outstanding advancement of optical networking technology, pervasive Grid computing will be a feasible option in the near future. As Grid infrastructure, optical networks must be able to handle different Grid traffic patterns with various traffic characteristics as well as different QoS requirements. With current optical switching technology, optical circuit switching is suitable for data-intensive Grid applications while optical burst switching is suitable to submit small Grid jobs. However, there would be high bandwidth short-lived traffic in some emerging Grid applications such as multimedia editing. This kind of traffic couldn't be well supported by both OCS and conventional OBS because of considerable path setup delay and bandwidth waste in OCS and inherent loss in OBS. Quasi-Circuit OBS (QCOBS) is proposed in this paper to address this challenge, providing one-way reserved, nearly lossless, instant provisioned wavelength service in OBS networks. Simulation results show that QCOBS achieves lossless transmission at low and moderate loads, and very low loss probability at high loads with proper guard time configuration.

All-optical ultrafast switching is demonstrated in a two-dimensional polystyrene photonic crystal, which is based on the shift of photonic gap under optical pumping. Both the response and recover time constants are measured to be less than 20fs.

The paper studies the performance of an all-optical packet switch (OPS) for different packet length. The packet loss ratio, considered the OPS without FDL and with FDL buffer, with different length distribution are analyzed under the Poisson arrival process and a burst super exponential arrival process respectively. The experiment and analysis show that the performance of different length packet is influence by the packet arrival process and the buffer size. When the OPS without buffer, the PLR under bursty traffic yields the higher packet loss ratio (PLR). PLR with Poisson arrival packet is not influenced by the packet length distribution, while with the super exponential traffic, the packet length distribution influences the PLR, and the fixed length packet yields higher PLR compared to the variable length exponential packet and the experiential Internet traffic. When the OPS with buffer, PLR with Poisson arrival packet yields lower than with super exponential packet under the same load, and the experiential length super exponential packet leads to highest PLR, while the fixed length Poisson process packet brings on lowest PLR.

OOK signals with certain phase modulations can show improved performance without incurring much complexity and cost. In this paper we report recent progress on the performance study of high-speed OOK data signals with different phase-modulation schemes.

This paper presents several ways to implement cost-effective optical duobinary transmitters using low-cost optical modulators. We show that, compared to the conventional scheme, the proposed schemes can be used to implement cost-effective duobinary transmission systems without a significant sacrifice of system performance.

In this paper, we report a wavelength- and time-division reconfigurable optical add/drop multiplexer using a time-space conversion and a MEMS optical switch. We verify the drop function and the reconfigurability of the proposed ROADM. The experimental results show the operability of the 20 channels (5 wavelength channels × 4 time slots) ROADM.

The performance of the OSNR monitoring technique based on the improved polarization-nulling method has been thoroughly evaluated in a highly nonlinear transmission link and a re-circulating loop. This was to verify the possibility of using such technique in the field.

We propose a control technique for erbium doped fiber amplifier (EDFA) gain transient occurring in optical burst network. In the optical burst network, optical packets are generated randomly and packet stream has idle time when no packet exists. This bursty nature of optical burst network causes gain transient in the EDFA with slow gain dynamics. To resolve this problem, the optical power of a channel delivering the burst control packets (BCPs) is modulated based on the information of the burst data (BD) packets contained in the BCP itself to feed constant power to the EDFA. We experimentally demonstrate 2.0-dB and 2.6-dB reductions in power fluctuation and gain transient in EDFA output, respectively. We show the stable transmission of 2.488 Gbit/s rate BCP and 9.953 Gbit/s rate BD packets using the proposed technique. The proposed method can be applied to optical burst switching (OBS) network which adopt out-of-band control channel covering small number of BD channels.

The gray code based all-optical analog-to-digital conversion (ADC) using optical interconnection is described. Recent tremendous growths of optical communications and digital signal processing have encouraged the demand of high-speed and high-resolution ADC. To pursue a high-speed and high-resolution ADC, optical approaches have attracted much attention recently. ADC generally consists of three procedures: sampling, quantization and coding. Whereas the optical sampling techniques have been proposed and realized, the optical quantization and coding techniques have investigated depending on various applications. For the application to the binary detection of a high-speed digital signal, we previously proposed an all-optical ADC which consists of optical quantization using self-frequency shifting in a fiber and optical coding using optical interconnection for general binary code. In addition, since we can easily prepare optical interconnection patterns corresponding to the various codes, this technique can be used in any coding schemes. In this paper, we demonstrate the gray code based all-optical ADC to verify its scalability. Experimental results show that the 8-levels analog signals can be successfully converted into the bitwise allocated 3-bit gray code.

Multi-Protocol Label Switching (MPLS) is a promising technology in next generation network. It is also an important technology to be combined with the Optical Burst Switching (OBS). Forwarding Equivalence Class (FEC) assembly in OBS merging label switching is introduced and FEC assembly at ingress node using Intel IXP1200 is achieved in this paper.

This paper studies the wavelength assignment algorithm in WDM networks. On the basis of discussing the difference between two typical wavelength allocation policies (Complete Sharing policy and Complete Partitioning policy), the dynamic threshold is introduced to solve the restrictions of the Complete Partitioning policy. Then a novel wavelength assignment algorithm based on the dynamic threshold is proposed. Besides, the constraints of dividing the wavelength set into separate subsets are given as well as a series of dynamic threshold values. In the network simulation, the NSF net is chosen as the network model. Two kinds of wavelength assignment algorithms with and without adopting the dynamic threshold are compared. The numerical results show that both the blocking probability and fairness improve by adopting the dynamic threshold.

The complete partitioning policy (CP) for the wavelength resource in optical networks is now widely focused on. The dynamic threshold is one of the ways to make CP policy more efficient. Furthermore, an optimized threshold will be better for reducing the blocking probability and improving the utilization of the wavelength resource. Hence, the genetic algorithm is selected as the optimal policy on virtue of its excellent global search performance for getting optimized value of the dynamic threshold. Moreover, a maximal threshold as the high limit for the dynamic threshold is needed to be decided for making wavelengths shared between different wavelength classes, because the class with higher priority can share its wavelengths with the lower one after its own call setups are satisfied. Therefore, a neural network predictor that can predict the number of the next call setup is designed on the basis of the genetic algorithm to solve this problem. The values of the dynamic threshold and the maximal threshold are calculated, and the simulation results show that they take good effect in reducing the blocking probability and improving the utilization of the wavelength resource.

This paper proposes and experimentally demonstrates the use of a quadruple fiber-optic collimator rather than a dual fiber-optic collimator to form a much improved low optical coherent crosstalk 1×2 fiber-optic add-drop filter. The use of the dual fiber-optic collimator together with a single fiber-optic collimator forms a passive noise rejector that eliminates unwanted optical noises at both output ports. Our experimental results show a much improved > 39.10 dB optical isolation at the Drop port. In addition, the measured average optical losses at the Drop and the Thru ports are 1.40 dB and 1.23 dB, respectively, when only the TF is in the opticap path. A low polarization dependent loss of < 0.17 dB is also determined. Furthermore, our design concept can be used to form a high optical-isolation fixed three-port add-drop filter and a high dynamic-range wavelength sensitive variable fiber-optic attenuator.

This paper shows how an angle multiplexing concept permits one thin-film filter (TF) to multiplex or demultiplex N wavelength optical beams, leading to cost-effective wavelength division multiplexers/demultiplexers (MUXs/DeMUXs). Our first TF-based wavelength MUX/DeMUX structure is arranged in a reflective design in which only the λ_N wavelength optical beam cannot be recombined into the main channel. Fully multiplexing and demultiplexing operations can be accomplished by using our transmissive TF-based wavelength MUX/DeMUX architecture. Experimental proof of concept for our reflective TF-based wavelength MUX/DeMUX structure using one TF, a triple fiber-optic collimator, and an optical circulator separates two wavelength optical beams with their channel spacing of 0.8 nm from the main channel. In this case, measured optical losses of 0.67 dB, 1.66 dB, and 2.59 dB are obtained for the first, the second, and the remaining wavelength optical beams, respectively. Optical crosstalk and polarization dependent loss of < -18 dB and < 0.08 dB are also investigated, respectively.

