With extraordinary increase of bandwidth requirement nowadays, service providers are compelled to push fiber deeper into the access networks in order to be able to offer the emerging services that customers are demanding. The economic aspects from the business side also put pressure on operators to deploy single network for multiple applications in order to maximize profitability. There are various technologies used today to increase the capacity of the access network. Fiber-To-The-Home (FTTH) utilizing ATM Passive Optical Networking (APON) has been developed for and field-tried in numerous networks all over the world. The advantages of APON solution lies in its future proof fiber based access technology (high bandwidth and long service life). Fiber and passive components used in the network are very reliable and low in the maintenance cost. The APON technology is standardized in ITU-T G.983, which is originated from Full Services Access Network (FSAN) initiatives. It supports multiple applications and provides powerful quality of service (QoS). In this paper, the requirements for APON FTTH network architecture, transport protocol and operation are examined in line with G.983 standards. The comparisons of APON to other existing high data rate access network solutions are then briefly discussed. Case studies of APON FTTH field trials are used as examples. These trials represented current status of APON FTTH development in US, on both network system elements and fiber plant components. A complete picture of APON FTTH network system and Outside Plant (OSP) design from CO to the residential homes is illustrated by these field trials. Discussions are given on topics related to APON FTTH system elements, such as OLT/ONT design, lifeline support, optical power budget and operation/fault management. Considerations regarding the design and deployment of OSP components focus on the configuration and construction of the fiber plant. Common issues related to field installations such as trouble shooting of fiber plant through splitter point, drop management, and long term reliability, etc. General design guideline on APON system and OSP requirements will be outlined based on the experience gained through limited deployment and field trials. Advances in APON technology in the near future will be debated at the end.

HomeFiber Networks has successfully demonstrated the use of Coarse Wavelength Division Multiplexing in a residential Fiber-To-The-Home installation in Northern California. The primary driver for using Wavelength Division Multiplexing is to allow multiple Ethernet switches to be served over a single pair of fibers. Today, the available dark fiber in and around small metropolitan areas comes at a premium, as it's primarily serving business customers. For Fiber-To-The-Home to even begin to become economically feasible, it's paramount to aggregate as much traffic as possible over as few fibers as possible. In this respect, the local network is somewhat similar to metropolitan area as well as long-distance networks. However, for a local network the cost must be brought to an absolute minimum. Coarse wavelength division offers such a solution, which also is scalable.

This paper discusses the network architecture for delivering voice, data, and video signals over broadband passive optical networks (BPON). We, first, explain BPON physical and media access control layer protocols. Then, we summarize the network configuration for carrying voice, data, and video. In particular, the necessary equipment and interfaces upholding these functions are introduced. Finally, we examine closely its limitation on transporting analog video signals in a wavelength overlay. This paper concludes that BPON can support around 400 TV channels with 64QAM using cost-effective optical components. Yet, it is limited to 30 or so analog channels with AM-VSB even with 'ideal' optical components. The paper also highlights the current standardization effort in FSAN and its initiatives on future BPON enhancements.

The scalability of multi-purpose fiber-optic access network (MFAN) has been improved significantly by using active components at the remote nodes. Unlike the passive MFAN, this network could now support more than 64 optical network units (ONU's).

Recently a lot of studies have been focused on the impact of the wavelength converters in a DWDM network for different switch architectures and different traffic load conditions. In addition, most of the previous studies, has assumed that the holding time of the wavelength connection is similar to that of telephone calls. This is not practical in the current optical network applications since most of connections stay permanently. In this paper, we consider new switch architecture with partial wavelength conversion capability. We investigate the benefit of the wavelength conversion in a DWDM network with limited number of wavelength converters assuming that the holding time of the wavelength connection is infinite. Our simulation results indicate that the performance of an optical network with only 25% of wavelength conversion capability is very close to that of an optical network with full wavelength conversion, for the network topology and traffic pattern considered. In another Metro application, we observe that the use of wavelength converters does not have any impact on the network performance.

