In this tutorial overview, we survey some of the challenging problems facing Optical Transport and their solutions using new semiconductor-based technologies. Advances in 0.13um CMOS, SiGe/HBT and InP/HBT IC process technologies and mixed-signal design strategies are the fundamental breakthroughs that have made these solutions possible. In combination with innovative packaging and transponder/transceiver architectures IC approaches have clearly demonstrated enhanced optical link budgets with simultaneously lower (perhaps the lowest to date) cost and manufacturability tradeoffs. This paper will describe: *Electronic Dispersion Compensation broadly viewed as the overcoming of dispersion based limits to OC-192 links and extending link budgets, *Error Control/Coding also known as Forward Error Correction (FEC), *Adaptive Receivers for signal quality monitoring for real-time estimation of Q/OSNR, eye-pattern, signal BER and related temporal statistics (such as jitter). We will discuss the theoretical underpinnings of these receiver and transmitter architectures, provide examples of system performance and conclude with general market trends. These Physical layer IC solutions represent a fundamental new toolbox of options for equipment designers in addressing systems level problems. With unmatched cost and yield/performance tradeoffs, it is expected that IC approaches will provide significant flexibility in turn, for carriers and service providers who must ultimately manage the network and assure acceptable quality of service under stringent cost constraints.

Presently, formerly independent autonomous optical links are growing increasingly into meshed networks making measures for performance monitoring and automated error analysis a necessity. At the same time adaptive electronic equalizers are becoming integrated by default into optical receivers in order to mitigate distortions which are ubiquitous on every fiber link. In this paper we investigate the possibility of using these already available components for identifying and quantifying optical distortions and thus, performing network monitoring without additional expensive optical equipment.

Digital electronic equalization is a key enabler for the cost-effective deployment of 10 Gb/s optical transmission in metro and regional core networks. The wider dispersion tolerance enabled by digital equalization eliminates the need for Dispersion Compensation Modules (DCMs) and amplifier gain-stages, with correspondingly simpler installation, lower capital cost, and reduced operating footprint and power costs. Digital equalization counters both propagation and component distortion, and so permits the relaxation of component specifications. The use of digital-equalization-enabled Adaptive Distortion Tolerant Transponders on 10G DWDM line cards results in significant system-wide cost savings.

The new requirements of Metropolitan Area Networks (MAN) have resulted in the emergence of new data transport technology, Resilient Packet Ring (RPR). The Key performance objective of RPR is to simultaneously achieve high bandwidth utilization, spatial reuse, and fairness. In this paper, a Rate Estimation Based Fairness Algorithm (REBFA) in Resilient Packet Rings is described and simulations are given to prove REBFA can achieve better performance than the fairness algorithm described in draft of IEEE standard for RPR.

A kind of optical multi-wavelength label switching adopting Gigabit Ethernet technology is introduced. In this switching, optical header is labeled by several optical pulses at different wavelengths in the same optical communication channel band as optical payload, and a specific Gigabit Ethernet adapter is used as the optical payload sender and receiver. The principle of optical switching is explained, and a demonstration experiment is described.

We report here a scalable, multichassis, 6.3 terabit core router, which utilizes our proprietary optical switch. The router is commercially available and deployed in several customer sites. Our solution combines optical switching with electronic routing. An internal optical packet switching network interconnects the router’s electronic line cards, where routing and buffering functions take place electronically. The system architecture and performance will be described. The optical switch is based on Optical Phased Array (OPA) technology. It is a 64 x 64, fully non-blocking, optical crossbar switch, capable of switching in a fraction of a nanosecond. The basic principles of operation will be explained. Loss and crosstalk results will be presented, as well as the results of BER measurements of a 160 Gbps transmission through one channel. Basic principles of operation and measured results will be presented for the burst-mode-receivers, arbitration algorithm and synchronization. Finally, we will present some of our current research work on a next-generation optical switch. The technological issues we have solved in our internal optical packet network can have broad applicability to any global optical packet network.

This paper is devoted to dimensioning and optimizing optical buffers for asynchronous, variable length packets in GMPLS-based WDM optical packet-switched networks, which aims to lower network congestion and increase utilization efficiency of optical buffers. In GMPLS-based WDM optical packet-switched networks, the critical issue is how to improve network flexibility and limit network congestion. Comparing to optical buffers with a single operating wavelength, WDM optical buffers in an optical packet switch (OPS) have demonstrated superior performance in solving congestion, which perform buffering by exploiting both time and wavelength dimensions. However, the realistic buffering capacity of optical fiber-delay-line (FDLs) buffers is currently limited. Hence, the algorithm for wavelength sharing and design of the optimal value of the basic time unit of the FDLs in WDM optical buffers are indispensable to shorten FDLs and put OPS into practice. In this paper, the algorithm for wavelength sharing among multiple connections is proposed, which aims to minimize congestion and shorten the queue length. And, in this scenario, the optimal value of the basic time unit of the FDLs is designed. Finally, the proposed algorithm is compared with others in terms of performance.

The main challenge concerning the design of an optical packet switch is the issue of contention resolution. This paper introduces an optical packet switch architecture, with a two-tiered partially shared buffer using a novel buffer scheduling scheme. The traffic scheduling in the partially shared buffers is done in a hierarchical fashion to ensure the maximum router throughput. We compare the performance of the proposed strategy with single and multiple groups of partially shared buffers, keeping the overall buffer size constant. The performance analysis has been done for a Bernoulli distributed traffic at varying loads. The simulation results indicate a significant improvement in the packet loss probability. A remarkable advantage of the suggested architecture is that it obviates the need for wavelength conversion, thereby achieving a reduction in the infrastructural cost involved in the employment of wavelength converters. The secondary benefits include the elimination of the delay incurred due to searching for the appropriate wavelength and tuning of the wavelength converter. Also, the complexity associated with the accommodation of packets in partially shared buffers is reduced from O(n) to O(1) in comparison to n-groups of partially shared buffers as the search for the least queued partial buffer is replaced by a Boolean decision for every contending packet. Thus, in addition to enhancing the router performance in terms of the packet loss probability, the proposed strategy also augments the router speed.

We present both optical and electrical compensation of Differential Mode Delay (DMD) in Multimode Fiber (MMF) links. Based on the rigorous analysis of optical compensation of DMD, a new kind of Dispersion Compensating Fiber (DCF) is proposed. We show that 10Gb/s Ethernet reach for conventional MMF can be extended to 300m by splicing 10~35m of this DCF. We prove that ISI resulting from DMD can be mitigated with electrical compensation by using linear or nonlinear equalizers. Moreover, we demonstrate that combination of optical and electrical compensation can dramatically enhance the bandwidth of installed MMF. Simulations carried out on six types of typical installed MMF for Overfilled Launch (OFL) as well as Restricted Mode Launch (RML) condition exhibit that transmission distance for conventional MMF with bandwidth-distance product 500MHz-km can be extended to 1000m or beyond with 2dB Eye-Opening Penalty (EOP).

A numerical model considering carrier heating and spectrum hole burning effects is given to study the pattern effect of clock recovery based on injection mode-locked SOA fiber laser, which is applicable for ultrashort pulses with several picoseconds pulsewidth. The reduction of pattern effect in the recovered clock at 40GHz successfully demonstrated by using a CW assist light at the transparency wavelength or gain region to SOA. The obtained minimum recovery time is less than 10ps, amplitude fluctuation will reduce from 64% to 5% and timing jitter will reduce from more than 3ps to 0.5ps. Moreover, the dependence of pattern effect on the power and wavelength of CW assist light and the bias current of SOA is analyzed, which is a useful guide to optimize this system. In order to get lower pattern effect, the required injection power of CW light at transparency wavelength will be higher than that of a wavelength in gain region, the minimum pattern effect occurs at a wavelength around the gain maximum. In this sense, a CW assist light in gain region is superior to the assist light at transparency wavelength. This scheme is an attractive method for SOA-based clock recovery and other all-optical signal processing.

Storage Area Network (SAN) has gradually developed as the demand for storage capacity and fast access has increased. The traditional way of attaching storage directly to the servers over a SCSI bus has limited scalability. Several drawbacks and limitations have turned up. Switched Fibre Channel SAN resolves all of these issues. In this paper, the architecture of the switch fabric for the SAN is discussed. The complete design of the free-space optical switching core based on the diffractive element and the PLZT shutter is proposed.

Proposed in this paper is an extension from a basic 2×2 contention resolving scheme to a complex N×N scheme for asynchronous variable length packets, through which contentions among multiple packets will be resolved. At first, a fast comparison algorithm is presented to build N×N all-optical contention resolution node based on 2×2 contention resolving scheme. Given the comparison algorithm, the number of basic 2×2 switching node that an N×N node demands are calculated and discussed, corresponding simulations are conducted and the packet-loss probability is tested. Moreover, further simulation is performed to evaluate the performance of the 8×8 bufferless all-optical contention resolving scheme equipped with limited-range wavelength converters (LRWCs).

A long fiber ring as the simple filter for carrier- suppression of microwave photonics signals is proposed. It is based on stimulated Brillouin scattering. The filter can be used to reduce RF loss in radio on fiber system. The modulation depth and RF signal gain increase with the power of input modulated signal. A modulation depth of 100% can be easily attained at a lower input power. The experimental results are well in agreement with the theoretical predictions. Approximately 26 dB of RF signal gain in the band from 1 GHz to 18GHz is achieved in the experiment with this technique. Finally, the transmission experiment shows that the performance of radio on fiber system is effectively improved with the filter.

This paper focuses on the problem of designing a large 256×256 high-performance broadband packet (or ATM) switch. Early packet switch research concentrated on using N×log₂N structures with buffers at each switch element. As a result, the nude number and the cost of such kinds of switches become prodigious and consequently could not have more practical significance. In this paper, we provide a method to design a kind of architecture containing fewer nodes. The distribution network of this new architecture based on the knockout theory has very low internal blocking, similar to that found in large 256×256 single crossbar networks, but the complexity and cost are less than that of single crossbar networks. A Balanced Incomplete Block Design (BIBD) in Combinatorics is provided to help construct the architecture. With this method, the mapping function between the input ports and the crossbars can be realized by a serial of mutually orthogonal Latin Squares (MOLS). Moreover, statistical tools are used to calculate the possibility of internal blocking of the network and to make a comparison between the improved architecture and the common one.

This paper introduces a design for a high capacity multicasting capable Optical Packet Switched Router (OPSR). It incorporates an Arrayed Waveguide Grating (AWG) as its core switching matrix to manage the ever increasing packet switched traffic within the optical network. The OPSR has the capability to switch optical packet traffic through the AWG with a switching speed of a few nano-seconds (nsec) at a bit rate of 10Gbps and above with multicasting capability. The capabilities offered by this OPSR will speed up all-optical packet switching and reduce packet latency and consequently packet loss.

We consider the problem of routing and wavelength assignment problem (RWA) in wavelength-routed networks, which consists of the Wavelength Division Multiplexers (WDM) and optical cross-connect (OXCs) devices interconnected via point-to-point WDM links. Assume that none of the OXCs has wavelength conversion capability. In this paper, we develop a set of dynamic distributed source-based DLE algorithms to provide primary and backup paths for connection requests with protection guarantee under single-link failures with local information. For better resource utilization, we employ multiplexing techniques, named primary backup multiplexing, to maximize the wavelength sharing among independent protection lightpaths. This technique allows a wavelength channel to be shared by a primary and one or more backup paths. We formulate the problem in link-based restoration context. The algorithm support on-demand path computation. Another main contribution is the introduction of the light-weight aggregated link metrics termed "barrel", we provide guidance to construct the protection routes with minimal wavelength consumption. We also compare the performance of various routing algorithms through simulation studies.

This paper proposes a novel 4×4 free-space bi-directional fiber optical switch. The switch is independent of the optical polarization, which can lead to polarization-dependent loss. When the input light beam incident on the polarization beam splitters (PBS) coating satisfies Brewster Condition, it is decomposed into two orthogonal linear polarized lights: P-polarized light and S-polarized light. In the paper, the free-space optical switches elements include a polarization control unit and a routing unit, and the former will be discussed emphatically. The polarization control unit can be acted by Ferroelectric Liquid Crystal (FLC) or half-wave plate array. By comparing the two methods, half-wave plate array will be chosen in the experiment. The controlling circuit of driving half-wave plate array is designed and polarization-dependent loss (PDL) in 4×4 optical switch using half-wave plate array is analyzed. The experiment results indicate that the insertion loss is less than 4.4 dB, the interchannel crosstalk is about 32 dB and the switching time of the optical switch is about 2 ms.

The Internet is growing, and terabit routers are needed. We will discuss two scalable architectures, and their performances: the packet switch with input buffers based on a cross-bar fabric, and the packet switch based on a Clos fabric and load balancing. Both architectures provide non-blocking, and, therefore agile admission control for unicast and multicast traffic. They will be also shown to provide rate and delay guarantees to the sensitive applications.