We use the expression relating the output state of polarization and PMD vector. Based on this expression we get the power fading including first-order PMD and chromatic dispersion, which is dependent on the angle of precession of output state of polarization around the PMD vector. From the expression for power fading, we get the average power penalty for chromatic dispersion and PMD. We propose a novel and fast PMD and chromatic dispersion monitoring technology. Measured results agree well with theoretical analysis.

Using the degree of polarization (DOP) ellipsoid method, differential group delay (DGD) is obtained from limited sampling data in the experiment, which is important for the feed-forward compensation scheme. Results indicate the values of DGD obtained from the experiment are accordant with the theoretical values when the DGDs are less than 30ps.

The performance of 40Gb/s three different formats (return-zero format (RZ) with 33% and 50% duty cycle respectively and carrier-suppressed RZ (CSRZ)) systems are compared when polarization mode dispersion (PMD) is considered. Analytic results indicate that CSRZ modulation format is the best in mitigating PMD and improving system performance among three formats.

We propose a polarization mode dispersion (PMD) compensation scheme for wavelength-division-multiplexing (WDM) system, in which two WDM channels share one PMD compensator at the receiver site. The effect of different modulation formats on multi-stage PMD compensators is studied and compared by numerical simulations in 40-Gb/s WDM optical fiber communication system. The degree of polarization (DOP) of single state of polarization (SOP) optical signal is used as the feedback signal in PMD compensators. The particle swarm optimization (PSO) algorithm is used as the searching algorithm in WDM systems. The compensated DOP values of return-to-zero (RZ) format and nonreturn-to-zero (NRZ) format by multi-stage PMD compensators have been increased distinctly compared to the corresponding cases without compensation. It is shown that the PSO algorithm is implemented successfully in adaptive multi-stage PMD compensation in a 40-Gb/s optical WDM system. The compensated eye diagrams for the two channels by multi-stage PMD compensators indicate that the three-stage PMD compensator which eliminates the influence of second-order PMD completely in WDM system takes the best efficiency in the multi-stage PMD compensators for RZ format and NRZ format.

A simple two-section PMD compensator is proposed for multichannel PMD compensation, which can compensate two or even more channels in DWDM systems simultaneously. It is shown that the outage probability of a 40×43Gbps DWDM transmission system using this compensator is decreased significantly from 3.6E-3 to 3.6E-5.

The simple structure of a tunable polarization mode dispersion (PMD) compensator based on a cantilever beam and a high-birefringence linearly chirped fiber Bragg grating is proposed. A cantilever structure is used to introduce a linear strain gradient on the grating, and we can tune the compensated differential group delay (DGD) at a fixed signal wavelength just by changing the displacement at the free end of the beam. Based on numerical simulations, the performance of the cantilever structure as a PMD compensator is assessed for 10-Gbits/s nonreturn-to-zero transmission systems with a large DGD. With this compensator, a significant improvement of system performance can be achieved in the eye pattern of a received signal.

We first compare different polarization mode dispersion (PMD) models for evaluating PMD impairments. We show that, compared with the all-order PMD model, most of the known first- and second-order PMD models overestimate the PMD induced distortions. Then we discuss various PMD mitigation techniques. We show that modulation formats have a big impact on choosing the most effective PMD compensation techniques. Not only do the benefits of using PMD compensators (PMDCs) depend on modulation formats, but the designing of a PMDC is largely affected by modulation formats as well. Some multi-channel PMD mitigation techniques for wavelength division multiplexing (WDM) systems are also discussed.

Wavelength converters for phase sensitive modulation formats based on semiconductor optical amplifiers are discussed. Advantages and limits are discussed based on an actual implementation.

We present some progress in the field of optical signal processing that could be utilized in all-optical packet switching. We demonstrate error-free 160 Gb/s optical wavelength conversion employing a single semiconductor optical amplifier. The gain recovery time of the semiconductor optical amplifier is greater than 90 ps. Assisted by an optical bandpass filter, an effective recovery time of 3 ps is achieved in the wavelength converter, which ensures 160 Gb/s operation. This optical wavelength converter can be controlled by a monolithically integrated optical flip-flop memory to route 80 Gb/s data-packets all-optically. The routing is realized without electronic control. The integrated optical flip-flop is based on two-coupled lasers, exhibits single-mode operation, has 35 dB contrast ratio between the states and switches state in about 2 ns. We demonstrate that the integrated flip-flop is able to control the optical wavelength converter up to 160 Gb/s. The system is capable of routing 80 Gb/s data packets with duration of 35 ns, separated by 15 ns of guard time.

Analysis of an all-optical 3R regenerator for DPSK signals is presented. The phase regeneration is performed in the amplitude domain by the use of a fiber-based 2R regenerator. Strong reduction of phase fluctuation is predicted.

We demonstrate a novel all-optical RZ decision gate by using one dark-optical-comb injected SOA which is controlled by a reshaped optical clock with extremely high cross-gain-modulation (XGM) depth and narrow gain window. Such a decision gate exhibits improved 3R regeneration performances including timing tolerance of 33.5 ps, Q factor of 8.1 and extinction ratio of 13.6 dB. The correlation between backward injected dark-comb and input data wavelength region for optimizing the extinction ratio of the decision gated RZ data is determined. Under a threshold operating dark-optical comb power of 7 dBm, such a decision gate can recover the degraded RZ data with a bit-error-rate of less than 10^-9 at 10 Gbits/s.

In order to increase the channel bit rate of commercial submarine systems from 10Gb/s to 40Gb/s in a cost effective way, new technologies have to be implemented. We will review them including Raman amplification, modulation formats and PMD mitigation techniques. The introduction of Wavelength Division Multiplexing (WDM) has triggered a tremendous capacity growth in submarine systems, both by the increase of the number of WDM channels and by the increase of the channel bit-rate. Starting from 2.5Gbit/s in the mid-nineties, the bit-rate was upgraded to 10Gbit/s by the end of the century in commercial prudcts. The next generation of submarine systems will likely be based on 40Gbit/s bit rate. However, transmissions at 40Gbit/s rate are more challenging than transmission sat 10Gbit/s. The goal of this paper is to provide an overview of the technologies which could be required or used in next-generation submarine systems. In the first part of this paper, an overview of the history of submarine links is provided. Then the technologies used in current Nx10Gbit/s systems are described. Eventually, the challenges to overcome are discussed, whether they concern the type of fiber, the type of optical amplifier, or the nature of the modulation format.

Along with the development of information society, the requirement of high speed, large capacity transport system is more and more important. 40 Gb/s based optical transport system could provide more high speed and more large capacity information transport. So it is development trend of optical transmission system. In this paper, the research and development progress of 40 Gb/s (STM-256) SDH optical transmission system and 80×40 Gb/s DWDM system in China have been introduced, including Key technologies of 40 Gb/s optical signal transport and some test results of 40 Gb/s SDH optical system and 80×40 Gb/s DWDM system have been showed.

We report the transmission of 40Gb/s channels over DWDM optically amplified communications systems employing RZ, NRZ, CS-RZ amplitude and differential phase shift keying modulation over 320 km standard SMF and 328 km dispersion compensating modules. Impacts of optical filtering of 10 Gb/s DWDM transmission system on these 40G modulation formats are reported. The mutual impacts of 10 Gb/s and 40 Gb/s co-transmission are proven to be minimum.

For the future broadband Fiber-To-The-Home (FTTH) services, it will be revealed to be a myth that the low bit-rate uplink may be deployed, while only the downlink has to be high bit-rate. Current FTTH system forces the customers a stressful access in the uplink due to its MAC based on TDMA under always-on service provisionings. Without an abundant bandwidth of uplink available, peer-to-peer applications such as exchanging gigabyte files of uncompressed 1.2 Gbps high-definition (HD) TV class or even 6Gbps super-high-definition (SHD)class digital movies as well as teleconferencing and bi-directional medical applications such as tele-diagnosis and -surgery won't become widewpread. With a narrowband uplink, even non peer-to-peer customers will be put in a disadvantageous position by being forced to share the limited bandwidth with a limited number of bandwidth-hungry users.

In this paper, an optical code division multiple access (OCDMA) system is demonstrated. The ultrashort light pulse is encoded and then decoded by amplitude sampled fiber Bragg gratings with equivalent phase shift (EPS). Compared with traditional superstructured fiber Bragg grating (SSFBG) with real phase shift (RPS), FBG with EPS is much easier to fabricate. In our experiment, it shows its full ability to perform encoding and decoding in OCDMA systems, and good encoding/decoding performance is achieved.