The widespread introduction of optoelectronic and multi- wavelength technologies is new network segments depends strongly on the mass production of low-cost, highly functional, integrated and standardized optical components. This paper describes the demand for advanced and application-specific optical components for evolving Ultra- Long-Haul (ULH), Metro, Access, Cable and Wireless Networks, as well as LANS/WANs and Dark-Fiber Networks.

High-margin component technology for the optical layer of transport networks has been dominated in the past half decade by the lucrative long-haul sector. For wavelength-division multiplexing, the optical metro market has been pulled in conflicting directions by the need for lower cost but almost the same performance as long-haul. High component costs has in turn inhibited adoption of pure optical solutions in the metro network. In the past year, crowding the optical component space, a slow-down in North American long-haul deployments, and the rise of a new generation of metro equipment has driven increased efforts from suppliers to produce components specific to metro needs. In this talk, we will examine how metro components differ from their long-haul counterparts and the factors that will affect their market acceptance.

As is known, SDH has become the backbone of today's communication networks. Meanwhile, Ethernet remains the dominating technology in LAN (Local Area Network) for data transmission. Therefore, how to exchange data among separate 10M/100M Ethernets over SDH has become very attractive. This paper gives a single chip solution, which provides an interface between a 10M/100M Ethernet and N E1 links (or a high-speed link) and consequently transmits the inter-Ethernet frames onto the backbone. In addition, this paper sets up the queueing model for the buffer and analyzes the performance of FIFO under self-similar traffic. Also suggestions on congestion control are given.

We propose new tunable liquid crystal Fabry-Perot filters for fiber-optical telecommunication application. The filters have a low insertion loss, fast response time, wide tunable range to cover total c-band, or L-band, or both. They are solid-state filters without moving parts and the tuning voltages are low.

We report the experimental results in transmission of 32 DWDM channels at OC48 over standard single mode fiber of 600 km using Directly-Modulated Laser (DML) transmitters. Fiber loss is compensated with 7 cascaded Erbium Doped Fiber Amplifiers (EDFA). Large chirp of a DML transmitter imposes a transmission limit of distance at about 200 km. Dispersion Compensating Fiber (DCF) is used to recover signal from distortion due to the interaction of fiber dispersion and laser chirp. In our test bed, we achieve the successful transmission with 1.5-dB power penalty for 600 km. Our study also shows that the transmission with DML transmitter incorporated with dispersion compensation exhibits economic advantage over the one with Electric-absorption Modulated Laser (EML) based networks. In addition, the dispersion compensated network can be upgraded to OC192.

Metro networks have evolved dynamically since its position in the network infrastructure. To gain competitive advantage in this attractive market, carriers should emphasize not only just the power of their networks in terms of the speed, number of channels, distance covered, but also the network's versatility in supporting variety of access interfaces, flexibility in bandwidth provisioning, ability of differentiated service offering, and capability of network management. Based on an overview of four emerging metro network technologies, an intelligent metro network control platform is introduced. The intelligent control platform is necessary for carriers to meet the new metro requirements. Intelligent control and management functions of the platform are proposed respectively. Intelligent metro network will bridge the metro gap and open up a whole new set of services and applications.

With the rapid pace of technology evolution, it is easy to allow complexity to overcome simplicity. Service providers are optimizing their metropolitan optical networks as well as extending this bandwidth capacity all the way to their business enterprise customers. However, in doing so, many have also increased the complexity associated with delivering and managing this bandwidth. This talk discusses how recent technologies help service providers design, deploy, and manage a state-of-the-art optical metropolitan network that avoids complexity and optimizes the flow of bandwidth. This talk describes how the needs of the service provider (to deploy an effective, flexible, scalable, and affordable multi-service network) and their customers (to receive quickly provisioned, reliable, and affordable services) are what drive the choices of technology used. A combination of these technologies, including wave-division-multiplexing (WDM), bit-error-rate performance monitoring, and management software, is necessary to promote harmony. This talk will familiarize the audience with how key technologies translate to real-life benefits for the delivery of high-bandwidth optical services.