In this paper, we propose a novel all-optical AND gate using frequency shift accompanied by cross-phase modulation in semiconductor optical amplifier. Theoretical model is built to analyze the performance of the proposed gate. Through numerical simulations, we find the bandwidth of the optical band filter can be optimized to improve the quality of the output signals. The AND gate with short pulse-width and high energy of the control signals is adaptive to high-speed optical packet-switched and label-switched network.

Using a conventional optical receiver model, we analyzed the penalties on the optical signal performance due to low-frequency subcarriers with a new approach, regarding the subcarrier as a sinusoidal distributed random variable of which the standard deviation is applied in the formula derivation. Then, the Q factor is derived as a simple analytic function of modulation index, OSNR, and the number of subcarriers. By using numerical analysis, our method agrees well with the proved theory and is more efficient and convenient for calculation.

Stimulated Brillouin scattering (SBS) is one of important nonlinear effects in optical fiber communication systems. Due to its low threshold, it limits the optical power that couples into the fiber. In this paper, Stimulated Brillouin scattering and the effect of phonon loss to SBS in optical fiber communication systems are investigated. In experiments we measure first the power of backward Stokes wave and output light under various input powers of the systems with different lengths of fibers. Then for observing the effect of the phonon blocker to the SBS threshold, we repeat the same experiments but with a phonon blocker inserted at different locations of the fiber link. The experiments indicate that adding phonon blocker to the fiber is able to increase the SBS threshold, and the blocker at different locations makes different effects.

We study the optical bistable properties in one-dimensional photonic crystal with symmetric structure (AB)^N(D)^M(BA)^N, where A is high refractive index medium, B is low refractive index medium, D is defect layer, N is the number of layer periods at the left and right of the defect layer. M determines the thickness of the defect layer with Kerr medium. First, we analyze mechanisms producing bistability by the structure consisted of a nonlinear layer sandwiched between Bragg reflectors. Then by using the transfer matrix method, we calculate numerically the threshold value of the bistability switching in one-dimensional photonic crystals with structure (AB)^N(D)^M(BA)^N. The investigations show that the number of layer periods (N) and the thickness of the defect layer (M) influence the threshold value of the bistability switching. We confirm the parameter of the bistability switching to produce the lower threshold value.

Optical signal quality monitoring is an important function for optical transport networks and future all-optical networks. To monitor the optical signal-to-noise ratio and/or waveform distortion transparently with respect to the signal format, data format, and signal bit rate, we introduce an optical signal quality monitoring method that uses asynchronous sampling, which is a sampling technique that does not use timing extraction. The use of high-speed asynchronous sampling and the adjustment of the sampling rate enable simple open eye-diagram monitoring and evaluation of a fixed-timing Q-factor (Q_t) at the maximum eye opening timing phase. This method was experimentally verified using an optical signal quality monitoring circuit, and obtained a good relationship between the measured Q_t and Q (which is a Q-factor calculated from the bit error rate (BER)). Moreover, we also introduce an average Q-factor (Q_avg) evaluation method, which measures the Q_avg value from an asynchronous eye-diagram (timing drifted eye-diagram). This method is useful when the sampling rate is low or when adjusting the sampling rate is difficult.

Based on the atmospheric optical irradiance scintillation channel and an avalanche photodiode detector (APD) receiver communication system, this paper analyzes the error probability performance of digital pulse interval modulation (DPIM). Both the theory and the simulation results show that compared with PPM, the DPIM has a similar transmission efficiency, less bandwidth requirement, a marginally inferior error probability, and less complicated to implement. Hence, DPIM is more favorable in wireless optical communications system.

Based on the analyses of the key limiting factors for duobinary signals in dispersion limited fiber systems, we propose a novel optical receiver design that can further overcome the deterministic timing 'jitter' resulted from the dispersion-induced inter-symbol interference. Our idea is based on over-sampling of the received waveform and a novel decoding scheme. Our technique would provide better estimations of the received bits based on a decision feedback and feedforward scheme that takes advantage of the knowledge of adjacent bits. We study, through extensive simulations, the effectiveness of our novel receiver design in increasing the dispersion tolerance, and our promising results show that our innovative solution can provide significantly improvement over even the best reported schemes thus far.

Semiconductor Optical Amplifiers (SOAs) are vital elements in future optical networks whether as amplifying elements to boost the optical signal. In addition to be used as amplifiers, SOAs can also be used as switching elements operating either as ON-OFF switches or as wavelength converters for wavelength routing switching. Additionally, their performance is compatible with nowadays and future transmission rates of 10, 40 and 80Gb/s and beyond. Hence, the impact they will have in all future optical networks is paramount and it is unimaginable that there will be any future optical network without them. With this knowledge a look into future optical networks reviles that the modulation format will be influenced by this development due to the physical impacts of the SOAs on the signal transmission. Hence the modulation format needs to be investigated. Here we pursue the investigation of different optical formats by means of optical simulation and do a comparison of the modulation formats with respect to the performance selected for this paper.

In today's terrestrial long-haul optical fiber communication systems, high channel powers are required to obtain a large transmission distance with reasonable optical amplifier spacing. In such systems, however, the presence of nonlinear effects such as the self-phase modulation (SPM) and the fiber dispersion as well as their combined effects, called SPM-induced nonlinear limitation or SPM limit, will seriously degrade the system performances in respect of the effective transmission distance and ultimately become a limiting factor in high-speed, long-haul optical fiber transmission.In this paper, a new transmission format: chirped duobinary PolSK transmission, has been proposed to generate a pre-chirped duobianry signal with fixed polarity (either positive or negative), which is modulated by a PolSK modulator. This format is based on a transmitter setup consisting of a duobinary PolSK Modulation transmitter followed by an additional phase modulator. The chirped duobinary PolSK transmission reduces the signal degradation and spectral broadening in the nonlinear regime significantly. Thus it shifts this SPM nonlinear limit to enable more relaxed dispersion compensation at high optical power compared to the conventional duobinary schemes.The simulation results show chirped duobinary PolSK transmission enlarges the dispersion limited transmission distance, increases the dispersion tolerance and overcome the SPM nonlinear limit.

A novel method to generate carrier-suppressed return-to-zero (CSRZ) pulses by using a sampled fiber Bragg grating (FBG) is proposed, and the transmission characteristics of the CSRZ pulses are simulated and discussed. This CSRZ pulse has a duty cycle of 50%, narrower than conventional CSRZ pulse, which is due to the filtering and shaping of FBG, and leads to increase of nonlinearity and dispersion tolerance in high-speed transmission systems. Through proper design of FBG and adoption of short pulse source, this technique provides a promising way to modify the pulse shape and consequently improve system performance.

In this paper, an overview of promising solutions for multi-level modulation formats, primarily based on phase shift keying (PSK), in optical WDM-transmission is presented and some recent results are summarized. With PSK modulation the option of multi-level signalling is offered which means that more than one bit is mapped on one transmitted symbol and thus offering bandwidth reduction for increased spectral efficiency. We investigate the performance with respect to receiver sensitivity, non-linear fibre impairments, chromatic and polarization mode dispersion. Last but not least the implementation-effort at high speed must comply with economic constraints.

The paper compares the effectiveness of four dispersion compensating schemes such as using dispersion compensating fiber (DCF), optical phase conjugator (OPC), OPC and distributed Raman amplifier (DRA) combination and apodized linearly chirped fiber Bragg grating (CFBG) in terms of the eye-opening penalty (EOP) at the end of the link. The comparison is also carried out with different modulation formats.

Using the degree of polarization as feedback signal to detect the polarization mode dispersion (PMD) variation in the fiber links, an experiment of dynamic adaptive PMD compensation in 4×10 Gb/s optical time-division multiplexed (OTDM) transmission system is reported. The PMD compensator consists of polarization controller and variable delay line with four degrees of freedom. Based on the variable step-size peak search algorithm, the performance of the PMD compensator is assessed in the eye pattern of a received signal. It is shown that the maximum compensated DGD is 25 ps, i.e. one bit period, and the compensation time can be less than 50 ms.

In this paper we report a successful experiment of adaptive polarization mode dispersion (PMD) compensation up to second-order in a 40Gb/s optical time-division multiplexed (OTDM) communication system by using two-stage compensator. In the experiment the PMD monitoring technique based on degree of polarization (DOP) was adopted. And the Particle Swarm Optimization (PSO) algorithm was introduced in adaptive PMD compensation, with the desirable features of fast convergence to the global optimum point for compensation without being trapped in local sub-optima and with good robustness to noise. The comparison was made to estimate the performance effectiveness between PSO algorithms with global neighborhood structure (GPSO) and with local neighborhood structure (LPSO). The LPSO algorithm is shown to be more effective to search global optimum for PMD compensation than GPSO algorithm. The ability of tracking changed PMD using PSO algorithm was also investigated. The two-stage PMD compensator is shown to be effective for both first- and second-order PMD, and the compensator is shown to be bit rate independent. The compensation time is within several hundreds of milliseconds. The response time for recovery from a sharp disturbance is about 11ms.

In this paper we investigate quaternary combined amplitude- and differential binary phase-shift keying (ASK-DPSK) and 8-ary combined amplitude- and differential quadrature phase-shift keying (ASK-DQPSK) at a bit rate of 40Gbit/s. For both formats, optimal values for amplitude ratios of the signal points and electrical and optical receiver bandwidths are given. Required optical signal-to-noise ratios (OSNR) at the receivers for bit error probabilities of 10^(-9) and chromatic dispersion tolerances are compared to amplitude-shift keying (ASK), binary differential phase-shift keying (DPSK) and quadrature phase-shift keying (DQPSK). For our numerical investigations we use a semi-analytical method for bit error probability computations.

In this paper, the performance of mid-link spectral inversion (MLSI)-based transmission is quantitatively compared to the performance for 'conventional' dispersion compensating fiber (DCF)-based transmission. Bit error rates (BER) are measured at optimized residual dispersion for all sixteen channels after 800km transmission in standard single mode fiber (SSMF). We show that for the MLSI-based configuration up to a decade of improvement in BER is obtained compared to the DCF based system. MLSI has the potential of lowering the cost of long haul transmission systems since instead of having multiple DCF modules per span only one spectral inverter is required for the whole transmission line. For spectral inversion, a polarization independent magnesium-oxide-doped periodically-poled lithium-niobate (MgO:PPLN) based subsystem was used.

Analytic and approximate traffic model is proposed for asynchronous optical packet switching (AOPS) networks to serve the network performance analysis. The study shows that AOPS network traffic could be modeled as Poisson arrival with Gaussian-distributed packet size, when the timeout-based optical packet assembly algorithm is adopted and the incoming IP traffic at AOPS edge routers is short-range dependent. These findings are helpful to accurately dimension the traffic model for AOPS networks.

This article discusses the basic structure and physical characteristics of liquid crystal and gives a description about the application and research status of liquid crystal technology in the field of optical communications. It studies mainly on several practical applications of liquid crystal optical component, simulation technique of turbulent atmosphere and the Space Light-beam Modulator (SLM) in the satellite optical communications system. The scheme of a simulating high precision beaconing light source modulated by liquid crystal is discussed in the paper in detail, at the same time presenting some analyses of the development trend on new and most advanced correlative technology, and pointing out that the liquid crystal technology will become one of the master technologies platform in the field of optical communications.

Analytical model to characterize fiber nonlinearities in the WDM Simulation Software have been applied in this paper, and the simulation results of software are shown. This model is very instructive to system simulation, network programming and design.

Photonic Design Automation (PDA) describes the design methodologies, software tools and services used to engineer complex photonic networks and products. It offers software integration along the signal path from transmitter to receiver, across the value chain from component and systems manufacturer to network operator, and allows to reduce number of costly and time consuming lab experiments and field trials investigating new system architectures and characterizing optical components. This presentation discusses various modeling aspects along the optical propagation channel. The concept of multiple optical signal representations is briefly introduced, which allows distinguishing between data signals, optical noise, distortions and crosstalk throughout the fiber transmission. The trends of modeling optical transmitters, fibers, doped fiber amplifiers, and of estimating the bit-error rate are outlined. The optimization of multiple pump lasers to achieve high-gain bandwidth fiber Raman amplifiers is discussed. The link of simulation models with real-world component measurement is discussed on an exemplary basis.

In this paper, for the first time, we analyze the optical exclusive OR (XOR) gate based on cross-polarization modulation (XPolM) effect in semiconductor optical amplifier (SOA) using nonlinear polarization rotation (NPR) theory. The extinction ratio (ER) of optical XOR gate, which reflects the performance of the gate, is calculated in consideration of the injected current, the length and the polarization angle of input signals of the SOA. The performance of the optical XOR gate can be optimized for some proper parameters of the SOA.