A theoretical analysis is presented to evaluate the impact of fiber chromatic dispersion on the bit error rate (BER) performance of a two dimensional wavelength hopping time spreading optical CDMA system. A bipolar-unipolar coding scheme has been considered for 7 chip m-sequence and 31 and 127 chip Gold sequence considering a standard single mode fiber. The numerical results show that there is an order of magnitude increase in BER due to the effect of dispersion. For 100 simultaneous users, at a received power of -4dbm, average BER occurs at about 10^-6 and 10^-12 with and without chromatic dispersion respectively for 7-chip m-sequence at a chip rate of 10 Gchip/sec for a fiber length of 245.05 km, with dispersion coefficient of 16 ps/km-nm. For 31-chip Gold sequence, the corresponding average BER is of the order of 10^-17 and 10^-21 with and without dispersion respectively. It is further noticed that for 10 simultaneous users, the system operating at a chip rate of 10Gchip/sec, suffers a power penalty of 9.15 dB for 7-chip m-sequence and 1dB for 31-chip Gold sequence at a BER of 10^-9.

Two novel multi-granularity optical cross-connect (MG-OXC) architectures for dense wavelength division multiplexing (DWDM) networks are proposed to realize transparent waveband switching, and hence to reduce the switching and transmission costs. The performances of the two architectures are studied and compared in detail. The functionality of the proposed interconnecting architecture is investigated for a 10-Gb/s WDM system.

For reasons of survivability, many methods have been introduced to Automatically Switched Optical Network (ASON), such as path protection. In path protection, the protection path is risk-disjoint with the working path, which can improve the survivability of network. The complete risk-disjoint routing algorithm (CRDRA) is the most popular algorithm in protection path selection at present, in which the links sharing same risk with those contained in the working path are pruned away from the available resources when calculating protection path. By using this algorithm, the survivability of network can be enhanced; but the number of connections that can be successfully routed over current network drops. In this paper, a partial risk-disjoint routing algorithm (PRDRA) based on shared risk link group (SRLG) is proposed. In this algorithm, the risk-sharing links are not pruned away but had their weight adjusted to a proper value that is larger than normal value but still comparable. Selecting protection path with PRDRA, improved survivability can be achieved while the number of connections that can be successfully routed over current network is kept from serious decline. Routing simulations have been done over mesh networks to compare the two different algorithms. With the simulation results, a conclusion can be made on the performance of different algorithms.

As bandwidth-intensive and time-sensitive streams applications such as high-definition television (HDTV) get popular, there rises a demand of supporting multicast communication directly at optical layer on next-generation optical networks. The multicast-capable optical cross-connect (MC-OXCs) is the necessary device to implement multicasting at optical layer. The functional building blocks for MC-OXC are light splitters (space splitters) or multi-wavelength converters (frequency splitters). The space splitter has no wavelength conversion capability, whereas the frequency splitter has limited fanout, decreased optical signal-to-noise ratio and also is expensive. We therefore proposed a tradeoff architecture called joint multicasting capable optical cross-connect (jMC-OXC) integrating both space splitters and multi-wavelength converters. In this paper, we investigate the network performance and the physical transmission performance on the jMC-OXC architecture by dynamic simulation and experimental demonstration. Three simulation schemes are designed and the results show that the network performance of the jMC-OXCs with limited multi-wavelength converters can obtain a close performance to that with full multi-wavelength converters. A prototype of jMC-OXC is examined and its bit error rate (BER) performance is tested after passing the multi-wavelength converters.

This paper describes the interworking between 4K digital cinema and a GMPLS/OXC network in JGN II. Through three trials in JGN II, we confirmed that 4K real-time streams were successfully transmitted in GMPLS paths and that the GMPLS/OXC technology can be used for transmission in such a broadband application.

A thorough study of an all-optical chaotic communication system, including experimental realization real-world testing and performance characterization through bit-error-rate analysis, is presented. Pseudorandom data that are effectively encrypted in the chaotic emitter and sent for transmission are recovered at the receiver with bit-error-rate (BER) values as low as 10-7 for 1 Gb/s data rate. Different data code lengths and bit-rates at the Gb/s region have been tested. Optical transmission using 100km fiber spools in an in-situ experiment and 120km in an installed optical network showed that chaotic communication systems does not act as a considerably deteriorating factor in the final performance.

The key of analyzing the light signal transmitting in fiber is the interaction and correlation between them. At present, we must consider the medium and the light field as one united system. Its strict physical theory is the quantum electrodynamics (QED), to process the light field and medium to second quantization. The quantum theory is considering both the medium and the light field as quantum wave function, so we can discuss the extreme limit frequency and distance. In the analyzing of extreme limit frequency, we can get it by considering two situations: the first is the photoelectric effect between light signal and fiber medium, the second is the transmission effect of high frequency incident light. In the analyzing of extreme limit of light signal distance, we will consider the interaction of the electron and the light field. We apply the theory and formula of scattering section. We analyze the colliding mechanism of particles in medium. We calculate the transmitting situation of photon energy according the energy and momentum reservation laws. According to the maximum frequency shift of light which is allowed by optical receiver, we can get the extreme limit distance of signal transmitting.

The authors have developed a multi-wavelength IP packet sender/receiver to be inserted into a PCI-bus slot of a PC or a workstation. When the workstations or PCs with the sender/receivers are connected to an internet access network configured on the PON as client terminals, a pee-to-peer connection-oriented communication path can be set up between them. The PCs can afford a real-time interactive communication with QoS fully guaranteed.

The number of FTTH subscribers in Japan has reached 5 million as of March, 2006. The dominant technology in Japan is GE-PON, which was standardized in IEEE in June, 2004. We present optical and system technologies in GE-PON and future PON technologies utilizing higher bit-rate and WDM technologies.

We discuss wishes for next generation access networks and proposed and demonstrated WDM-PONs based on wavelength locked F-P LDs. These WDM-PONs guarantee color-free (or wavelength independent) operation of ONTs and OLT transceivers. The scalability of the WDM-PON is also investigated in terms of transmission length, split ratio, and transmission speed.

We evaluate the effects of the downstream modulation formats (such as NRZ, inverse RZ, and Manchester) on the performance of the WDM PON implemented by using the reflective semiconductor optical amplifiers (RSOA's). The results show that, when we modulate the downstream signals in Manchester format, the performance of the upstream signal becomes insensitive to the operating condition of the RSOA.

Bidirectional HDTV/Gigabit Ethernet/CATV over hybrid dense-wavelength-division-multiplexing passive optical network (DWDM-PON) based on injection-locked vertical cavity surface emitting lasers (VCSELs) and distributed feedback (DFB) laser diodes as transmitters were proposed and demonstrated. Services with 129 HDTV channels, 1.25 Gb/s Gigabit Ethernet connection, and 77 CATV channels were successfully demonstrated over 40 km single-mode fiber (SMF) links. Good performance of bit error rate (BER)/carrier-to-noise ratio (CNR)/composite second order (CSO)/composite triple beat (CTB) were achieved in our proposed bidirectional DWDM-PON.

The concept of a nonlinear transfer function of a fibre-optic communication link is reviewed. Also an approximation of the nonlinear transfer function is introduced, which allows to define an equivalent single-span model of a dispersion-managed multi-span system. In this paper we will show its limits of validity and try to extent these limits by enhancing the theoretical model. In this respect we will discuss the impact of dispersion precompensation and show the influence of residual dispersion per span, number of spans and local dispersion on transmission systems with on-off keying and differential phase-shift keying modulation formats. This approach allows fast assessment of the performance of a given modulation format over various dispersion maps by reducing the need for extensive numerical simulations.