Network architecture design is the operational framework that begins, at the top level, with capacity design, then works through to the transport and nodal function designs, and finally generates directions for engineering design. In the process, a network architect is faced with the question of balancing the pursuit of his own architecture while managing commercial vendors' influences. In this article, we seek to develop techniques and methodologies for modeling and optimizing the metro network architecture with an emphasis on network capacity planning. Capacity planning directs subsequent architectural and engineering designs. We first discuss how to build metro network models and generate executable procedures of the network analysis. The metro network is described as an object-oriented network with an integration of traffic sites and network entities. The traffic demand, network connectivity and traffic processing are the main object functions that have to be modeled. Cost modeling, which is used to link the architecture with its capital expense, is briefly reviewed. Cost models are established with capacity demands allocated to network functioning points. Afterwards, equipment allocation is performed to determine the network devices needed. Finally, the authors show how to perform network design optimization. The optimization processes are executed on a search engine to produce the best combination of variables of the network entities.

Multi-service Access on the narrow-band DDN (Digital Data Network) leased lines to use the bandwidth more efficiently and reduce the cost has attracted much interest. In this paper, one novel multi-service multiplexing scheme based on DTSA (Dynamic Time-Slot Allocation) is given. This scheme can guarantee the QoS of the multiplexed services such as FAX, Voice and data and adapt to different link rates (64kb/s, 128kb/s, 256kb/s), A model is given in this paper to analyze the data behavior under this scheme. The simulation result and the model result have shown that the QoS guarantee to voice and FAX doesn't compromise the QoS of data service much in the meaning of delay and delay variance when the data load is not too high. The simulation result agrees with the model well when data load is not too high.

Ethernet was first developed as a low cost multi-access protocol for local area networks (LANs). In the last 15 years, there has been tremendous development in the Ethernet technology. Modern Ethernet is no longer just a LAN protocol. As a result of the Internet proliferation, Ethernet has taken a much broader meaning in today's fast growing telecommunication world. High speed Ethernet standards using optical fiber as the transmission medium have been developed to meet the increasing bandwidth demand. These systems are penetrating into metropolitan area networks (MANs) and wide area networks (WANs). They will also be used with DWDM technologies. In this paper, we describe the development and application of modern Ethernet technologies.

IP-based Virtual Private Network becomes more and more popular. It can not only reduce the enterprise communication cost but also increase the revenue of the service provider. The common IP-VPN application types include Intranet VPN, Extranet VPN, and remote access VPN. For the large IP-VPN market, some vendors develop dedicated IP-VPN devices; while some vendors add the VPN functions into their existing network equipment such as router, access gateway, etc. The functions in the IP-VPN device include security, QoS, and management. The common security functions supported are IPSec (IP Security), IKE (Internet Key Exchange), and Firewall. The QoS functions include bandwidth control and packet scheduling. In the management component, policy-based network management is under standardization in IETF. In this paper, we discuss issues on how to integrate the QoS and security functions in an IP-VPN Gateway. We propose three approaches to do this. They are (1) perform Qos first (2) perform IPSec first and (3) reserve fixed bandwidth for IPSec. We also compare the advantages and disadvantages of the three proposed approaches.

The popularity of the Internet has caused the traffic on the Metro Area Network (MAN) to grow drastically every year. It is believed that Wavelength Division Multiplexing (WDM) has become a cornerstone technology in the MAN. Solutions to provide a MAN with high bandwidth, good scalability and easy management are being constantly searched from both IP and WDM. In this paper we firstly propose a metro optical network architecture based on GMPLS--a flexible, highly scalable IP over WDM optical network architecture for the delivery of public network IP services. Two kinds of node including Electronic Label Switching Router (E-LSR) and Optical Router (O-LSR) are involved in this metro optical network architecture. Secondly, we mainly focus on design issues of OR including multi-granularity electro-optical hybrid switching fabrics, intelligent OTU, contro l plane software and etc. And we also discuss some issues such as routing, forwarding and management of OR. Finally, we reach conclusions that OR based on GMPLS and hybrid-switching fabrics is suitable for current multi-services application environment of MON and optimistic for IP traffic transfer.