A theoretical model of clock recovery based on SOA fiber laser is presented, through which a complete numerical analysis about the clock characteristics at 40Gbit/s rate is given, which is an effective guide for experiment and necessary to optimize the system performance. The crucial parameters that determine the pulse width, energy and frequency chirp of the recovered clock pulses are investigated, including the pulse width and energy of the external data signals, the small signal gain and carrier lifetime of SOA, and the cavity loss. Injection mode-locked SOA fiber laser is not only suitable for bit clock recovery but also for frame clock recovery. A frame clock recovery scheme based on SOA is proposed, which requires the data packet to have a fixed header marker to generate a stronger frame clock frequency component. Simulation results show a high-quality frame clock can be obtained using this scheme. In experiment, bit clock recovery is realized at bit rates up to 20Gb/s using injection mode-locked SOA fiber laser. In addition, 2.5GHz frame clock is also extracted from 8x2.5Gb/s and 16x2.5Gb/s OTDM signals. The recovered clock pulses are wavelength tunable and very stable, which can be used for high-speed all-optical signal processing.

We demonstrate a method to generate ultra-high-speed phase correlated optical signals using a Kerr shutter. Two examples are provided: 160-Gb/s CSRZ with pair-wise alternating phase, and 320-Gb/s CSRZ signals. We also discuss other phase correlated data signals including RZ-VSB, DPSK, and CSRZ with alternating phases in groups.

A high-performance switched optical interconnection network is designed. The top layer of optical fiber ring network builds upon wavelength division multiplexing and hardware routing technologies. The bottom layer of star network has a throughput of 10 Gbit/s based on an 8-channel digital cross-point switch. Two types of optical network interface cards are developed to meet the requirements of interconnection bandwidth and support PCI bus full bandwidth of 1.056 Gbit/s and 4.224 Gbit/s, respectively. Field programmable gate array (FPGA) is adopted for frame head analysis, hardware routing and dynamic switch configuration. By adding a subnet to enlarge the network, the maximal delay increment is only 1.5 μs.

Currently existing dense wavelength division multiplexing (DWDM) networks start to migrate from numerous point-to-point links towards meshed, transparent, optical networks with dynamically routed light paths. This increases the need for appropriate network monitoring and supervision methods. Optical performance monitoring (OPM) has to be cost-effective. Thus additional spendings for OPM have to be significantly smaller than the savings in OPEX due to increased reliability and ease of operation, administration, and maintenance (OAM). We elaborate on different advanced monitoring concepts. First, we discuss general failure scenarios in meshed networks. Then we describe software based failure root cause analysis and its implementation. We conclude that by implementing appropriate software algorithms in the network hardware effort can be significantly reduced. Finally, we assess different advanced OPM methods which may show up as useful to enable OPM in future optical networks.

This overview will discuss core network technology and cost trade-offs inherent in choosing between "analog" architectures with high optical transparency, and ones heavily dependent on frequent "digital" signal regeneration. The exact balance will be related to the specific technology choices in each area outlined above, as well as the network needs such as node geographic spread, physical connectivity patterns, and demand loading. Over the course of a decade, optical networks have evolved from simple single-channel SONET regenerator-based links to multi-span multi-channel optically amplified ultra-long haul systems, fueled by high demand for bandwidth at reduced cost. In general, the cost of a well-designed high capacity system is dominated by the number of optical to electrical (OE) and electrical to optical (EO) conversions required. As the reach and channel capacity of the transport systems continued to increase, it became necessary to improve the granularity of the demand connections by introducing (optical add/drop multiplexers) OADMs. Thus, if a node requires only small demand connectivity, most of the optical channels are expressed through without regeneration (OEO). The network costs are correspondingly reduced, partially balanced by the increased cost of the OADM nodes. Lately, the industry has been aggressively pursuing a natural extension of this philosophy towards all-optical "analog" core networks, with each demand touching electrical digital circuitry only at the in/egress nodes. This is expected to produce a substantial elimination of OEO costs, increase in network capacity, and a notionally simpler operation and service turn-up. At the same time, such optical "analog" network requires a large amount of complicated hardware and software for monitoring and manipulating high bit rate optical signals. New and more complex modulation formats that provide resiliency to both optical noise and nonlinear propagation effects are important for extended unregenerated reach. More sophisticated optical amplifiers provide lower noise for increased reach and increased spectral bandwidth for higher wavelength count lower wavelength blocking probability. Optical analog networks also require methods for mitigating optical power transients, for controlling optical spectral flatness, and for dynamically managing changes (e.g. in chromatic dispersion and polarization mode dispersion.) Since signals stay in the optical domain, optical performance monitoring techniques are required for fault isolation and correction. Efficient routing of optical signals also requires sophisticated switching nodes with an ability to selectively steer optical signals into several directions with single-channel spectral granularity. Most of these technologies are not modular and require interruption of service if not deployed at the initial system installation, thereby increasing first install costs substantially, even if initial capacity loading is small. Further, validation of systems and software targeting a specific network design is complex. Only a small fraction of the total network may be reasonably reproduced in the lab, and many field configurations are not predictable or even dynamic. Thus, extra system margin has to be allocated to handle the behavior uncertainty. To constrain the complexity of both hardware technology and software algorithms, regions of network transparency can be established with OEO forced at perimeters. Thus, "analog" regions are surrounded by "digital" interfaces. Following are some example tradeoffs that will be discussed. What is a good modulation format choice, and does increased reach and impairment resiliency justify hardware and controls that are more complicated? What are reasonable amplifier choices to make under specific network assumptions? Can high transport system capacity be leveraged to simplify optical switch node design by reducing spectral efficiency?

The MOEMS 2×2 optical switch with slant lower electrode and with torsion beam on silicon is designed and analyzed theoretically. Analytical formulae for the squeeze film damping coefficient and the squeeze film damping moment on the cantilever beam of the optical switch are derived. Based on the torsion dynamics theory, the technique and relative results are presented for analyzing the actuating voltage and the switch time. The optimized result of parameters is as: length, width and thickness of the torsion beam are 700, 12 and 10 um, length and width of the cantilever beam are 1900 and 1000 um, length and width of the balance beam are 100 and 1000 um, shortest spacing between the upper and lower electrodes is 0.05 um, and highness of the lower electrode is 55 um, respectively. The actuating voltage is less than 10 V, and the switching time of Ton and Toff are 1.30ms and 1.25ms, respectively. The computed results show that the air squeeze film damping is an important factor for the study of dynamic response on MOEMS optical switch.

All-optical switch is a key component for all-optical communication and switching; we demonstrate an all-optical switch with two switched output-ports like a nonlinear optical loop mirror (NOLM)based on a 3x3 collinear fiber coupler. After introducing the principle in this paper we give some pulse switching experimental results. In order to avoid high power laser system and save the cost the NOLM loop is made up of 12km G.652 fiber, the extinction ratio of switching signal is more than 20dB. Further we optimize the loop and EDFA, without reducing the extinction ratio we observe the pulse switching in a loop length less than 6km G.652 fiber. Compared to conventional optical switches, this type switch can support two switched output-ports which act complementarily at the same time, this character is practical for optical switching, and data signal can be selected to one of the multi-output ports by optical control signal.

Optical switching technologies exhibit many characteristics that the electronic switching technologies do not have. One such characteristic lies in switching states. An electronic crosspoint, having open and close as its two switching states, only allows one traversing signal. But an optical switching crosspoint, such as a directional coupler or a MEMS mirror, usually allows two traversing signals. This paper studies how the characteristic can expand the switching capabilities of an optical switch and shows new ways of building more crosspoint-efficient nonblocking switches.

A optical measurement system of Micro-electro-mechanical system (MEMS) motion is demonstrated to increase the efficiency of MEMS design and manufacturing. The system is developed based on the heterodyne laser Doppler technology, and frequency shift of 40MHz is introduced by a Bragg cell. The paper introduced MEMS vibration theory, designed optical structure and electronic circuit, and obtained the linear relation between the Doppler shift and the target velocity. Experiments on driving and measuring double-ended tuning fork resonator were carried out. The resonator is driven by using electrostatic force at its natural frequency 2.4KHz. Experimental results indicate that the measure system can be used to measure MEMS motion with high accuracy.

From element automatic control view, we propose a configuration algorithm for three-level cross-connects in data plane to handle with bypass, grooming and local add/drop traffic of fiber-level, band-level and wavelength-level by a abstract bipartite graph of MG-OXC and bandwidth utilization spectra graph. The configuration algorithm is evaluated by computer simulation as well as validated by experiment on our flexible Multi-functional Optical Switching Testbed (MOST).

In this paper four-wave mixing (FWM) and its impact on multi-span NRZ-modulated wavelength division multiplexing (WDM) systems are examined. The impact of the inline dispersion compensation map on FWM impairments is crucial. From re-circulating loop experiments as well as our analytical model worst-case conditions for the dispersion map are found. Furthermore, the impact of increasing the number of WDM channels is investigated. An analytical model is presented to assess the signal degradation. The impairments due to FWM are related to a Q-factor or an EOP. The presented formulas are applied to different dispersion compensation schemes and also mixed-fiber systems. The analytical model is verified by system simulations employing the split-step Fourier method as well as re-circulating loop experiments.

The nonlinear Schrodinger equation with variable coefficients is analyzed by means of projection matrix method. An exact analytical solution is obtained, which clearly shows how the variable fiber dispersion, nonlinear, and loss coefficients affect the propagation of ultrashort optical pulses. The obtained solution is used to analyze the propagation properties of ultrashort pulses in dispersion-decreasing fibers. It is found that the ultrashort pulse can realize stable soliton transmission if the fiber dispersions have some certain profiles related to the fiber loss and nonlinear properties. A small variation in the dispersion has a similar perturbative effect to an amplification or loss. The exponentially dispersion-decreasing fiber is studied exemplificatively to demonstrate the obtained results.

Under the influence of self-frequency shift, the interactions between in-phase and out-phase neighboring fundamental and second-order optical solitons are investigated numerically, and the impacts of soliton interactions to timing jitter are analyzed. It is found that under the influence of self-frequency shift, the periodic collision of neighboring fundamental in-phase soliton pair is broken. They are apart from each other rapidly after one collision and the self-frequency shift phenomenon is much more obvious after the collision. While for neighboring out-phase fundamental soliton pair, two solitons both shift to the dropping edge and the impacts of self-frequency shift are weaker than that of in-phase soliton pair. For second-order solitons, either in-phase or out-phase soliton pair will be split. Two split stronger solitons will collide with each other during the propagation in the optical fiber and the difference between in-phase soliton pair and out-phase soliton pair exists that the interactions of out-phase pair is weaker than that of in-phase soliton pair and the collision distance of out-phase pair is much longer than that of in-phase soliton pair. A nonlinear gain can be used to effectively suppress soliton interactions as well as effects of soliton self-frequency shift, and stabilize the soliton propagation.

Polarization mode dispersion (PMD) is the differential group delay between two differently polarized modes in single mode fibers, which is a serious limiting factor for high bit rate optical telecommunication systems. In our previous work, a novel measurement method of PMD in optical fibers has been proposed, which is based on coherent optical frequency domain reflectometry technique (OFDR). In the method, a frequency-shifted feedback (FSF) fiber laser was developed as light source; PMD is determined by self-delayed heterodyne detection from the beat frequency generated by interference between lights from the Fresnel reflection at the far end of the fiber. In this work, both theoretical and experimental comparisons are made between the proposed method and other International Telecommunication Union-Telecommunication standardization (ITU-T) recommended methods. Up till now, several kinds of methods have been proposed by different groups. According to the bandwidth of light sources, however, two categories can be distinguished: narrow band (quasi-monochromatic) wavelength tunable laser and broadband sources. The characteristic of our proposed method different from the two types of source in that it utilize a FSF laser, with a bandwidth between the narrow band and broadband. We analysis the difference from the view point of laser theory, and imply the effects on PMD measurements. Also experimental comparisons are made with ITU-T PMD Round Robin samples, and experimental results verified the effectiveness of the novel OFDR technique based method.

Bi-directional transmission of optical data over a single fiber reduces the first installed cost as well as the operational cost of DWDM systems. This paper discusses the special design issues, when a single amplifier and a single dispersion compensation module are used for both traffic directions. Results of loop transmission experiments over 5000 km of NZDSF are demonstrated.

The advantages of low cost amplifier solutions in single-channel link extender or loss compensator systems cannot be fully realized unless the ASE noise around the signal peak is removed. Doing so requires a cost-effective solution with high performance, including low insertion loss (<-2.5dB), low PDL (<-0.25dB), low power operation (<200mW), and fast tuning (<1sec). We have successfully fabricated and packaged a tunable ASE filter into a small form-factor 2-port package which meets these requirements. We obtain filter properties at both the chip and package-levels and examine filter performance operating under optically open and closed loop control.

VSR4 links use graded index multimode fibers (GIMMFs) as the transmission medium with operation wavelength 850nm. For cost reasons, VCSEL has been selected as the optical source to VSR4. The minimum bandwidth specification for 62.5um GIMMF in VSR4 is only 400 MHz•km for over-filled-launch (OFL) condition. The distance of 300 meters is limited over transmission rates of 1.25Gbit/s on the basis of this specification. In order to overcome the OFL bandwidth limit by selective excitation of a limited number of modes, conditioned launch technique is investigated. In this paper, based on a comprehensive dispersion theory of GIMMF, a model is built to simulate the transmission of optical signal in GIMMFs and a comparison between OFL and conditioned launch is analyzed. The result can be the guidelines for the best choice of techniques for various LAN and interconnect systems also.