In this work we investigate the influence of adaptive equalization techniques on the link design of long haul optical transparent transmission (up to 800 or 1600 km). Receiver-side electronic and optical equalizers can be utilized to mitigate all kinds of distortions generated in the transmission link e.g. by the fiber, amplifiers or other network elements (OADMs, OXCs) making the systems robust and more tolerant against these effects. However, the performance of the realizable equalizers might not be sufficient to compensate for all dispersion of a link in optical transparent core and long haul networks which are typically in the range of up to 1600 km. Therefore we present a link design concept which takes a distributed static dispersion compensation as well as adaptive equalization at link termination points (e.g. OADMs or final receivers) into account. In our concept only a coarse compensation with DCF-x modules which compensate only for e.g. 80, 20 or 5 km standard single mode fiber is used and the residual dispersion will be handled by an adaptive equalizer. The system behavior is analyzed for specific target lengths of 800 and 1600 km taking chromatic dispersion, nonlinear effects and amplifier noise into account. It is shown how the deployment of receiver-side equalizers may be used to allow for a coarse DCF granularity while maintaining the same overall system performance. Several amplitude modulation formats like NRZ, RZ or duo-binary are investigated within this concept.

We propose a novel method of tunable compensation for dispersion and dispersion slope based on a uniform fiber Bragg gratings (FBGs). A specially designed mechanical rotator and a metal beam with a thickness gradient can induce a nonlinear strain gradient along the length of the gratings so as to control the dispersion of the gratings and their dispersion slope as well. By way of changing rotation angle, the dispersion value was controlled from -894.6ps/nm to -365.2ps/nm while the dispersion slope from 1049.2ps/nm² to 134.8 ps/nm². The centre wavelength was measured to shift by less than 0.07nm. In simulated 80Gb/s NRZ transmission system, we showed to compensate dispersion and dispersion slope with system power penalty of about 2.5 dB by using the proposed method.

A 40Gb/s dynamic dispersion compensation system employing the tunable dispersion compensation device and the spectral shift in the semiconductor optical amplifier is demonstrated. The compensation accuracy can reach 5ps/nm, and the range is 400ps/nm.

We demonstrate a 10.7Gb/s-line-rate L-band WDM loop transmission over 1890km standard single-mode fiber (SSMF) with 100km amplifier spacing as well as non-return-to-zero (NRZ) format. For the first time, dispersion compensating fiber (DCF) plus chirped fiber Bragg grating (CFBG) is employed for hybrid inline dispersion compensation. The power penalty of each channel is less than 3dB after three loop transmission. The experimental results show that high-performance-CFBGs can be successfully used in ultra-long haul (>1000km) WDM systems. We also point out that all-CFBG compensation scheme is not suitable for re-circulating loop transmissions.

Recent demonstrations have shown the benefits of using Raman amplification for transmitting high bit rates over long distances for both repeatered terrestrial and unrepeatered submarine applications. A multi-reach terrestrial field trial using Raman amplification was recently completed in conjunction with Verizon Business. The field trial verified the feasibility of upgrading an existing network to transmit a mixture of 10 Gb/s and 40 Gb/s over a record setting distance of 3040 km. In addition, an unrepeatered submarine demonstration was recently completed to show the versatility of all-Raman amplification and its ability to transmit a mixture of 40 Gb/s (OC-768) and 10 Gb/s signals over a long unrepeatered link. Together these demonstrations show the versatility of all-Raman amplification to support high-bandwidth requirements over increasingly longer distances. This paper reports on the record setting multi-reach field trial and provides the results of the recently completed demonstration to transmit a mixture of 40 Gb/s and 10 Gb/s signals over an unrepeatered distance of 400 km.

We develop an efficient design method for the bi-directionally pumped distributed Raman amplifier (DRA). Using this method, we have designed a bi-directionally pumped DRA having a small gain ripple (< 1.1 dB) over the bandwidth of 90 nm.

All-optical gain-control Raman fiber amplifier (RFA) based on ring cavity technique using a pair of circulators is investigated experimentally. A pair of 1465nm laser diodes is used as the pump source to the RFA which provides a gain peak at around 1565 nm. The analyses of gain and noise figure are carried out using lasing wavelengths of 1550 and 1565 nm. Gain and noise figure variations are 0.4 and 0.5 dB respectively. We observed strong influence of pump-signal relative intensity noise transfer that results in noise figure penalty. The analyses are also carried out on hybrid configuration of RFA and EDFA and the characteristics of the dynamic range of the configurations are being compared.

All-optical wavelength conversion of NRZ signal by SOA-MZI with bidirectional data injection is numerically studied. The numerical results showed good agreement with experimental results.

We give the expression of the input experimental pulse, numerically study formation and propagation of chirped soliton according to the nonlinear propagation theory by use of the split-step Fourier method. The numerical results are consistent with the experimental data. The pulse can evolve into chirped soliton after propagating over 3.5 dispersion length, the temporal FWHM and chirp gradually decreases with increase of input power. The temporal FWHM of chirped soliton executes a kind of damped oscillation with increase of propagation distance whose period and amplitude decrease with increase of input power, the temporal waveform always maintain near hyperbolic secant pulse during nonlinear propagation.

By using thin membrane plating technology in collimating lens, we successfully manufacture a 160Gbit/s time-division multiplexer basing on the combination of fiber and space structure method. The multiplexer is flexible according to your practical need, which can output 20, 40, 80, or 160Gbit/s signal. For 20, 40, 80 or 160Gbit/s multiplexer, insertion losses are 4dB, 6dB, 8dB and 16dB respectively. The low insertion loss provides maximum transmission power. It also has such features: polarization insensitivity, very short coherence length, high time-delay accuracy, and excellent temperature stability. An experiment of two degrees PMD compensation in 160Gbit/s RZ optical communication system is achieved. The PMD monitoring technique is based on DOP as error signal. A practical adaptive optimization algorithm was introduced in dynamic adaptive PMD compensation. The experimental results show the improvement in PMD. With this compensator, a significant improvement of system performance can be achieved by auto-correlative curves. The 2.5ps first-order and 15ps² second-order PMD are compensated. The PMD compensating time is less than 100ms.

We report on components and techniques for single wavelength channel transmission at data rates up to 2.56 Tbit/s. The enabling technologies were OTDM technology, phase modulation formats, and precise dispersion management of the fiber link.

Transmission of 40 Gb/s alternating phases 0 and π duo-binary modulation format with 67%, 50% and 33% pulse width (modulation duobinary carrier suppression-CS and carrier maximum pulse carving) is demonstrated. A minimum 136 ps/nm dispersion tolerance can be achieved at this transmission rate. At least 5 dB and 1.2 dB receiver sensitivity improvement are achieved for CSRZ 67% and RZ 50% duobinary modulation respectively as compared with the experimental measurement under CSRZ-DPSK transmission over 320 km 3-span optically amplified dispersion compensated standard single mode fiber system.

We propose two novel transmitter configurations for generation of linear and non-linear phase modulated optical minimum shift keying (MSK) signals based on approaches of continuous phase frequency shift keying (CPFSK) and I-Q component structure of offset/staggered DQPSK formats. The precoder for I-Q optical MSK structure is derived. The direct detection of optical ligtwave is utilized with implementation of the well-known differential non-coherent balanced receivers. Simulated spectral characteristics and dispersion tolerance to 40 Gb/s transmission are presented and compared to those of ASK and DPSK counterparts.

The generation and detection of linear, weakly non-linear and strongly non-linearly optical MSK long-haul transmission systems have been reported. The simulation test-beds based on Matlab Simulink platform have been successfully developed. The following significant results have been achieved: (i) BER of 10^-38 is achieved for an OSNR of 22.2dB; (ii) At BER of 10^-12, linear MSK is approximately 1 dB more tolerable to nonlinearity than weakly nonlinear MSK and 5.5dB more durable than strongly nonlinear MSK; (iii) (c).The sideband suppression ratio for linear MSK is approximately 22dB. New technique for calculation of performance evaluation implementing expected maximization (EM) theorem is utilized.

This paper describes high-speed integrated optical modulators for advanced modulation formats, such as differential quadrature-phase-shift-keying (DQPSK), continuous-phase frequency-shift-keying (CP-FSK), etc. 80Gbit/s optical DQPSK modulation was demonstrated by using the modulator consisting of four phase modulators.

This paper proposes a method of wavelet analysis for de-noising at receiver system in WDM laser inter-satellites communication. Background noises such as galactic noise, sunlight and etc make the received power reduce. The noisy signal is decomposed using wavelets and wavelet packets; then is transformed into wavelet coefficients and the lower order coefficients are removed by applying a soft threshold. De-noised signal is obtained by reconstructing with the remaining coefficients. In this paper, we evaluate different wavelet analysis for de-noising at receiver system in inter-satellites laser communication. Simulation results indicate that if the wavelet de-noising method is used with different wavelet analyzing functions, it will improves the signal to noise ratio (SNR) about 2 dB when the signal frequency is 1.5 GHz.