Two optical packet switching metropolitan area networks are introduced. They adopt different ring architectures and related media access control protocols. Their simulation system is set up to study both their physical and logical characteristics. Simulation system adopts self-similar traffic that has been demonstrated to accord to real traffic. This simulation system can layout the MAN with optimized network architecture, node scale, wavelength multiplexing number and system physical and logical response parameters according to the design requirements including bandwidth, QOS, traffic load and so on.

The new definition and implementation of new communication network architectures and elements in the present data-centric world are due to dramatic change in technology, explosive growth in bandwidth requirement and de-regulated, privatized and competitive telecommunication market. Network Convergence, Disruptive Technology and Convulsive Market are the basic forces who are pushing the future network towards Packet based Optical Core Network and varieties of Access Network along with integrated NMS. Well-known Moore's law governs the result of progress in silicon processing and accordingly the present capacity of network must be multiplied by 100 times in 10 years. To build a global network which is 100 times powerful than present one by scaling up today's technology can not be a practical solution due to requirement of 100 fold increase in cost, power and size. Today's two network (Low delay, fixed bandwidth, Poisson voice traffic based, circuit-switched PSTN/PLMN and variable delay, variable bandwidth, no-guaranteed QoS based packet switched internet) are converging towards two-layer network (IP and ATM in lower layer; DWDM in network layer). SDH Network which was well drafted before explosive data traffic and was best suitable for Interoperability, Survivability, Reliability and Manageability will be taken over by DWDM Network by 2005 due to 90% of data traffic. This paper describes the way to build the Communication Network (either by migration or by overlay) with an overview of the equipment and technologies required to design the DWDM Network. Service Providers are facing tough challenges for selection of emerging technologies and advances in network standard for bandwidth hungry, valued customers. The reduction of cost of services due to increased competition , explosive growth of internet and 10GbE Ethernet (which is being considered as an end-to-end network solution) have given surprise to many network architects and designers. To provide transparency to data-rate and data-format the gap between electrical layer and Optical backbone layer has to be filled. By partitioning the Optical Bandwidth of Optical Fibre Cable into the wavelengths (32 to 120) Wavelength Division Multiplexing can transport data rate from 10MB/s to 10GB/s on each wavelength. In this paper we will analyze the difficult strategies of suppliers and obstacles in the way of service providers to make DWDM a reality in the field either as Upgrade or Overlay or New Network. The difficult constraint of protection scheme with respect to compatibility with existing network and network under development has to sorted out along with present standard of Optical Fibre to carry DWDM signal in cost effective way to Access , Edge and Metro part of our network. The future of IP under DWDM is going to be key element for Network Planners in future. Fundamental limitation of bit manipulation in Photonic domain will have implication on the network design, cost and migration to all optical network because Photons are computer un-friendly and not mature enough to give memory and logic devices. In the environment of heterogeneous traffic the DWDM based All Optical Network should behave as per expectation of users whose primary traffic will be multi-media IP type. The quality of service (QoS), Virtual Path Network (VPN) over DWDM, OXC and intelligence at the edge will play a major role in future deployment of DWDM in our network . The development of improved fiber characteristics, EDFAs and Photonic component has led the carriers to go for Dense WDM Network.

In this paper we present the topology of an optically transparent DWDM network testbed (part of the KomNet field-trial). We show results of our investigation on performance monitoring in the optical layer of this complex network. The testbed consists of a remotely configurable DWDM metropolitan area ring network (capable of transport up to 80x10 Gb/s) with two OADMs and one HUB/OXC and a long-haul bi-directional DWDM link (16x10 Gb/s, 8 spans over 750 km).

Optical Add/Drop Multiplexer (OADM) is an important network element. In the ring architecture, OADM can be introduced to make efficient use of network capacity, network protection, wavelength routing and many more good features. In this paper, an OADM with high performance realized by us is demonstrated. The key technical problem , solving method and design rule for the OADM are given. The experiment results of long distance transmission by use of the OADM are illuminated by some figure . It shows that the OADM realized by us is advanced, practical, reliable, and applied in China Advanced Info-Optical Network (CAINONet).