We present a novel multiple access scheme for Passive Optical Networks (PON) based on optical Code Division Multiple Access (OCDMA). Di erent from existing proposals for implementing OCDMA, we replaced the predominating orthogonal or weakly correlated signature codes (e.g. Walsh-Hadamard codes (WHC)) by convolutional codes. Thus CDMA user separation and forward error correction (FEC) are combined. The transmission of the coded bits over the multiple access fiber is carried through optical BPSK. This requires electrical field strength detection rather than direct detection (DD) at the receiver end. Since orthogonality gets lost, we have to employ a multiuser receiver to overcome the inherently strong correlation. Computational complexity of multiuser detection is the major challenge and we show how complexity can be reduced by applying the turbo principle known from soft-decoding of concatenated codes. The convergence behavior of the iterative multiuser receiver is investigated by means of extrinsic information transfer charts (EXIT-chart). Finally, we present simulation results of bit error ratio (BER) vs. signal-to-noise ratio (SNR) including a standard single mode fiber in order to demonstrate the superior performance of the proposed scheme compared to those using orthogonal spreading techniques.

This paper introduces a novel tunable external-cavity diode laser based on the grating light valve (GLV) technology, which has the advantages of high speed, high reliability, low cost and compactness. A theoretical model is present on this tunable laser and the configuration parameters are calculated to illustrate the practical signifiance of the laser.

A concept for a new optical Ethernet interface card for FTTD (fiber-to-the-desktop) is proposed and application areas of the proposed card are discussed. As a first step towards realizing the concept, a one-fiber 100-megabit optical Ethernet PC card was fabricated by applying a hybrid integrated optical transceiver module and an MU connector. As a second step, a gigabit optical Ethernet PC card was fabricated for the first time and its performance was evaluated. These cards could become the standard in the coming optical FTTD/FTTN (fiber-to-the-notebook PCs) age because they are suitablie for notebook PCs and can operate in one-fiber WDM systems.

In this paper, we present a simple marker pulse extraction scheme using a modified terahertz optical asymmetric demultiplexer and experimentally demonstrate marker pulse extraction from 20Gbit/s all "1" packet. Particular attention is given to eliminate the effects of polarization rotation in semiconductor optical amplifier.

In this paper, self-phase modulation (SPM) effect in cascaded Mach-Zehnder Interferometers (MZIs) is studied thoroughly, first in single stage and then in multi-stages. Its application in all-optical self-synchronization is proposed. The SPM effect in cascaded MZIs is strengthened by assisting semiconductor optical amplifiers (SOAs), making the proposal simple, integrable and needing no special marker pulses. Furthermore, corresponding system performance is examined and advice is given on the designing of key system parameters.

In this paper, a novel scheme of stabilizing all-optical clock recovery is proposed, which uses a mode-locked fiber ring to extract the same frequency clock pulses from the uneven multiplexing 4×10Gbps optical signals. The super-mode suppression ratio reaches 60dB (for signals of 10Gbit/s) and an improvement of 20dB is achieved by adding stabilization devices. Wavelength shift is less than 0.01nm within 2 hours, and this value is about one sixties of that without adding stabilization devices. The results show much better cavity length stability and noise suppression performance.

To research the influence of optical switches on the optical true time delay network (OTTDN), the delay differences between the bar and cross states of a series of 2x2 optical switches are measured, and the results show that the average delay difference between the bar state and cross state is about 5.4ps. To verify this effect, a 2bit optical delay line based on 2x2 optical switches is constructed and measured. The time delay error is measured to be 5.34ps. Both analysis and experiment indicate that by replacing 2x2 optical switches with 1x2 optical switches, no error caused by optical switches remains.

A modification to the transmission profile of filter in the pre-filtering transmission is proposed. The system employing optimized filter surpasses the common choice in the performance of filtering and transmission. The improvement reduces the filtering penalty from 0.6dB down to 0.05dB, and maintains this advantage through transmission avoiding the interference with intra-channel nonlinearity. The conclusion has been approved by numerical simulation.

Deployment of 40-Gbit/s technology in metro and core networks is still attractive, to bring down costs and to increase transmission capacity. This paper summarises design issues of 40-Gbit/s WDM systems for their application in wavelength division multiplexed metro and core networks. Parameter tolerances and transmission distances for different modulation formats are numerically and experimentally investigated. Based on Deutsche Telekom's fibre infrastructure, upgrade studies show that polarisation mode dispersion will be the main obstacle when installing 40-Gbit/s technology in deployed fibre infrastructure.

The optimal length of the fixed delay line (FDL) for a one-stage optical polarization mode dispersion (PMD) compensator with two degrees of freedom (DOF) is investigated for a 40 Gbit/s NRZ optical fiber transmission system by implementing genetic algorithm in the logic-control unit. It is found that there exists an optimal length of FDL for such an optical PMD compensator (PMDC), and the optimal length of the FDL is equal to the PMD value of the preceding transmission fiber to be compensated for. Such an optimized one-stage PMDC with a fixed delay line can achieve a close performance to that of a one-stage PMDC with a variable delay line.

In this paper, for the first time, we analyze theoretically the wavelength conversion based on semiconductor optical amplifier (SOA) using nonlinear polarization rotation (NPR) theory. The non-inverted wavelength conversion is obtained through numerical simulation. The extinction ratio (ER) of the converted signal is over 20 dB, which accords with the experimental result. We investigate the performance of non-inverted wavelength conversion considering some parameters. The research result in this paper shows that we may accomplish wavelength conversion using NPR in SOA well.

Starting from the definition of the linewidth enhancement factor of the semiconductor optical amplifier (SOA), the linear relationships between nonlinear effects of cross-gain modulation (XGM), cross-phase modulation (XPM) and cross-polarization modulation (XPOM) in SOA are verified. Wavelength conversion based on XPOM in SOA is numerically analyzed using a simple rate equation model, which shows a high performance of this kind of converter. These results are practical for nonlinear applications of SOA. The advantages of this scheme include simple operation, inverted and non-inverted signal available, etc.

We proposed a new approach to enhance supercontinuum (SC) spectrum generation in dispersion-decreasing optical fiber with a new dispersion profile. The approach is based on the change of the chromatic dispersion D(λ,z) as a concave function of wavelengths that has no zero-dispersion wavelengths over the whole or part of the fiber as D(λ,z) is positive. A flatter and broader spectrum can be produced in our scheme. The general criteria parameters are obtained for characterizing SC shapes in the optical fibers.

We propose a polymeric variable optical attenuator using a modified wide angle X-junction structure. The device's size is only one sixth of that using conventional Y-junction structure. A commercial 3D-Beam Propagation Method simulation software was used to optimize and analyze our design, which was then fabricated using benzocyclobutene polyimide. The experimental results well match with the simulation results. The VOA has an attenuation range larger that 34dB for TE polarization by means of thermo-optic tuning. The measured response time was about 100 micro second, which is fast in compared with other thermo-optic device. The compactness of our device makes it very suitable for high density integrated photonics circuits.

A novel method for fast pump configuration of flat-gain Raman fiber amplifier is provided in this paper. With neglect the weak difference of Raman gain profile dependent on pump wavelength, the total gain of amplifier in dB unit can be viewed as output of a special linear time invariant (LTI) system. The LTI system characters its unit impulse response as the Raman gain coefficient spectrum, its input as pump configuration spectrum, and its output as the amplifier gain. Given the target amplifier gain, the continuous pump configuration spectrum can be easily derived by using fast Fourier transform (FFT). Then with clustering approximating of the continuous pump configuration spectrum, the pump configuration is obtained. The clustering approximation and its impact are also discussed. The numerical results reveal that for 13THz bandwidth Raman fiber amplifier, the maximum gain-inflatness of LTI system method is only 0.3dB worse than that of conventional Generic algorithm (GA) method, but the computation time is at least one magnitude saved.

The error probability of differential quadrature phase-shift keying (DQPSK) systems is analyzed considering both self- and cross-phase modulation (SPM and XPM)-induced nonlinear phase noise. XPM-induced nonlinear phase noise is treated here as Gaussian distributed phase noise. In DQPSK systems where group velocity dispersion (GVD) is perfectly compensated span after span, the influence of XPM-induced nonlinear phase noise on the error bribability performance can be dominant compared with that of the SPM-induced nonlinear phase noise. The 10-Gb/s systems, whose walkoff length is larger than that of the 40 Gb/s systems, are more sensitive to XPM-induced nonlinear phase noise.

The random deviation of the periodicity of the gratings will affect the performance of the fiber gratings. The random errors would not accumulate when the gratings were cascaded. But we found a kind of fabricating system errors induced by the method for the side writing of fiber gratings, which would accumulate when cascaded. So laser with the less pulse energy should be used to write the gratings to developing the system's performance.

To meet the demand for ever-increasing transmission capacity led by the increase in Internet traffic, up to 10 Tb/s transmission capacity experiments have been demonstrated using wavelength division multiplexing (WDM) and up to three transmission bands. Most of today’s commercial WDM systems, however, are capable of 80 channels at 10 Gbit/s in the C-band and similar capacity in the L-band. 40 Gbit/s channel rate WDM systems are not yet widely commercially deployed. To achieve the aforementioned multi-terabit capacity systems for the future high spectral efficiency and the opening of additional transmission wavelength bands will be necessary. Besides the already used conventional C-band and the long-wavelength L-band the short wavelength S-band is the most promising candidate for a third transmission window. A key technology for accessing a new transmission band is the availability of optical amplifiers, which is fulfilled for the S-band by using either gain-shifted thulium doped fiber amplifiers or new erbium doped fiber amplifiers. In this paper we will provide an overview of amplifier types and their possible usage to upgrade to multi-band transmission as well as we will discuss general design options for upgrading transmission bands. In particular, we will show numerical results for Raman based C- and L-band amplification with multiple Raman pumps and different pumping schemes and an experiment for opening up the S-band by a fiber amplifier approach.

Based on the assumption of a quasi-monochromatic wave for light in fibres and 2□2 coherency matrix, a relationship of DO P with several parameters as well as DGD is derived from the common definition of DOP. A normalized coherent coefficient, showing the degree of correlation of optical signals in both principal states of polarization (PSPs), is derived theoretically. The value of this parameter indicates directly the degree of polarization of optical signals within the communications fibers. The correlation is also the straightforward express of differential group delay between two PSPs. With a pulse shape of Gaussian, we give a useful expression of DOP with power splitting ratio and differential group delay. Two curves of calculations are also given. When DGD becomes bigger, the coherent coefficient is smaller, meaning that the DOP is lower. DOP is one of feedback control signals for dynamic PMD compensation.

Polarization dependent characteristics of the fiber-optic transmission systems become more important than ever before as the bit rates increase to 10 Gb/s and beyond. We derived the Jones matrix and Muller matrix of a polarization controller (PC), and in our experiment, we can get every wanted state of polarization (SOP) at the PC output with all kinds of input SOP. Using PC as a key component, we designed a polarization stabilizer. With the unknown change of the input SOP and random shift of the environment, the polarization stabilizer can maintain a fixed SOP through controlling the PC in real-time.

We completed a 5-degree of freedom PMD compensator with a novel feedback control algorithm. The tracking algorithm, which successfully solves the problem of easily been trapped into local optima is adopted in the tracking course with no visible worsening of the feedback signal in the compensation process. Long time testing shows good performance of our compensator for fast tracking and compensation.

The relationships between Degree of Polarization (DOP) and Polarization Mode Dispersion (PMD) for different modulation format signals are discussed not only in theory but also in experiment, which include the NRZ, RZ, CSRZ, DPSK, RZ-DPSK and CSRZ-DPSK signals. The RZ signals with different duty cycle are also analyzed and compared. The results are found that the optical signals DGD/DOP relation depends on their optical spectrum structure as well as the modulation formats, and the phase-shift-keying signal with free chirp will provide better performance in the PMD compensation based on the DOP feedback control because of the DOP sensitivity and wide monitoring window, besides, the CSRZ has the stalest character for the variety of modulation chirp.