In this paper, differential pulse-position modulation (DPPM) over an indoor optical wireless link is represented as a Markov source transmitted over a finite-state machine channel. Then, the MAP, Max-Log-MAP and SOVA algorithms are employed for soft-decision decoding of DPPM. The error performance with intersymbol interference is evaluated and compared to hard-decision decoding and MLSD. It is shown that the performance of Max-Log-MAP and SOVA is closed to that of MLSD and MAP with lower complexity.

Fully coherent laser beams are very sensitive to the turbulence medium. Studies have shown that using partially coherent beams is an effective method to reduce this unsatisfactory feature. Unfortunately partially coherent beams are more expanded when they propagate thought turbulence medium. In this paper we have determined a special degree of global coherency, which optimizes beam propagation with respect to limiting factors (sensitivity to the turbulence and broadening)

By injecting the optical NRZ data into a Fabry-Perot laser diode (FPLD) synchronously modulated at below threshold condition or a semiconductor optical amplifier (SOA) gain-depleted with a backward injected clock stream, the all-optical non-return to zero (NRZ) to return-to-zero (RZ) format conversion of a STM-64 date-stream for synchronous digital hierarchy (SDH) or an OC-192 data stream for synchronous optical network (SONET) in high-speed fiber-optic communication link can be performed. Without the assistance of any complicated RF electronic circuitry, the output RZ data-stream at bit rate of up to 10 Gbit/s is successfully transformed in the optically NRZ injection-locked FPLD, in which the incoming NRZ data induces gain-switching of the FPLD without DC driving current or at below threshold condition. A power penalty of 1.2 dB is measured after NRZ-to-RZ transformation in the FPLD. Alternatively, the all-optical 10Gbits/s NRZ-to-RZ format conversion can also be demonstrated in a semiconductor optical amplifier under a backward dark-optical-comb injection with its duty-cycle 70%, which is obtained by reshaping from the received data clock at 10 GHz. The incoming optical NRZ data-stream is transformed into a pulsed RZ data-stream with its duty-cycle, rms timing jitter, and conversion gain of 15%, 4ps, and 3dB, respectively. In contrast to the FPLD, the SOA based NRZ-to-RZ converter exhibits an enhanced extinction ratio from 7 to 13 dB, and BER of 10^-13 at -18.5 dBm. In particular, the power penalty of the received RZ data-stream has greatly improved by 5 dB as compared to that obtained from FPLD.

We present semiconductor optical amplifier-based optical signal processing for high bit-rate optical communication. All optical wavelength conversion and logic operation of 40-Gb/s PSK and OOK signals are demonstrated. High-speed all-optical signal characterization methods and the application to all-optical signal processing are also presented.

We propose and demonstrate a novel four-wave mixing (FWM) compensator made of highly-nonlinear fiber and a pump laser. We experimentally confirm that the proposed technique can suppress the FWM components generated in the transmission fiber by as much as 13.3 dB.

This paper focuses on the experimental demonstration of all-optical wavelength conversion at 10 Gbit/s based on cross-bias modulation (XBM) in electroabsorption modulators (EAMs). It is shown that the input 1552 nm pump signal can be converted to 1562.2 nm at 10 Gbit/s by XBM. The optical spectra of both the pump and the probe signals after wavelength conversion is given and the eye diagram of the converted signal is demonstrated. The advantages of the proposed wavelength conversion scheme are also discussed.

This paper presents a study on an all-optical multicasting switching matrix design based on SOA and AWG technology for all optical DWDM network. The technique not only has the capability to control the degree multicasting optically but also perform double stage of wavelength conversion. Experimental results have shown the design's capability to multicast an incoming 10Gbps optical signal onto 16 outgoing signals using Cross Gain Modulation and then using Cross phase modulation as a second stage of conversion to perform 2R (re-amplified and re-shape).

We have studied an all-optical wavelength converter utilizing cross-polarization modulation in a semiconductor optical amplifier, and presented a feedback scheme for polarization based on the monitoring the light intensity of converted signal and adjusting the voltages of the polarization controller, which realized an error-free conversion of 10Gbit/s NRZ signal for more than 2 hours.

We proposed a novel microwave-band radio-over-fiber system with wavelength-division multiplexing (WDM) bus architecture. An electrooptic phase modulator combined with Fiber Bragg Gratings (FBGs) can realize frequency multiplying. Here the FBGs are used to fulfill three main tasks: convert phase modulation into intensity modulation, filter away unwanted optical frequency components, and work as demultiplexers in the WDM radio-over-fiber system. Simulation demonstrates a system capable of multiplying the frequency of differential phase-shift keying (DPSK) modulated signals from 2.7GHz to 5.4GHz (two times), 10.8GHz (four times) and even higher frequency values after transmitting the signals over 0.5 km of polymer optical fiber. Such a RoF system is cost-effective and suitable to install in in-house environment.

A radio-on-dense-wavelength-division-multiplexing (DWDM) transport system for personal handy system (PHS)/vehicle information and communication system (VICS)/electronic toll collection (ETC)/satellite broadcasting (SB) applications was proposed and demonstrated. Improved performance of bit error rate (BER) over a-100 km of large effective area fiber (LEAF) was obtained. Signal qualities meet the PHS demand with a data signal of 1.9GHz/10Mbps, VICS demand with a data signal of 2.5GHz/12Mbps, ETC demand with a data signal of 5.8GHz/20Mbps, and SB demand with a data signal of 12GHz/25Mbps.

A radio-on-DWDM transport system based on injection-locked Fabry-Perot laser diodes (FP LDs) and large effective area fiber (LEAF) transmission was proposed and demonstrated. Improved performance of bit error rate (BER) over a-50 km of LEAF was obtained. Signal qualities meet the personal handy system (PHS) demand with a data signal of 1.9GHz/10Mbps, vehicle information and communication system (VICS) demand with a data signal of 2.5GHz/12Mbps, electronic toll collection (ETC) demand with a data signal of 5.8GHz/20Mbps, and satellite broadcasting (SB) demand with a data signal of 12GHz/25Mbps.

Using four wavelength mixing (FWM) effect in high-nonlinear dispersion-shifted fibre (HNL-DSF), a novel up-conversion technique for binary on-off-key (OOK) radio-on-fibre (ROF) system is proposed in this paper. By injecting a non return-to-zero (NRZ) format data signal and a carrier-suppressed return-to-zero (CSRZ) format carrier signal into HLN-DSF, a signal with a new frequency component generated by FWM effect will have the phase characteristic of CSRZ carrier. The simulation results indicate that the signal has the better transmission performance because of the strong ability of resisting dispersion.

The simultaneous generation of multi-wavelength optical single sideband (SSB) modulation for wavelength division multiplex (WDM) Radio-over-Fiber systems using a differential group delay (DGD) element and a polarizer is constructed in the paper. Simulation proved the principle shows the capacity of such filtering scheme works for at least 8 channels. Experiment results accorded with simulations. Experiment also shows the operating range of the system is more than 40nm, which covers the whole C-band. The sideband suppression ratio for all possible channels is greater than 20dB.

A novel photonic microwave notch filter with both negative and positive coefficients is proposed and demonstrated using a single optical source as a probe beam. Due to the change in birefringence induced by another modulated pump beam, both non-inverted and inverted signals are generated in a highly nonlinear fiber. These are separated by using a polarization beam splitter and recombined with optical time delay given to one of them. From the generation of both coefficients and the orthogonal property between them, no resonance peaks were observed at baseband and the optical coherence could be avoided independently of the source coherence length. The experimental results show a stable transfer response with the free spectral range of 65 MHz and 1.32 GHz when the optical path difference is given to 3.19 m and 0.19 m, respectively.

A novel optical beamforming network scheme based on broadband optical source and Chirped Fiber Grating as true-time-delay module is proposed in the paper. The principle of this scheme is illustrated by theoretical analysis and the feasibility is validated by experiment. In experiment, Er-Doped Fiber Amplifier (EDFA) is used as broad band optical source, chirped fiber grating as dispersive element is used for time delay unit, and tunable optical filter is used to choose operation wavelength. In 9.25~10.25 GHz microwave band, optical true time delay performance is measured and results shows that the delay time consistency and the ratio of the RF signal to its noise (SNR) meet the practical demands, and such system has many potentials in optical complexity, cost, tune speed and the system expansibility.