To learn the surrounding conditions in the fiber link and its effect on PMD, and to provide the first-hand design basis, we have carried out the data observation of PMD in a fiber link for a long time. We have tested the first-order and second-order PMD. The fiber tested is the G652 fiber produced by Corning Co. of USA, and the testing distance is 1000km; n segments of same fibers are linked into one, and n equals to 40, that is to say, the length of every segment is 25km; for the requirement of dispersion compensation in the high-speed and long distance fiber optical communication system, one fiber grating dispersion compensator is added in the place of every 200km, and there are five compensators; one EDFA is added in the place of every 100km, and there are eleven EDFA. The result suggests that, with the increase of length of fiber link, the distribution of PMD intends to be stable, that is, with the number n increasing, the relative error of PMD becomes less. The testing methods are the Jones matrix eigenanalysis technique and interference technique. HP8509B fiber polarization analyzer of Agilent in USA is used for measuring instrument of the Jones matrix eigenanalysis technique; FPMD-5600 Femtosecond PMD Analyzer of EXFO in Canada is used for measuring instrument of interference technique. The difference between these two testing methods is analyzed. With the Jones matrix eigenanalysis technique, fibers of 1000km are inspected through 48 hours, and the result suggests that, at nine o'clock in the morning, PMD reaches the maximum, at nine o'clock in the evening, it reaches the minimum, during other time, its change is very little. So it can be concluded that, PMD in the long distance fiber link is affected by temperature of the lab. Stress testing is carried in the ultra-short fiber (less than one meter). PMD has no obvious change in the range of stress which can be endured by the fiber.

It is important to know how the degree of polarization (DOP) changes with differential group delay (DGD) when DOP is the feedback parameter for an adaptive polarization mode dispersion compensator (PMDC). We derived the mathematical expression of DOP for pulse sequence with arbitrary and Gauss waveforms under the condition of quasi-monochromatic wave, then analyzed theoretically the relation between DOP and DGD with different power splitting ratio and pulse width. Then we carried out the experiments of DOP with 10Gbit/s RZ/NRZ pseudo random sequence pulses and made comparison between theoretical calculation and experimental data. The comparison showed that they matched closely. We also verified by experiments that DOP is independent of fiber dispersion.

A novel serial scheme of header extraction for optical unslotted networks using SOA-MZI with differential modulation scheme was proposed in this paper for the first time. Numerical analysis and simulation show, that more than 15dB contrast ratio of the separated header at 2.5Gbit/s to the suppressed payload at 40Gbit/s could be achieved. In addition, the SOA-MZI parameters are discussed and designed to optimize the performance of the proposed scheme.

Wavelength conversion and optical buffer are the two most commonly used and most useful contention resolution schemes. This paper presents a comprehensive study of them in a multiwavelength optical burst switching network. Specially, we focus on 1) how does the wavelength conversion degree affect the burst blocking performance? 2) how does the optical buffer affect the burst blocking performance? Different from the studies in literature, we will not only investigate the effect of optical buffer length on the blocking performance, but also we investigate the effect of the granularity, which is defined the length of an FDLs and measured in terms of a delay unit, of optical buffers on the burst blocking performance. 3) how does various combinational schemes can be architected and how they perform.

Optical burst switching (OBS) is more efficient and feasible solution to build terabit IP-over-WDM optical network by employing relatively matured photonic and opto-electronic devices and combining the advantage of high bandwidth of optical transmission/switching and high flexibility of electronic control/processing. Channel scheduling algorithm is one of the key issues related to OBS networks. In this paper, a class-based scheduling algorithm is presented with emphasis on fairly utilizing the bandwidth among different services. A maximum reserved channel numbers and a maximum channel search times is introduced for each service based on its class of services, load and available bandwidth resource in the class-based scheduling algorithm. The performance of the scheduling algorithm is studied in detail by simulation. The results show that the scheduling algorithm can allocate the bandwidth more fairly among different services and the total burst loss ratio under high throughput can be lowered with acceptable expense on delay performance of services with lower delay requirement. Problems related with burst loss ratio and the delay requirement of different services can be well solved simultaneously.

The performance of an optical burst switching network is determined by the comprehensive effect of various parameters. This paper presents analysis results of an all-optical burst switching network which consists of edge nodes, which are responsible for traffic aggregation and distribution, and core nodes, which transmit payload data burst in the optical domain. Some essential parameters are compared on their effects on implementing switching operation, contention resolution and routing functions. We have examined the interrelationship of these parameters. Typical performance criteria are investigated in a sample OBS network.

Optical Packet Switching (OPS) has been envisioned as a prominent future optical networking technology for data-centric IP over Wavelength Division Multiplexing (WDM) networks, or optical Internet. Such OPS technology however raises significant transport and Quality of Service (QoS) challenges due to technological limitations. To circumvent OPS limitations, we have proposed a new Optical Coarse Packet Switching (OCPS) paradigm, which uses in-band-controlled per-burst switching and advocates traffic control enforcement to achieve high bandwidth utilization and Quality-of-Service (QoS). Based on OCPS, we have constructed an experimental IP-over-WDM network, referred to as OPSINET. OPSINET consists of two major types of nodes- edge routers, and Optical Label Switched Routers (OLSRs). In this paper, we first introduce the OCPS paradigm. We then present the architecture of OPSINET, describe the in-band header/payload modulation technique, and detail the operations of the edge routers and OLSRs.

Optical buffers are critical for low packet-loss probability in future photonic packet-switched networks. In particular, they would be required to store packets during rate conversion and header processing, and to overcome the receiver's bottleneck. They would be required for queuing packets while transmitters await access to the network. In this paper, we present a novel structure of optical buffer with compact size. This kind of optical buffer is based on a collinear 3x3 fiber coupler in which three fibers are completely in the same plane and weakly coupled. A SOA is used as its nonlinear element as well as an amplifier in it.The experiment result will be also given in the paper. Storage results obtained with this novel structure optical buffer at 100Mb/s will be presented first and then its capacity is extended to higher data rates of 2.5Gb/s, more compatible with present optical networks. Storage has been observed for time up to 1.568ms(more than 32 circulations) in both cases without obvious degration. The novel structure of optical buffer could be a more compact device which makes it possible to be integrated in a chip. SOA in the buffer is used as a nonlinear element as well as an amplifier to compensate loss in the buffer loop. The buffer needs low control power for switch operation. It is easy to control 'write' and 'erase' operation because the same TOAD switch in the buffer can be used for both 'write' and 'erase' operation.

OBS is more flexible and has higher bandwidth utilization than OCS; on the other hand, it is more feasible than OPS. OBS is the ideal switching technology at present. In this paper, we study OBS ring network with tunable transceivers. To reduce the high cost per node of an OBS ring network, a possible solution is to use tunable transceivers at the nodes. Efficient tuning strategies of these transceivers are required to maximize the ring bandwidth capacity and mitigate the effect of tuning latency. We propose a transmitter tuning strategy and a receiver tuning strategy and simulate the proposed tuning strategies and present the result and analyze their performances in OBS ring network.

The current routing and wavelength assignment (RWA) algorithms in optical switching (OBS) networks usually adopt the shortest path between the source-destination pairs as the routes and assign wavelengths hop-by-hop. There are two main problems exist in theses algorithms: (1) If there are common links among the shortest paths of different source-destination pairs, the one-way reservation protocol may cause congestion on these links while other links are underutilized, which may deteriorate the network performance, especially in an unsymmetrical network. (2) Few RWA algorithms take the fault recovery into consideration, which is important for the network to operate smoothly. An ant system based RWA algorithm is proposed in this paper to resolve these two problems. The destination nodes send ACKs back for each successfully received burst control packet (BCP). The ACKs are feed back along the same path as the one through which BCPs are forwarded. ACKs leave some "pheromone" along the path. The coming bursts will choose the output links with the probability proportioned to their pheromone intensity. Numerical results obtained from simulation show that our RWA algorithm can find the optimal routes adaptively and get a better burst drop probability performance compared with current RWA algorithms in an unsymmetrical network. Furthermore, our RWA algorithm is robust for fault recovery. When there are failures on some fibers, the bursts can be dynamically deflected to a suitable route without any extra information exchange among the switching nodes.

Deflection routing is one of contention resolution of optical burst switching networks. But the offset time maybe not enough for reserve resource if deflection routing adopted. Too much deflection adopted will deteriorate the network performance, another issue is how to determine if a contending burst will be deflected or discarded. In this paper, little Fiber Delay Lines (FDL) is used to assure the offset time will be compensated in time, and an optimum scheme is proposed from three aspects as if the network situation permits deflection, if contending burst is worthy to be hold continuously and the impaction of alternative route on deflection. Numerical results show that our optimized deflection scheme can achieve not only preferred deflection, but also to keep wavelength link from overloading. It balances the network load and stabilizes the network performance some degree.

Optical Burst Switching (OBS) is a new paradigm for future all-optical networks. It has been noted that performance of an OBS node depends on the wavelength assignment algorithm that is used. In this paper we present a new class of wavelength allocation algorithms called cost-based algorithms. We note that bursts compete for more types of resources than wavelengths alone. For example if a given burst is to be allocated successfully, a Fiber Delay Line (FDL) or a wavelength converter may have to be used. It can be expected, however, that the set of available resources will be limited. If at a given time all the converters are used, then any arriving burst will have to be allocated on the same wavelength - if it is available. Similarly, the unavailability of FDLs will decrease the probability of a burst being accepted. In a cost-based algorithm, each resource is assigned a metric (or price). Channels are priced according to their suitability for a particular burst. When a control packet arrives at a core node, all the possible ways of handling the corresponding burst are found (the outgoing channel, with or without a FDL or wavelength converter), and the one with a lowest metric is chosen. To show how the performance of a cost-based algorithm compares to other algorithms we present the results of our simulations for a node with full conversion capability and a shared FDL.

This paper presents a quality of service (QoS) enable optical delay line (ODL) architecture to solve the problem of resource contention and support multilevel priority queues in an optical packet switch. ODL has been used in optical packet switches to resolve resource contention; however, the packets travel continuously in ODL limits the management of random access of the packets and increases the packet loss probability. Moreover, multiple ODL sets usually are needed to realize multiple priority queues in order to support QoS. In this paper, a new Unicast Recirculatiion ODL (URODL) architecture is proposed to resolve the output contention problem in an input-queued optical packet switch. To improve relatively poor throughput due to the head of Line (HOL) blocking in the input-queued switch, we modify URODL to support a more efficient window-based lookahead scheduling algorithm. Furthermore, a control strategy is designed to turn a single set of URODL into multiple logical queues to hold different priority packets. The simulation results show our URODL model reduces packet loss effectively with the capability to support QoS. This URODL model can be easily implemented and managed in a fast optical packet switch.

The National Aeronautics Administration of China is planning lunar exploration. The National Key Laboratory of Tunable Laser Technology of Harbin Institute of Technology is responding to this challenge by developing a optical communication system with pulse-position modulation signaling and APD Receivers. So in this article, we mainly calculate the bit-error rate (BER) of detector uncoded pulse-position modulation signals using a Webb-Gaussian modeled avalanche photodiode (APD) receiver in the presence of additive Gaussian thermal noise for the lunar-earth laser link. Performance curves for 4-ary pulse-position modulation obtained through Monte Carlo simulation as well as a numerical calculation result are presented. The result indicates that the number of background photons per slot must be restrained below 20 in order to achieve 10-6 BER when the number of signal photons per signal slot is 120.At the same time, the pointing error is the important factor that influences the BER .It must be restrained under 15urad when the signal optical beam width is 100urad in order to achieve 10-6 BER. And we find that optimized the APD gain can improve the performance of the optical communication system. The number of background photons per slot can be more than 20 when BER need to be the order of 10-6 if the number of signal photons per signal slot is 120.If we take the effect of atmosphere on the optical pulse shape into account, the BER will rise. In order to improve BER, we present the analysis of return-zero (RZ) pulse position modulation scheme for the optical communications with non-rectangular pulses. The results show that RZ PPM will restore performance losses due to reduced peak intensity during the detection process.

In order to overcome fading in radio-optical communication channel, such as attenuation and optical intension twinkle, this paper introduces Turbo codes to improve the performance, although the results have a larger gap to Shannon's limit remains, the advantageous of using Turbo codes in a radio-optical communication environment is obvious and it can be put into practical use. In this paper we introduce a modified BCJR algorithm to adapt to the radio-optical channel.

We review recent progress on optical switches that employ silica-based planar lightwave circuits (PLC) including a 16 x 16 matrix switch, a 1 x 128 switch, an 8-channel 1 x 8 switch array, and a 32-channel 2 x 2 switch array. These switches are composed of silica-based waveguides and have a solid-state structure with no moving parts, thus they exhibit low insertion loss, high repeatability, and excellent long-term reliability. We also describe IC-on-PLC technology, where IC chips are integrated on the PLC surface for higher functionality and greater compactness. We report the results of reliability tests on the PLC switches, and show that they can put to practical use.

Mach-Zehnder interferometer (MZI) 2×2 thermo-optic switches, based on parabolically taped MMI couplers, have been designed in this paper. The theory of the parabolically tapered MMI coupler is studied to estimate the optimal length, and then the effective index method (EIM) and the beam propagation method (BMP) are utilized to simulate these devices to find out the optimal length. The values of waveguide parameters are set as: the refractive index difference is chosen to be 0.75%, the width of the MMI field is 30 micron, the distance between the MZI arms is 100 micron and the separation between the access (input/output) waveguides is 250 micron. The simulated results show that the performances of the tapered coupler is better, whose the second width is equal to 25 micron. MMI coupler is also simulated by BPM. Although the tapered coupler can not offer performances as well as those of other MMI coupler, it can offer compact devices. The switches, based on conventional MMI couplers and tapered couplers, are simulated at the same of access waveguide separation and MMI width. Simulated results show that although the performances of this new structure are not as good as those of the switches based on conventional MMI couplers, but can reduce the length of the device effectively. This compact characteristic is important for integrated planar optical circuit.

A microphotonic switch with three input waveguides and two output waveguides and integrated with an optical power splitter has been proposed. It is fabricated on the multimode interference principle in Si-based SiGe material system and configured for a 3x2 symmetrical structure of the three input waveguides and the two output waveguides of the device. The central waveguide section is based on a multimode interference and incorporated with an activated carrier injection element. The operating wavelengths of the device are specially designed for 1545, 1550, and 1555 nm conventional-band wavelengths. The measured crosstalk is at around -17 dB and the average insertion loss is about 2.3 dB. At switch-ON state, the measured injection current is 370 mA corresponding to an injection current density of 950 A/cm².

Thin film filter (TF)-based 1x2 wavelength selective switch (WSS) architectures are introduced. Our key idea is to locate a movable mirror orientated at a desired angle close to the TF to switch the desired wavelength optical beams to the wanted switch ports. Our first proposed WSS is in the transmissive mode where the surfaces of the TF and the movable mirror are parallel to each other and it provides a moderate optical isolation. Another WSS structure is in reflective configuration in which the movable mirror is tilted with respect to the surface of the TF and when combined with the optical circulator leads to a very low optical coherent crosstalk. Our experiment using a commercially available TF and a movable mirror shows that our transmissive-mode WSS provides a -18.87 dB optical coherent crosstalk while a much improved < -53 dB optical coherent crosstalk can be obtained between the two switching ports in our reflective-mode WSS structure. Our reflective 1x2 WSS also gives a higher optical loss due to the use of an optical circulator. Low polarization dependent loss of < 0.1 dB is determined for both WSS structures.

In this paper, a kind of tunable wavelength selective optical switch was proposed with two-input/two-output fiber ports. It is based on tunable Fabry-Perot cavity by a pair of multi-layered piezoelectric ceramics. Each fiber carries N wavelengths, one of which can be selected. The tunable span can reach 5.43 nm under 10 V DC voltages. The relation of wavelength tuning ability and driving voltage is linear. The maximum of difference between theoretical and experimental results is less than 0.08nm. The quantities of maximum insertion loss, switching time and on/off ratio are about 3 dB, 1 ms and 28 dB

In this paper, the propagation properties of the electromagnetic (EM) waves in one-dimension photonic crystals (PC) with nonlinear defects layer closed by periodic layered structures are reported. It is shown that in the presence of the nonlinearity, the transmission properties are strongly modulated by both the frequency and intensity, and the system exhibits bistability and multistability. Moreover, we find that both the switching-up and switching-down intensities of the bistable response can be made very low, when the intensity of the incident wave matches the defect mode of the structure. By use of the nonlinear finite-difference time-domain (NFDTD) method, we simulate the bistable threshold of the nonlinear photonic crystals with the kerr medium and the feasibility of making a switch utilizing such a structure. Bistable swithching with a low threshold intensity of 0.0007 kW/cm² is obtained. The numerical experiments describenoptimal bistable switching in a nonlinear photonic crystal system. A photonic crystal enables the device to operate innsingle-mode fashion, as if it were effectively one-dimensional. This provides optimal control over the input and outputnand facilitates further large-scale optical integration.

We present an analytical model and numerical experiments to describe optimal bistable switching involving photonic crystals nonlinear cavity. The involving photonic crystals nonlinear cavity is particularly suitable for large-scale all-optical integration. It is interesting to show that the refractive index of the system can be altered by the incident intensity when the incident frequency is tuned to the evanescent mode and the cavity contains a nonlinear Kerr medium. The response of this system to high-intensity incident waves is studied in detail. A cavity host a localized mode at 1.55 um with a Q of 4000 and a mode volume of 0.055 um3, we predict the onset of bistable reflection at incident powers of ~40 mW. The downstream reflections lead to hysteresis loops in the reflectivity that are topologically distinct from conventional Lorentzian-derived loops characteristic of isolated Fabry-Perot cavities. We provide a stability argument that reveals the unstable branches of these unique hysteresis loops, and we illustrate some of the rich bistable behaviors that can be engineered with such downstream sources, this fact may be used to obtain more flexibility in designing nonlinear devices.

The PMD at a wide range of optical frequencies and statistics of all-order PMD, especially the first-, second- and third-order PMD, of a n-segment DGD units combination separated by PCs or PRs are fully analysed with Poincare sphere, Mueller matrix and Fourier transform methods. The results show that, to simulate all-order PMD in DWDM systems as truly as possible, the number of n should be as large as possible, the least value of 4 is preferred. The DGD units should be variable or unequal. Choosing PC or PR just affects the probability density function of PMD when n is not too much and the DGD values of the DGD units are unchangeable. At last, a simple, practical, and low cost all-order PMD emulator in DWDM systems is proposed.

Optical switches made with Micro-electro-mechanical Systems (MEMS) technology bear advantages of both MEMS and traditional optics. It is of very importance to control position misalignments and angular misalignments, which have heavy effects on insertion loss of the MEMS optical switches. In this paper, a 4×4 MEMS optical matrix switches is proposed, in which on-off micro-mirrors are arranged with Crossbar switching network. Ball-lensed single-mode optical fiber collimators are used as input and output ports. With overlap integral calculus theory of Gaussian beams coupling, the insertion loss of the optical matrix switches, and also its main origins, i.e. the position misalignments and the angular misalignments, are calculated and analyzed. Some useful conclusions are drawn to guide designing such devices. As for the misalignments tolerances: 2μm of input/output lateral positioning misalignment, 0.15°for input/output angular misalignment, and 0.15°for micro-mirror angular misalignment, the insertion loss of the pilot study are 2.77dB for the highest value.

A novel micro high-speed 2x2 magneto-optic switch and its optical route, which is used in high-speed all-optical communication network, is designed and analyzed in this paper. The study of micro high-speed magneto-optic switch mainly involves the optical route and high-speed control technique design. The optical route design covers optical route design of polarization in optical switch, the performance analysis and material selection of magneto-optic crystal and magnetic path design in Faraday rotator. The research of high-speed control technique involves the study of nanosecond pulse generator, high-speed magnetic field and its control technique etc. High-speed current transients from nanosecond pulse generator are used to switch the magnetization of the magneto-optic crystal, which propagates a 1550nm optical beam. The optical route design schemes and electronic circuits of high-speed control technique are both simulated on computer and test by the experiments respectively. The experiment results state that the nanosecond pulse generator can output the pulse with rising edge time 3~35ns, voltage amplitude 10~90V and pulse width 10~100ns. Under the control of CPU singlechip, the optical beam can be stably switched and the switching time is less than 1μs currently.

The performance of normalized throughput in WDM+OCDMA system is studied. When all simultaneous users are equally allocated to different wavelength channels, upper bound of normalized throughput is obtained. When all simultaneous users are first allocated to same wavelength channels, lower bound of normalized throughput is obtained. However, if there is no central control to allocate wavelength channels equally, WDM+OCDMA system has the in-between performance of normalized throughput.

Dense wavelength division multiplexing (DWDM) system is the tendency of optical fiber communication systems because of its high speeds and capacities. DWDM systems with channel data rates of 40 Gbit/s bring us both advantages and challenges. With broader spectrum, the signal suffers more from chromatic dispersion, optical fiber nonlinear effects and polarization mode dispersion (PMD). Simultaneously, the combination of PMD and nonlinear effects results in more complexity. In this paper, amended nonlinear Schrodinger equations, which include group velocity dispersion (GVD), third order dispersion (TOD), self-phase modulation (SPM), cross-phase modulation (XPM), four-wave mixing (FWM) and PMD synthetically, are derived, then the system degradation induced by PMD and nonlinear effects is investigated via numerical simulation. The results show that when the power of input signal is low, PMD has more affection on system comparing with nonlinear effects, and nonlinear effects become dominating with power increasing, but mild PMD may mitigate their impact to some extent. The results in this paper are valid for design and analysis of long haul DWDM systems with high bit rate.

As the key apparatus, the All optical Wavelength Converter (AOWC) will play an important role in future optical communication and optical signal processing system. In this paper, switching characteristics of wavelength converter based on saturable absorber in semiconductor lasers will be researched. This kind of conversion mechanism possesses some advantage, such as simple structure, low cost, high stability and so on. This paper is organized as follows: Firstly, utilizing rate equations, a new theoretical model on wavelength conversion based on saturable absorber is put forward. Nextly, the frequency modulation response of wavelength conversion will be discussed under the small-signal analysis based on the theoretical model. Lastly, Numerical value solution results will be given out when external signal light injects in saturable absorber region of semiconductor lasers. The characteristics of wavelength conversion are simulated in different optical parameters including the injection current, the input signal optical power and bit rate. Those results are useful to realization and the optimal design of the wavelength converter based on saturable absorber.

As the key apparatus, the All optical Wavelength Converter (AOWC) will play an important role in future optical communication system. In this paper, a theoretical model based on mutual coupling effect of laser modes for all-optical wavelength converter in single-model homogeneously broaden laser is presented. With the theory the paper analyze the stability of single-model homogeneously broaden lasers and the influence on its characteristics when there is parameters variety. Numerical value of saw tooth wave is given also. Those results are useful to realization and the optimal design of the wavelength converter.

The speed of high-performance switches and routers is often limited by the bandwidth of commercially available memories. Meanwhile, the rapid growth in network bandwidth accompanied by the slowly increasing memory speed makes the problem even harder over time. There are, in fact, several techniques to build faster memories. However, some are based on the ideas from computer systems, such as parallelism, interleaving and banking, which can hardly be applied directly to packet buffering, while the others like hybrid SRAM-DRAM packet buffers are restricted by the speed of SRAM and inapplicable as the link rate exceeds the speed of SRAM. Motivated by increasing the throughput of packet buffers with only common memory arrays, we present one particular packet buffer architecture called Tri-Stage Memory Array (TSMA) that can speed up the packet buffering and retrieving processes to an arbitrary high speed theoretically. To replenish TSMA, a memory management algorithm called Most Urgent Queue First (MUQF) is also described and analyzed. It is proved that TSMA architecture coupled with MUQF algorithm can guarantee a bounded delay for each packet under any traffic arrival pattern or scheduling algorithm. Moreover, we provide an alternative architecture of TSMA to achieve simple implementation.

The burst contention has a strong impact on the performance of optical burst switching (OBS) network, and currently, various contention resolution schemes have been proposed from time domain, space domain and wavelength domain, respectively. Among the schemes, the deflection routing is more attractive, which can loosen the requirement of optical buffer in capability and quantity. With deflection routing, the contending burst is rerouted to an available port other than the intended output instead of being dropped. However, when the traffic load is high, there should not be only more contentions but also frequent deflection routings, some of which might aggravate the network status. To improve the status, this paper proposes the tunable-parameter deflection routing (TPDR) algorithm, which can control the deflecting operation of a contending burst by a tunable-parameter, i.e., deflection probability, which is settled by the priority and traffic load of contending burst, and search the optimum deflection path in terms of the burst loss probability and the path length. The simulation results show that it can improve the overall BLP and the BLP of each individual priority, and alleviate the negative effect of the offset-time deficit on QoS guarantee.

A novel optical add-drop multiplexer (OADM) based on Mach-Zehnder interferometer and fiber Bragg grating (FBG) is proposed. In the structure, the Mach-Zehnder interferometer acts as an optical switch. The OADM can add/drop one of the multi input channels or pass the channel directly by adjusting the difference of the two arms of the interferometer. The channel isolation is more than 20dB.

In this paper, the optimization algorithms are introduced in adaptive PMD compensation in 10Gb/s optical communication system. The PMD monitoring technique based on degree of polarization (DOP) is adopted. DOP can be a good indicator of PMD with monotonically deceasing of DOP as differential group delay (DGD) increasing. In order to use DOP as PMD monitoring feedback signal, it is required to emulate the state of DGD in the transmission circuitry. A PMD emulator is designed. A polarization controller (PC) is used in fiber multiplexer to adjust the polarization state of optical signal, and at the output of the fiber multiplexer a polarizer is used. After the feedback signal reach the control computer, the optimization program run to search the global optimization spot and through the PC to control the PMD. Several popular modern nonlinear optimization algorithms (Tabu Search, Simulated Annealing, Genetic Algorithm, Artificial Neural Networks, Ant Colony Optimization etc.) are discussed and the comparisons among them are made to choose the best optimization algorithm. Every algorithm has its advantage and disadvantage, but in this circs the Genetic Algorithm (GA) may be the best. It eliminates the worsen spots constantly and lets them have no chance to enter the circulation. So it has the quicker convergence velocity and less time. The PMD can be compensated in very few steps by using this algorithm. As a result, the maximum compensation ability of the one-stage PMD and two-stage PMD can be made in very short time, and the dynamic compensation time is no more than 10ms.