The average queue delay of optical packet increases when the packets block in the virtual output queue header in the asynchronous optical packet switching. We analyze the character of variable length Internet packet and propose the preemptive short packets priority (PSPP) algorithm for reducing the total packet waiting time in the queue. In the PSPP algorithm, the short packet can preempt the transmission time of the long packet and can be served first. The analysis and the simulation shows the PSPP algorithm can make the average waiting time of the short packet decrease to zero almost and can decrease the total packet average waiting time largely when the traffic load is middle and low. The PSPP can guarantee the low average waiting delay for the real time TCP traffic implementation.

By taking advantage of statistical multiplexing in the burst level, Optical Burst Switching (OBS) technology enables optical internet to handle huge volume of data in an efficient manner without requiring optical memories in the optical domain. However, when the congestion is built in the optical network core, large amount of data can be lost. In this paper, we propose an efficient burst control algorithm, which operates based on the awareness on future traffic condition. Through the performance analysis, it is verified that the proposed algorithm effectively decreases the burst loss while maintaining high throughput as compared to the burst control algorithm based on current information.

In order to provide quality-of-service (QoS) in OBS, in this paper, a novel QoS classification based assembly algorithm for OBS is introduced. By employing assembly time and offset time based QoS classifier at edge node it supports QoS with more delicate time intervals and higher transmit rate. It also contributes to decrease the consisting congestion in the network. The design is implemented using Network Processor (NP) and evaluation results show the improvement for supporting multilevel QoS services.

Optical Burst Switching (OBS) is a switching paradigm to efficiently support the ever-growing broadband multimedia traffic over optical networks which has been proposed with the consideration that future optical networks should provide efficient support for bursty traffic. Although research in OBS networks has evolved from theoretical investigations to proof-of-concept demonstrations, several key issues need to be investigated further before OBS prototypes can clearly outperformed Optical Circuit Switched (OCS) networks. In OBS networks, data loss may occur when bursts contend for network resources. There have been several proposed solutions to resolve contention. Most contention resolution techniques only react to the contention, but do not share the contention information. Based on the presented offset time determination technologies, we propose a novel dynamic offset time mechanism which concerns how to adjust offset time at source. In this mechanism, the intermediate node will send a signal to the source to notice the contention. Source will adjust the offset time in different strategies according to the situation of the network. Numerical results demonstrate that the proposed mechanism effectively reduces the contention at the expense of the increase the packet average end-to-end delay.

In this paper, an architecture of Optical Burst Switched WDM ring network is studied. In the studied OBS ring network, every node is equipped with one tunable transmitter and one fixed-tuned receiver (TTFR) operating on a given wavelength that identifies the node. TTFR type ring network has an advantage that no receiver collisions occur and each node can detect channel collision by transmitter side. By computer simulations, we evaluate the performance of throughput, End-to-End delay and queue size. As a result, we show the performance effectiveness of our proposed access protocol.

We suggest a new protocol for Optical-Burst-Switched (OBS) ring network to support synchronous services (SONET/SDH, named TDM services) and asynchronous services (data services) expediently. In the new protocol, we use different bandwidth assignment methods for the two kinds of services. For synchronous services, we use fixed bandwidth assignment to broadcast the TDM slots to all remained nodes in optical domain; for asynchronous services, we use dynamic bandwidth assignment by OBS technologies. The data channel is frame based with 125μs frame period. In each 125μs frame, there are several reserved time slots for TDM services, and the remained time block can be used to transport data services. The main features of this protocol: 1) the bandwidth ratio of TDM and data services can be adjusted flexibly basing on demand; 2) the TDM services can be transported with guaranteed QoS; 3) the data services can be transported in data burst with statistic multiplexing profit; 4) we can transport TDM and data services in same data channel with OBS technology; 5) the total bandwidth utilization is about 70%.

The timing alignment between the pulse carver and data modulation signals in advanced modulation formats is monitored with two-photon-absorption processes in Si-APD. For 20 Gb/s signals, the monitor is sensitive to within ± 10 ps.

We report -31 dBm receiver sensitivity for heterodyne detection of 10 Gb/s OOK without using an optical pre-amplifier. These are the highest receiver sensitivities for unpreamplified heterodyne 10 Gb/s detection. We also show the development of a coherent heterodyne balanced fiber optic receiver. The receiver incorporates a DFB or a solid state laser local oscillator, balanced PIN photodiodes, RF post amplifier, automatic frequency control (AFC), phase locked loop (PLL), polarization control, and precision power supplies in a small instrument case. We will show shot noise limited detection of amplitude modulated signals, cancellation of laser RIN noise, performance improvement using balanced detection at 2.5 and 10 Gb/s, and IF linewidth reduction.

A new architecture of photonic link with balanced receivers is introduced. The newly proposed architecture can reduce both the influence of the relative intensity noise (RIN) of LD and the even order distortion of the optical modulator. Results show that 21dB RIN suppression is obtained.

An integrated VCSEL transmitter has been realized and designed in TSMC 0.18 μm CMOS process. The transmitter is compromised of a gain stage utilizing active feedback, active inductor, slew rate boosting techniques to achieve high gain, high bandwidth and an unbalanced degenerated driver stage to achieve data speed of 10Gb/s while ensuring the large peak to peak current into VCSEL. Simulated maximum bandwidth and power consumption of transmitter are 7.5 GHz and 40 mA, respectively.

We fabricated the standing metallic nano-stripe array structure on the thin ITO layer with a very narrow width and a relative high height of the stripe by using the electron-beam lithography and Ar ion milling process. The polarization properties were investigated as the period and the incident angle are varied. In transmission spectra of the TE polarization, the resonance with Fano-like resonance was observed. In case of TM polarization, there's no resonance near normal incidence because the electric field is perpendicular to the very thin metallic grating.

This paper presents a low-complexity parallel Reed-Solomon (RS) (255,239) decoder architecture using a novel pipelined variable stages recursive Modified Euclidean (ME) algorithm for optical communication. The pipelined four-parallel syndrome generator is proposed. The time multiplexing and resource sharing schemes are used in the novel recursive ME algorithm to reduce the logic gate count. The new key equation solver can be shared by two decoder macro. A new Chien search cell which doesn't need initialization is proposed in the paper. The proposed decoder can be used for 2.5Gb/s data rates device. The decoder is implemented in Altera' Stratixll device. The resource utilization is reduced about 40% comparing to the conventional method.

In this paper, we present the model of switching matrix based on the XPM in SOA for optical packet-switched networks, which is used to evaluate its performance. Cross-talk-induced bit error rate (BER), packet error rate (PER) and cascadability under unbalanced traffic are examined, which proves to be relevant to the properties of signal pulse such as initial chirp, shape and bit-rate. Several rules are drawn to facilitate the design of SOA-based switching matrix under unbalanced traffic.

We report the BER performance characteristics of 16-ary Multi-level Amplitude-Differential Phase Shift Keying (MADPSK) modulation formats. The modulation scheme is bandwidth efficient with an effective transmission symbol rate equal to 1/4 of the bit rate. The BER performance of NRZ and RZ MADPSK formats are investigated for long haul optically amplified transmission in the presence of optical noise, fiber residual chromatic dispersion and non-linearity.

A novel scheme for wavelength monitoring of DWDM systems was proposed in which the two transmission spectra of a polarization interference filter (PIF) are employed. The filter consists of a polarizer, a polarizing beamsplitter (PBS) and a birefringent plate sandwiched between them. The birefringent plate used is made of yttrium orthovanadate (YVO₄) crystal. The two transmission spectra with perpendicular polarization directions are complementary in light transmittivity. By designing the filter and controlling its temperature appropriately, the intersections of the two transmission spectra can be utilized to monitor operating wavelengths in a DWDM network. This scheme doesn't need reference light for wavelength sensing. In addition, the monitoring sensitivity is twice as large as that in the way that only one transmission spectrum is utilized.