Dynamic programming, the ordinary adaptive compensation in the operational research, is used to resolve extremum of functions under the constraint condition. In this paper, it is introduced that, fundamentals, methods and steps about the first-order PMD compensation by dynamic programming. The result shows that, for the first-order PMD compensation, dynamic programming is used to carry out optimized design, and the result is satisfied. Beginning with recursive relation of PMD vectors in the fiber and the compensation devices, mathematical model of the first-order PMD compensation is established. Through optimized algorithm using dynamic programming, simulation and experiment for PMD adaptive compensation are both implemented. Some optimized algorithms once used in the PMD adaptive compensation have slow speed to approximate the optimal value and tends to become the local optimal solution; but if optimized algorithm using dynamic programming can approximate the global optimal solution directly. Therefore, it has the advantage of fast optimized speed. On this basis, the image of optimal solution is given and analyzed. PMD compensation scheme, based on PMD compensation vector, is proposed. The algorithm principle of adjusting control and the direction of improvement are also put forward. These results are benefit for dynamic adaptive compensation of PMD.

In the paper, a new optical packet switch is proposed, which uses shared WDM buffers (fiber delay-lines) and shared tunable wavelength converters (TWCs) as contention resolution. In each fiber delay-lines, many wavelengths can be used to buffer packets at the same time. The shared TWCs have two kinds of functions, either as translators for free wavelengths in destination output fibers, or as translators for free wavelengths in the feedback FDLs. Based on this architecture, a lower packet loss rate can be achieved without the need of deploying a large number of FDLs and TWCs. The performance of the new architecture has been extensively studied by means of simulation experiments.

We present a scheme to realize entangled states, swap gate and quantum-information transfer for two atoms in cavity QED. The scheme does not require the transfer of quantum information between the atoms and cavity and that the cavity field is only virtually excited, thus the requirement on the quality factor of the cavity is greatly loosened.

This paper describes the design process and performance of the optimized parallel optical transmission module. Based on 1×12 VCSEL (Vertical Cavity Surface Emitting Laser) array, we designed and fabricated the high speed parallel optical modules. Our parallel optical module contains a 1×12 VCSEL array, a 12 channel CMOS laser driver circuit, a high speed PCB (Printed Circuit Board), a MT fiber connector and a packaging housing. The L-I-V characteristics of the 850nm VCSEL was measured at the operating current 8mA, 3dB frequency bandwidth more than 3GHz and the optical output 1mW. The transmission rate of all 12 channels is 30Gbit/s, with a single channel 2.5Gbit/s. By adopting the integration of the 1×12 VCSEL array and the driver array, we make a high speed PCB (Printed Circuit Board) to provide the optoelectronic chip with the operating voltage and high speed signals current. The LVDS (Low-Voltage Differential Signals) was set as the input signal to achieve better high frequency performance. The active coupling was adopted with a MT connector (8° slant fiber array). We used the Small Form Factor Pluggable (SFP) packaging. With the edge connector, the module could be inserted into the system dispense with bonding process.

This paper proposes a novel and innovative scheme for 10Gb/s parallel Very Short Reach (VSR) optical communication system. The optimized scheme properly manages the SDH/SONET redundant bytes and adjusts the position of error detecting bytes and error correction bytes. Compared with the OIF-VSR4-01.0 proposal, the scheme has a coding process module. The SDH/SONET frames in transmission direction are disposed as follows: (1) The Framer-Serdes Interface (FSI) gets 16×622.08Mb/s STM-64 frame. (2) The STM-64 frame is byte-wise stripped across 12 channels, all channels are data channels. During this process, the parity bytes and CRC bytes are generated in the similar way as OIF-VSR4-01.0 and stored in the code process module. (3) The code process module will regularly convey the additional parity bytes and CRC bytes to all 12 data channels. (4) After the 8B/10B coding, the 12 channels is transmitted to the parallel VCSEL array. The receive process approximately in reverse order of transmission process. By applying this scheme to 10Gb/s VSR system, the frame size in VSR system is reduced from 15552×12 bytes to 14040×12 bytes, the system redundancy is reduced obviously.

Optical Burst Switching (OBS) is considered as an efficient switching technique for building the next generation optical Internet. An offset-time-based scheme has recently been proposed in order to provide quality-of-service (QoS) in OBS networks. Unfortunately, the proposed service differentiation has several problems. In this paper, we propose a burst-length-based QoS scheme cooperate with Void-Filling, which is less complex and more valuable than existing offset-time-based scheme. Simulations are conducted to evaluate the performance of our burst-length-based scheme provisioning within OBS networks in terms of burst loss probability.

A new scheduling algorithm, which aims to provide proportional and controllable QoS in terms of burst loss probability for OBS (optical burst switching) networks, is proposed on the basis of a summary of current QoS schemes in OBS. With simulations, performance analyses and comparisons are studied in detail. The results show that, in the proposed scheme, burst loss probabilities are proportional to the given factors and the control of QoS performance can be achieved with better performance. This scheme will be beneficial to the OBS network management and the tariff policy making.

Optical burst switching (OBS) is one of the most important switching paradigms in the future optical WDM networks and Internet. OBS has better bandwidth utilization than OPS and has faster switching time than OCS. OBS is a promising switching technology in next generation optical internet. The issue of survivability has been among the most important research in designing an optical network. In this paper, we deliver optical protection & restoration and MPLS protection & restoration. Then we present a survivable OBS network using a multiple-ring approach.

Two schemes of linear combinatorial code (LCC) coding scheme with different and multiple wavelength bandwidth were proposed using fiber grating cascades in series and parallel, respectively. The system model and probability density function are investigated. Numerical simulation confirmed that bit error ratio (BER) performance largely improved due to that of sub-codeword system. Moreover, the preference of LCC with different wavelength bandwidth is better than that of multiple wavelength bandwidth.

An 8-channel switching node with routing function is designed in this paper using a high-performance digital cross-point switch. The throughput of the 8-channel routing node is 10 Gbit/s. Combined with time division multiplexing (TDM) technique, a high-speed optical interconnection network was designed. Using Optical Network Interface Cards (ONIC) and the designed routing node, a two-layer optical interconnection network with structure of ring and star topology is designed. Based on circuits cascading technique, the network can be expanded to meet the request of large scale data communication with low communication latency.

N-bit programmable photonic time delay structures are proposed for the compensation of polarization mode dispersion (PMD). Our key design idea is to have one switchable differential group delay (DGD) module, that specially contains the zero DGD and the minimum DGD resolution step of t, connected to the rest of the variable photonic DGD bit modules that offer DGD values of (+/-)2t to (+/-)[2^N-2]t. Our proposed variable photonic DGD architectures use 2x2 polarization maintaining optical switches or 90 degree polarization rotators to choose the desired DGDs. When all variable photonic DGD bit modules are attached to the polarization rotator, 2^N DGD values ranging from -{[2^N-1]-1}t to [2^N-1]t are obtained. Our simulation study points out that compact programmable photonic DGD modules can be achieved by using materials with strong birefringence and our reflective configuration. In addition, materials with low wavelength sensitivity are preferred for use in broadband wavelength operation. Mirror-based programmable photonic DGD structures are also introduced.

We investigate the performance of a differential amplitude pulse-position modulation with return-to-zero coding (DAPPM-RZ) over an indoor optical wireless channel. We compare the performance of DAPPM-RZ(A=2) with DAPPM(A=2), DPPM and DH-PIM. The result shows that, over a non-dispersive channel, DAPPM-RZ yields better power efficiency than DAPPM. It requires about 1.5 dB less transmit power. However, the bandwidth of DAPPM-RZ is double that of DAPPM. Compared to DPPM, the bandwidth of DAPPM-RZ is about the same as that of DPPM but DAPPM-RZ yields less power efficiency. When the number of bits/symbol(M) is above 3, the DAPPM-RZ is superior to DH-PIM_2 in terms of power efficiency but has less bandwidth efficiency. Over a dispersive channel, given the same value of M, DAPPM-RZ outperforms DPPM, DAPPM (without RZ) and DH-PIM₂ when the normalized rms delay spread is high.

In this paper, an inner wavelength method (IWM) is proposed to enlarge buffering capacity of shared FDL-buffers. In addition, an optical packet switch called Extended shared buffer type optical packet switch (Extended SB-OPS) is proposed to realize the inner wavelength method. In order to further improve performance of Extended SB-OPS, a greedy algorithm based on inner wavelength method (GA-IWM) is introduced. The performance of Extended SB-OPS is evaluated by simulation experiments.

Optical performance monitoring is a very import issue in the optical transparent network. We present an optical performance monitoring method, based on asynchronous sampling technology, and Q value and bit error rate can be calculated by the asynchronous histogram. Experiment result shows that the optical performance method is bit rate transparent and modulate format transparent.

In this paper, a novel method to reproduce the maximum DGD is proposed. In the model different PMD states are obtained by randomly varying the polarization coupling angle between sections. By confining the coupling angles' distribution to a shrinking range, the statistics of DGD will be distorted and the low probability events will be efficiently produced. The relationship between the mean DGD of changed DGD distribution and the angle range is given to make it convenient to choose a suitable angle range at a given maximum DGD and simulating sections.

We have developed a GMPLS-based software platform with Intelligence Optical Network (ION) capabilities to simulate different technologies and protocols of ASTN. For optical transport network (OTN) based ASTN, usually named as automatically switched optical network (ASON), two signalling schemes, Explicit Label Control (ELC) and Backward Wavelength Reservation with N retries (BWR-N), are considered. The ELC is associated with two routing and wavelength assignment (RWA) policies, the shortest path first routing (Dijkstra algorithm) with first-fit (FF) wavelength assignment and the fixed path least congestion (FPLC) routing with FF wavelength assignment. For the BWR-N, the route is determined by Dijkstra's algorithm and the wavelength is selected through signalling process. For SDH/SONET-based ASTN, we also consider two signalling scheme, the ELC and the Forward Timeslot Reservation (FTR). Similar simulations have been performed. These signalling and routing schemes are evaluated by several crucial metrics, such as the blocking probability, the setup time, the average routing length of the connections and the control overhead. Some significant conclusions have been presented by this paper.

Utilizing optical technologies for the design of packet switches and routers offers several advantages in terms of scal-ability, high bandwidth, power consumption, and cost. However, the configuration delays of optical crossbars are much longer than that of electronic counterpart, which makes the conventional slot-by-slot scheduling methods no longer be-ing the feasible solutions. Therefore, there must find some tradeoff between the empty time slots and configuration overhead. This paper classifies such scheduling problems into preemptive and non-preemptive scenarios, each has its own advantages and disadvantages, and decides whether the traffic in one input queue should be completely transmitted in one switch configuration. Although non-preemptive scheduling is inherently not good at achieving above-mentioned tradeoff, it is shown, however, that the proposed maximum weight matching (MWM) based greedy algorithm is guaranteed to achieve an approximation 2 for arbitrary configuration delay, and with a relatively low time complexity O(N2). For preemptive scheduling, a novel 2-approximation heuristic is presented. Each time in finding a switch configuration, the 2-approximation heuristic guarantees the covering cost of the remaining traffic matrix to have 2-approximation. Simulation results demonstrate that 2-approximation heuristic (1) performs close to the optimal scheduling; (2) outper-forms ADJUST[1] and DOUBLE[3] in terms of traffic transmission delay and time complexity.

Ant colony algorithm (ACA) is a novel simulated evolutionary algorithm. It is a population-based approach, which allows positive feedback to be used as the primary search mechanism and provides a new method for complicated combinatorial optimization problems. In this paper, it is used to optimize restoration routing for WDM optical networks. It is improved in three parts of selection strategy, local search, and information modification and it can process the problem of optimal restoration route search for different purpose in the cases of various failure conditions. Through the numerical results of practical networks: CHINANET, the practicability has been proved.

Scalability performance of optical node is an important issue in the design of optical cross-connect (OXC) architecture. A practical architecture is expected to be scalable easily with low cost. However, when OXC employed with wavelength converter (WC) to support virtual wavelength path (VWP), the OXC scalability degrade due to the limit number of the WCs and partial sharing characteristic. In this paper, we proposed a new scheme to expand a kind of limit wavelength conversion OXC, share-per-wavelength architecture, which shares the WC within the same wavelength. This scalable scheme is analyzed in detail through comparing various wavelength grouping methods. Furthermore, their blocking performance is studied by using discrete event simulation. The results of the simulation show that our scheme has comparable blocking performance to the architecture without scalability. The small improvement at high initial occupation is because small wavelength grouping lead to a little better sharing efficiency under high load. But when the group becomes too small, for example less than the number of WCs, the repeated WCs will reduce its performance.

Polarization effects such as polarization mode dispersion (PMD) and polarization dependent loss (PDL) have become the critical limiting factors for long haul high bit rate optical fiber transmission. This paper use waveplate models to simulate the statistical performance of optical link with PMD and PDL in 40Gbit/s optical system. Three DPSK modulation formats are compared each other for their tolerance against PMD and PDL. The 33% RZ-DPSK is superior to the other two DPSK formats, the 50% RZ-DPSK and the CSRZ-DPSK whose duty cycle is 67% when only PMD is considered. And in case of PDL, the outages for three DPSK formats do not show obvious changes with the increasing of PDL. Furthermore, the performance of two RZ-DPSK formats, whose duty cycles are 33% and 50% respectively, is superior to that of CSRZ-DPSK. This shows that modulation format with narrower pulse width owns more high tolerance against PMD and PDL, therefore the CSRZ-DPSK format shows poorer performance when it is compared with the other two DPSK formats.

Real-Time Oscilloscopes that are currently available go up only to 8 GHz. For transient signal measurements above 8 GHz, streak cameras have to be used. An instrument, developed by YY Labs, provides a convenient, low-cost method for such measurements. YY Labs' Single-Shot Scope captures single-shot signals in one or two channels and then regenerates 1000 copies of the one- or two-channel analogue signals to form a pulse train, with the aid of an optic-fiber delay line loop. A sampling scope recovers the original signals from the pulse train.

Frame clock is useful for packet processing such as header detection and payload demultiplexing. A novel all-optical frame clock recovery scheme based on "intensity reshaper" and mode-locked semiconductor fiber ring laser is demonstrated. The "intensity reshaper" including a polarization controller and a polarizer is the key element to realize frame clock recovery from equal-amplitude even-multiplexed OTDM signals. In theory, a mathematical expression is given to analyze the intensity of harmonic of clock-frequency component. The relative intensity of each clock-frequency component will change with the alterative angle caused by adjusting the PC in the "intensity reshaper", so the desirable clock-frequency component can be enhanced, which is helpful for clock recovery. Moreover, the intensity of harmonic of clock-frequency component is also related to the pulse amplitude, width and period in the multiplexed data. In experiment, 2.5GHz frame clock is extracted from even-multiplexed 4x2.5GHz and 8x2.5GHz OTDM signals respectively. At the same time, bit clock is also recovered by using this scheme. The extracted clock pulses have several desirable features such as low timing jitter, broad wavelength tuning range and polarization independence. This scheme simplifies signal generation and propagation in OTDM systems, which can be applied to clock recovery in high-speed OTDM network.

A novel full optical time division multiplexer (OTDM) based on fiber acousto-optic coupler (FAOC) with loop configuration is presented. FAOC with loop configuration is analyzed theoretically, and a 8X2.5 Gb/s OTDM which utilizes FAOC with loop configuration concatenated is investigated. Due to the tunable characteristic between coupled power ratio and operational wavelength of fiber acousto-optic coupler, OTDM in the paper can lower requirements for high performances of devices with respect to the common OTDM methods. Finally, it is pointed that OTDM using several RF drivers can generate simultaneously high speed optical pulses of multi-wavelength. The configuration is feasible in terms of analyzing and simulating theoretically.

Optical wireless communication (OWC) systems use the atmosphere as a propagation medium. In optical wireless communication links, atmospheric turbulence causes distortions in the laser beam wavefront, leading to fluctuations in both the intensity and the phase of the received light signal performance, i.e. signal fading. This paper presents a suppression method of turbulence noise on the basis of analyzing the optical signal effect of turbulence.In this paper, according to probability distributions of turbulence-induced intensity, we describe a BEP model. From the model, we develop an adaptive threshold detector technique to decrease the BEP. The BEP increases from 10^-11 to 10^-7 for a linear increase of normalized standard deviation σ_x of the log-amplitude from 0 to 0.5 which is shown in fig.1 at a stationary threshold. Finally, we describe the matched receiver to maximize the output signal-to-noise ratio (SNR). We also numerically compute the average bit-error probabilities for the filter case.

The performance of 10Gb/s or above optical single channel transmission system is mainly limited by the chromatic dispersion (CD) and polarization mode dispersion (PMD) effect, and hence, different modulation formats and equalization methods are required to increase transmission distance, and mitigate transmission impairment. In this paper, we numerically evaluate the performance of 10G bit/s optical single channel transmission systems, which combine with varied kinds of modulation formats, which includes non-return-to-zero (NRZ), return-to-zero (RZ), carrier suppressed return-to-zero (CSRZ), and two kinds of electrical signal process (ESP) techniques which includes linear equalizer (LE) and decision feedback equalizer (DFE). Numerical simulation results show that the combined use of varied modulation formats and equalizer, we can obtain complete different performance improvement.

Adiabatic transmission over 42.211 km of 10GHz, 1.5ps, ultra-short solitons in G652 standard fiber have successfully be demonstrated. The results show that optical pulses with input power of fundamental soliton can adiabatically transmit over more than one thousand of dispersion length, keep their shape unchanged, and the dispersion wave does not yet appear That is good agreement with theory of adiabatic soliton transmission,. When power of input pulses is less than fundamental soliton power, they can evolves into soliton and adiabatically transmit. Temporal width of the output soliton pulses is obviously more than that at soliton power input, but the spectral with of the output soliton pulses is very close to that at soliton power input. When power of input pulses is more than fundamental soliton power, temporal width of the output pulses is slight less than that at soliton power input, but the spectral with of the output pulses is obviously more than that at soliton power input.

In this paper, the principle of a cantilever-beam micro-mechanical switch with SiO2 optical waveguide on silicon is described, and the solid model of the switch is established and finite element method (FEM) is used in studying its dynamic characteristics. Under different parameters, the modal characteristics are simulated and analyzed. The influence of the switch structural dimension on the resonant frequency and dynamic characteristics of the switch are investigated, and the harmonic response of the micro-mechanical optical switch is achieved. The results of analysis offers reliable basis for the structure design of the micro-machined cantilever optical switch.

Transmission performances of 10Gbps DWDM NRZ, RZ and CSRZ system are compared by both theoretical simulation and experiment. Simulation shows the NRZ, RZ and CSRZ have different duty cycle, spectrum and chirp characteristics, which are the main reasons for different transmission performances. Simulation also shows that NRZ has the largest dispersion tolerance. The chirp of RZ brings out more jitter than the other two modulation formats do during transmission. CSRZ has less pulse distortion induced from dispersion due to its chirp characteristics. Our simulation is verified by experiments with a multi-channel 10Gbps ultra long haul system. The experiments of our transmission system show that both NRZ and RZ system can transmit above 3200km, and CSRZ can transmit as far as 5490km.

We present how to obtain the forward linear birefringence and the angle increment of birefringent axis distributions from the round-trip local birefringence vector by using P-OTDR technique. Further we can use these data to calculate the DGD distribution in two ways, one is to show the DGD for every fiber sections and the other is to show the DGD evolution along the fiber. The experimental results are in good agreement with the theoretical results and the result measured using JME method.

Polarization mode dispersion (PMD) and chromatic dispersion (CD) are two important limiting factors, which affect the performance of high speed fiber transmission system. The tunable PMD and CD compensation experiment on 10Gbit/s NRZ fiber transmission link is reported by using two nonlinear chirp sampling fiber gratings, which are fabricated on high-birefringence and standard single mode fibers by using uniform phase mask. The total system power penalty is less than 1dB (@BER 10^-12) after PMD and CD tunable compensation.

Nonlinear optical pulses with sub-picosecond width will exhibit the phenomena of timing jitter and pulse decay induced by the third-order dispersion and Raman self-frequency shift. The optical phase conjugation is proposed to compensate for the high-order dispersion and high-order nonlinear effect. However, the spectral phase conjugation can’t eliminate the time delay induced by Raman self-frequency shift and the temporal phase conjugation can’t avoid the pulse splitting into two pieces induced by the third order dispersion. Thus, the schemes of the combination of the spectral and temporal phase conjugation are supposed to use for recovering the distorted pulse and reducing the time delay induced by the third-order dispersion and Raman self-frequency shift. There exist the optimized schemes to deploy the relative position between the spectral phase conjugator and the temporal phase conjugator to obtain the optimization transmission qualities with the minimum penalty for timing jitter and pulse distortion. In addition, the spectral phase conjugator and temporal phase conjugator should be settled down after the midway and before the midway, respectively, in order to obtain the residual third-order dispersion to enhance the frequency red-shift trend of two polarization components after the temporal phase conjugator for reducing the time delay.

This paper focuses on the problem of buffering performance optimization for optical packet switch with shared WDM (Wavelength Division Multiplexing) optical buffers. For this purpose, several packet scheduling algorithms are proposed to make full use of buffering resources. Influences of these algorithms on optical packet switch are evaluated by simulation experiments. Simulation results show that under different traffic load, different packet scheduling algorithms should be adopted. In addition, the performance of optical buffers also depends on the granularity of FDL.

We demonstrate transmission of 36×10-Gbit/s WDM signals over 4134 km of standard single-mode fiber using all-Raman amplification and NRZ format. Wide-band fiber delay loop depolarizers are used in the Raman pump unit to reduce the depolarization dependent Raman gain.

An automatic control algorithm considering the pump interaction for flattening the gain of a fiber Raman amplifier is derived from the Raman scattering equations. A pseudo-inverse gain matrix is introduced to adjust the powers of the pump lasers. The pump-to-pump interaction is compensated by the pump coefficients. It is demonstrated experimentally that the convergence of this algorithm is faster than that of the algorithm without considering the pump-to-pump interaction. In a system with 50-km SMF and 13.4km DCF, flattened Raman gain 15dB with a fluctuation no more than 0.40dB over a 45-nm bandwidth and 26dB with a fluctuation less than 0.8dB are realized in the experiment.

The work presented in this paper gives performance comparisons of Carrier Suppressed Return-to-Zero Differential Phase-Shift Keying (CSRZ-DPSK) with Return-to-Zero Differential Phase-Shift Keying (RZ-DPSK) against both the effects of Intra-Channel Four-Wave Mixing (IFWM) and inter-channel Four Wave Mixing (FWM) in 16×40 Gb/s WDM systems with symmetrical fiber link schemes. IFWM gives smaller phase perturbation to CSRZ-DPSK signal than to RZ33-DPSK, although BER performance of CSRZ-DPSK is worse than that of RZ33-DPSK. The numerical comparison indicates that RZ33-DPSK is a better candidate for 40 Gb/s WDM systems than CSRZ-DPSK.

We design a novel tunable interchannel dispersion-slope compensator using a single broad-band nonchannelized sampled fiber Bragg grating (FBG) with chirp in the sampling period. Tunability of the dispersion-slope can be achieved by stressing or heating the grating uniformly, which has a third-order time-delay variation with wavelength, thereby causing a second-order variation in dispersion. By using the equivalent-chirp method, i.e. by chirping the sampling function, we get the time-delay within -1^st order Fourier reflection band and consequently realize dispersion-slope compensator.

At present, polarization mode dispersion (PMD) has become an important limiting factor for the performance of long distance transmission ,high rate data and analog communication systems. This paper introduces the principle and characteristics of PMD, and tells the causes and effects of it. We also report the standardized techniques for measurement of PMD according to the ITU-T Recommendation G.650.At the end ,we show some experimental result in our project.

With the drastic data traffic requirements of Internet, the demand of high bandwidth optical communication for local-area-network (LAN) and storage-area-network (SAN) are increasing. A novel design of Z-axis pluggable 4x10Gb/s fiber-optic transceiver module, which owns the advantages of high port density, low cost, thermal stability and compact size is studied in this paper. The Xenpak-like transceiver modules with 4 channels LC type connectors are designed in the Xenpak mechanical dimension specified in Xenpak Multi-Source Agreement (MSA) to accomplish signal transmission. The Xenpak-like 850nm transceiver modules make use of the 70-way two-row electrical connector, which is originally designed for 10Gigabit Extended Attachment Unit Interface (XAUI), to transmit and receive the 4x10G/bs signal. A signal compensated circuit would be not only implemented to solve and overcome the loss and distortion of the connector interface but to improve signal integrity of 10Gb/s data transmitted over 100mm distance with standard FR4 printed circuit board. Moreover, in order to improve the performance of high-density printed circuit, the effects of electromagnetic interference (EMI) will also consider and discuss in this study. The optical performance of each lane of transceiver module is evaluated to pass the requirements of 10Gigabit Ethernet, respectively. The transmit eye diagram of each lane in the Xenpak-like transceiver modules is tested compliant with 10GE eye mask. And the extinction ratio of each transmitter is also evaluated above 3.5dB, averagely. Besides, the receiver sensitivity of each lane is demonstrated below -11dBm. To fulfill the future traffic demand of Ethernet, the authors have provided an alternative and novel design of 4x10Gb/s transceiver module for LAN/SAN applications.

A novel fiber-in-line optical transceiver structure by focusing the outgoing radiation from a tilted and chirped fiber Bragg grating is proposed. The chirped gratings with 45° tilted angle is designed for an optimum input coupling at 1.55 μm and a FDTD method is used for analysis. Experimental results on beam focusing are shown. We expect this novel design can be used for a low-cost receiver in the bi-directional fiber-in-line optical transceiver.