A novel scheme of optical packet power equalization is proposed by utilizing the gain characteristic of FP-SOA sensitive to input optical power. The steep gain descending curve with the carrier density provides that the high-power input optical signal obtains the low gain and the low-power one achieves the high gain. Real-time controlling mechanism according to input optical power is aborted in this scheme. The working principle of optical power equalizer based on FP-SOA is introduced in detail. The simulations show that 10 dB pulse peak power variation can be clamped in less than 1 dB. The impact of the injecting current, pulse width and pulse period are investigated and discussed.

In this paper, a new architecture of the networks based on the proposed optical CDMA based on PON (OCDMA-PON) is presented, the design of the Optical Line Terminator (OLT) and the Optical Network Unit (ONU) for the OCDMA-PON are studied. The performance of networks based on the OCDMA-PON is analyzed in detail. Its bit error rate (BER) performance, networks scalability and transmission performance of such a system are studied under consideration of the multiple user interference (MUI) and the difference noise contributions. As shown in the study, the number of ONU/ONT, the channel link length, the number of available codes and the optical component characteristics deeply impact both the scalability and the transmission performance of such a system. Simulation results indicate that the proposed scheme is feasible and that the novel design can improve the scalability and transmission performance of the optical access networks.

Optical Burst Switching (OBS) is a promising solution for the future Internet and has attracted many attentions. In OBS network a key issue is channel scheduling. The objective of existing channel scheduling schemes is to maximize the number of bursts to be transmitted. Unfortunately it does not mean the maximum channel utilization. To maximize channel utilization we introduce an optimization scheduling scheme, in which the problem of channel scheduling is mapped to a problem of finding maximum weight independent set (MWIS). That is, the channel scheduling is equivalent to find a set of no-overlapping bursts such that the sum of bursts sizes is maximal amongst all sets of no-overlapping bursts. Our proposed optimization scheduling scheme can be solved as an integer programming problem in polynomial time, and is analyzed by a G^X/G^Y/1 queue system. Simulation results show outperforming in terms of channel utilization and bit loss probability compared with existing schemes. Discrimination to smaller burst and the alleviation with adjusted weight are discussed. The improvement with adjusted weight is showed as well.

In this paper, we propose Valiant Load-Balanced robust routing scheme for WDM mesh networks under the model of polyhedral uncertainty (i.e., hose model), and the proposed routing scheme is implemented with traffic grooming approach. Our Objective is to maximize the hose model throughput. A mathematic formulation of Valiant Load-Balanced robust routing is presented and three fast heuristic algorithms are also proposed. When implementing Valiant Load-Balanced robust routing scheme to WDM mesh networks, a novel traffic-grooming algorithm called MHF (minimizing hop first) is proposed. We compare the three heuristic algorithms with the VPN tree under the hose model. Finally we demonstrate in the simulation results that MHF with Valiant Load-Balanced robust routing scheme outperforms the traditional traffic-grooming algorithm in terms of the throughput for the uniform/non-uniform traffic matrix under the hose model.

We design a new Dynamic Power Equalization system by using OCM and DCE technology. The experiment results show that the new system has faster equalization speed and higher equalization precision than existing equipment. The equalization time is less than 200ms, and the precision is higher than 0.5dB.

A CWDM multi-service transmission system used in metro-networks is proposed and demonstrated. In the system, two wavelengths, with channel spacing of about 5 nm and 6 nm in old band and C band, respectively, are multiplexed by using fiber tapered couplers, instead of using thin film filter couplers. The two multiplexed bands are doubly multiplexed in a fiber. The system is designed to transmit different kinds of signals including E1, IP, E3/T3 and etc in metro networks. So the system can be used to transmit four different kinds of network service in one fiber link. The designs of the whole system and the related subsystems are described in details, and the corresponding tested results are also given.

A novel label abstraction and erasion scheme based on a FP-SOA is proposed for all-optical separation of the bit-serial label from payload and its performance is investigated by simulation. Fabry-Perot semiconductor optical amplifier (FP-SOA) is attractive not only due to simple fabrication without anti-reflection coating step but also due to its signal amplification capability and strong nonlinearity to perform all-optical label processing. Important features of this scheme are that it does not make use of any high-speed electronics and that only one device is needed. Using this scheme, label abstraction and erasion can be realized with the extinction ratio of 9.72 dB and 7.05dB, respectively. And the influence of pulse width ratio between the label and payload pulse on the extinction ratio is also investigated. We find that for a given case, there must be an optimized pulse width ratio to make the extinction ratio largest. It depends on working points of the label pulses and payload pulses in the gain curve.

In OBS, JET (Just-Enough-Time) is the classical wavelength reservation scheme. But there is a phenomenon that the burst priority decreasing hop-by-hop in multi-hop networks that will waste the bandwidth that was used in the upstream. Based on the HPI (Hop-by-hop Priority Increasing) proposed in the former research, this paper will do an unprecedented simulation in 4×4 meshed topology, which is closer to the real network environment with the help of a NS2-based OBSN simulation platform constructed by ourselves. By contrasting, the drop probability and throughput on one of the longest end-to-end path lengths in the whole networks, it shows that the HPI scheme can improve the utilance of bandwidth better.

Pre-configuration Cycle (p-cycle) is a promising approach for protecting working capacities in wavelength-division-multiplexing (WDM) mesh networks because of its ability to achieve ring-like recovery speed while maintaining the capacity efficiency of a mesh-restorable network. The p-cycle design in WDM mesh networks is to determine a set of p-cycles requiring near-minimal spare capacity to fully protect a capacitated network with individually different working capacity quantities on each span. The spare capacity allocation for the p-cycles is to decide where to place spare capacity in the network and how much spare capacity must be reserved on the p-cycles to guarantee seamless communications services survivable to a set of failure scenarios. The p-cycles are formed in the spare capacity of the network, so a spare capacity allocation of the p-cycles can be done without affecting the working traffic. This paper investigates the spare capacity allocation problem for p-cycles in WDM mesh networks. We propose a heuristic method, called the Joint p-Cycle capacity allocation (JCCA), for p-cycles spare capacity allocation in WDM mesh networks. This method can allocate optimal spare capacity for p-cycles and ensure 100% restorability without using Integer Linear Programming (ILP). JCCA configures the p-cycles with considering the distribution of the capacity distribution of the networks and first assigns p-cycles with good actual efficiency. The performance of JCCA is evaluated by computer simulations on the real world network topology. And the comparisons of this method and traditional design algorithms are presented in the paper.

A novel label erasion scheme based on a FP-SOA is proposed for all-optical separation of the subcarrier multiplexing (SCM) label from payload and its performance is investigated by simulation. Fabry-Perot semiconductor optical amplifier (FP-SOA) is attractive not only due to simple fabrication without anti-reflection coating step but also due to its signal amplification capability and strong nonlinearity to perform all-optical label processing. Important features of this scheme are that it does not make use of any high-speed electronics. Using FP-SOA as optical SCM label eraser, more than 9db extinction ratio can be obtained. Meantime the payload can be amplified by more than 11dB. In this scheme, FP-SOA acts as a low-pass filter as while as an amplifier.

The dispersion of 8×10Gb/s wavelength division multiplex (WDM) system has been compensated by the cascaded chirped fiber Bragg gratings(CFBGs), with ITU-T standard wavelengths and wavelength grid. The ASE of the EDFA could be reduced, the OSNR of the transmitted signal could be increased and the fluctuation of the EDFA gain could be controlled in the certain scope by the dispersion compensated CFBGs' WDM system. Impact of cascaded CFBGs' delay ripple on dispersion compensation has been analyzed. Experiment of error-free 8×10Gb/s 2015km transmission without forward error correction (FEC) and electronic repeaters were demonstrated. In the transmission, simplex CFBGs compensators were used and no other form of dispersion compensators were adopted. The experiment result showed that the consistency of the dispersion compensating in each channel is perfect over 2015km optical fiber transmission. The experiment result does agree with the theoretic analysis.

This paper describes the transmission performance of the 10Gb/s phase-modulated (PM) duobinary modulation format, compared with the conventional duobinary modulation format using a low-pass filter (LPF) in cascaded optical networks. We theoretically investigate the transmission characteristics of both modulation formats with considering the effect of narrowing bandwidth of multiplexers (MUX) and demultiplexers (DEMUX) at cascaded nodes with different fiber launching powers (FLP). It's shown that PM duobinary signals, at the FLP of 9dBm, can be transmitted over 3 nodes (up to a distance of 1,200km) with only 1.25dB power penalty.

Multicast at WDM layer by light splitting is considered as an efficient way to support increasing multicast traffic, but the enormous power losses brought by light splitting must be seriously taken into account. In this paper, we study the scheme of light-tree power recovery under optical layer power budget constraints. A modified slitter-and-delivery (MSaD) switch with the capability of power recovery is proposed. To minimize the costs and delay brought by power recovery, we develop an algorithm called optimal power recovery nodes allocation (OPRPA). Simulation results demonstrate that the power recovery scheme can guarantee power budget constraints and make the costs and delay acceptable.

We adopt the worst channel equalization (WCE) scheme to compensate polarization mode dispersion (PMD) in wavelength-division-multiplexing (WDM) systems. The degree of polarization (DOP) of single state of polarization (SOP) optical signal is used as the feedback signal of the PMD compensators, and the particle swarm optimization (PSO) algorithm is used as the searching algorithm. Simulation results show that the DOP of RZ format of the worst-performance channels after compensation is greatly improved, and the PSO algorithm is successfully experienced into adaptive multi-stage PMD compensators in a 40-Gb/s optical WDM communication system. The WCE scheme is an effective way for PMD compensation in WDM system.

In an optical packet switch, wavelength allocation algorithm has to be considered when wavelength resources are shared among all optical packets. This paper addresses the wavelength allocation algorithms for optical packet switch with limited wavelength converters and output WDM optical buffer. Under the condition of limited wavelength conversion, two wavelength control strategies, i.e., greedy mode and conservative mode, are proposed. According to these two wavelength control strategies, four wavelength allocation algorithms are presented. Performances of these algorithms are compared in detail by simulation experiments.

Optical buffers and wavelength converters are two effective methods to resolve optical packet contentions in an optical packet switch. This paper focuses on the performance of fixed wavelength converters in resolving optical packet contentions. An optical packet switch with fixed wavelength converters and WDM optical buffers is proposed. The fixed wavelength converters and the WDM optical buffers are all configured as recirculation form. The performances of the optical packet switch are evaluated by simulation experiments.

A novel header recognition scheme using variable keyword generator and optical logic XOR gate with loop is proposed. The proposed optical header recognition is realized all-optically, and is expected to operate in over 40 Gbps. The simulation results show its feasibility to recognize bit patterns of header at 10 Gbit/s and beyond.

We have proposed a new model of the rate equation for laser diode performance analysis. A modulated signal of LD is analyzed by the rate equation using the finite difference method (FDM) in time domain. For the analysis of modulated optical signal, an injection current term in the rate equation is altered to an appropriated electrical function. L-I curve, intensity modulation frequency response, and nonlinearity of laser diodes are analyzed by the proposed rate equation model. The proposed model can be applied to determine the operating point of LD and to design the analog optical transmitter.

The formulas for calculating nonlinear phase noise are proposed for both pre-compensation and post-compensation schemes. Based on these formulas, the phase noise, power tolerance and optimal signal peak power of both dispersion compensation schemes are analyzed and discussed in detail, respectively. The result shows that pre-compensation is more effective in reducing the nonlinear phase noise when compared with post-compensation. Its suppression ability improves with signal energy, ASE power spectral density and transmission distance increasing. The pre-compensation system possesses higher power tolerance than post-compensation system and the optimal signal power is increased when dispersion is taken into account, which results that the optimal phase shift is larger than 1rad. And the optimal signal power for pre-compensation system is larger than post-compensation system.

We proposed a novel linearization technique for DFB LD in Radio-over-Fiber. The proposed scheme is using light injection based on cross-gain modulation (XGM) effect. We experimentally demonstrated and evaluated the enhanced CNR performance using the proposed scheme.

A feedforward over an analog optical link is known as difficult technique in spite of its superior linearization. We proposed a broadband analog feedforward optical transmitter using a wideband 180° hybrid coupler instead of conventional frequency-sensitive phase shifter with the narrow bandwidth property. Using the wideband 180° hybrid coupler, the wideband linearization technique enhances the wide operation frequency and simple adjustment of the feedforward optical transmitter. Our experimental results show the 3^rd-IMDs enhancement more than 10 dB at 385 MHz range (1.375~1.76 GHz).

A novel waveband switching node architecture is proposed using tunable optic filters for arbitrary waveband add/drop. The simulation results show this new architecture achieves lower blocking probability and reduce the scale of ports at switch fabric.

We present an incoherent optical spectral CDMA (OS-CDMA) multiple-access system that uses in-fibre Bragg gratings for encoding/decoding. The system comprising four channels, one desired and operating at a 2.5Gbps and 10Gbps. Our measurements verify that this technique suffers from severe chromatic dispersion. Nevertheless with appropriate dispersion compensation, the system is able to achieve a span of 100km.

For future high speed indoor wireless communication, diffuse wireless optical communications offer more robust optical links against shadowing than line-of-sight links. However, their performance may be degraded by multipath dispersion resulting from surface reflections. We have developed a multipath diffusive propagation model capable of providing channel impulse responses data. It is aimed to design and simulate any multi-beam transmitter under a variety of indoor environments. In this paper, a multi-beam transmitter system with semi-sphere structure is proposed to combat the diverse effects of multipath distortion albeit, at the cost of increased laser power and cost. Simulation results of multiple impulse responses showed that this type of multi-beam transmitter can significantly improve the performance of BER suitable for high bit rate application. We present the performance and simulation results for both line-of-sight and diffuse link configurations.

The study of an optical wireless communication link using illumination light-emitting diodes (LEDs) has been performed. The eye patterns and bit error ratio of LED transmitter are measured and compared according to the modulation frequency and the distance between optical source and receiver. From the experimental demonstration, it is shown that the optical wireless communication link using illumination LEDs can perform 20 Mb/s data transmission.

Two kinds of electronic compensation of polarization mode dispersion (PMD) are demonstrated in this paper, including transversal filter(TF) and feed forward equalizer(FFE) + decision feedback equalization(DFE). Compared with the normalized least-mean-square (NLMS), the least-mean-square (LMS) is the best algorithm to control the equalizers. The tap weights are adapted using the LMS algorithm until they reach stationary values. Then we focus on the structure of FFE+DFE and find out the relationship of the number of taps and BER with different structures in the same conditions. From the comparison, the work of 4-tap FFE+2-tap DFE is more effective, which has the most commercial value.

The paper considers the problem of establishing robust routes for multi-granularity connection requests in traffic-grooming WDM mesh networks and proposes a novel Valiant Load-Balanced robust routing scheme for the hose uncertain model. Our objective is to minimize the total network cost when assuring robust routing for all possible multi-granularity connection requests under the hose model. Since the optimization problem is recently shown to be NP-hard, two heuristic algorithms are proposed and compared. When implementing Valiant Load-Balanced robust routing scheme to WDM mesh networks, a novel traffic-grooming algorithm called MHF (minimal hop first) is proposed. We evaluate MHF by Valiant Load-Balanced robust routing with the traditional traffic-grooming algorithm by computer simulation.

Our amplifier using an all optical method and a fixed GFF achieved automatic gain flatness through all C-band without any NF degradation, and simultaneously a constant 25 dB gain, while input signals were varied between one channel and forty WDM channels.

Polarization Division Multiplexing (PDM) transmission system is simulated. Two types of PDM system are simulated in order to find important variables for transmission. Both 10 Gbit/s and 40 Gbit/s case are examined for comparison. Signal modulations are NRZ and RZ. PDM systems are sensitive for transmission length and polarization-related loss, especially Polarization Mode Dispersion (PMD). RZ signal has more degradation than NRZ.

A novel all-optical flip-flop is proposed and demonstrated experimentally. The flip-flop consists of a conventional multi-mode Fabry-Perot laser diode (FP-LD) and a specially designed master FP-LD which has a built-in external cavity and operates in single longitudinal mode oscillation. The triggering pulses were generated by external modulators with non-return-to-zero pulse pattern generators. The rising and falling times of the output signal in on-off operation of the flip-flop were about 50 ps. The required powers of both set and reset pulses were less than -9 dBm.

This paper describes the simulation approach of the improvement of dispersion tolerance for an electrical-binary-signalbased duobinary transmitter. We can transmit the duobinary signals over 200 km of singe-mode fiber by optimizing the relative time delay and the driving voltage.