Optical cross connects (OXC) play a key role in the contemporary metro DWDM networks. It provides optical interconnection between multiple interoffice metro fiber rings in addition to the simple wavelength add/drop functions. It also provides re-configurable routings and efficient recovery capabilities for mesh architectures with multiple access nodes over a metropolitan area. Any traffic traveling through such a multi-node ring and mesh network could potentially pass through a cascade of several optical cross-connects and several fiber spans. The cascadability is an important function of OXC in optical networks. This paper experimentally demonstrates the cascadability of all optical cross connects (OXC) in high speed (10Gb/s), DWDM (32l) optical networking through 300km SMF-28 fiber without regeneration. The power penalty is less than 0.5dB compared to the system without OXCs but with same amount of attenuation as the OXCs. The results conclude that by carefully managing the power budget and dispersion compensation, the penalty free transmission through cascaded OXC system can be achieved.

By designating different frequencies to different mark positions of the modified prime code, a frequency-time spreading code is proposed. The side lobe of auto-correlation and cross-correlation of the proposed code have relatively lower values. The scheme greatly improves the total number of the available codes in asynchronous optical code division multiple access (OCDMA) system. Numerical results show that the asynchronous OCDMA system using the proposed code has excellent performance in the presence of double hard limiters. We also introduce an encoding/decoding device which is based on the frequency-time spreading modified prime code.

Performance of asynchronous phase-encoded OCDMA system is evaluated on the condition that receiver noise and the impact of fiber channel are neglected. Phase-encoded optical signal (pseudo random optical signal with low intensity) is analyzed in the view of stationary random process. The pseudo random optical signal with low intensity is seen as a sample function of a certain stationary random process which is ergodic in strict sense. The analysis results reveal that the variance of the corresponding random process is only inversely proportional to the code length while the root-mean-square width of the phase-encoded optical signal is proportional to the width of initial optical pulse and the code length F. The numerical results demonstrate that the better system performance can be achieved in case of larger code length and shorter initial optical pulse.

In this paper, perfonnance of OCDMA network implemented using single wavelength with optical orthogonal codes (OOCs) and wavelength division multiplex (WDM) networks with embedded CDMA (WeCDMA) has been evaluated. Performance ofthis network is also evaluated when it is implemented with multi-wavelength optical orthogonal codes (MWOOCs) andbit stuffed unipolar m-sequence encoding schemes. In the above evaluation, the receiver is considered to be a PINphotodiodepreceededby an optical amplifier (PIN+OA). Expressions for the probability oferror are derived alter taking into consideration optical multiple access interference (0MM), optical amplifier noise andvarious sources ofreceiver noise. The numericalvalues are computed for practical values ofparameters andthe results are compared for all the above schemes. It is observed that OCDMA networks implemented using WeCDMA with OOC perfonn better thanthe single wavelengthwith OOC forthe samebandwidth expansion. It is also observedthat perfonnance of OCDMA networks using MWOOCs is better than bit stuffed unipolar rn-sequence based networks when number of simultaneous users are less, whereas for large number of simultaneous users the network perfonnance with bit stuffed unipolar msequence encoding scheme is better compared to that with MWOOCs encoding schemes.

Group velocity walk-off effect on frequency hopping optical code division multiple access (FH-OCDMA) system is analyzed. We develop a systematic method which employs transfer function considering encoder, fiber channel and decoder to analysis the FH-OCDMA system. The difference of group velocities affects system performance more seriously than dispersion effect. The results show that single mode fiber is not proper for the FH-OCDMA system even the transmission length is no more than 1 kilometer while the dispersion-shifted fiber is more suitable for FH-OCDMA system.

In this paper plastic optical fiber (POF) access network based on optical code-division multiple access (OCDMA) technology is proposed. Recently, optical transmission using POF has received much attention due to its low weight, large core diameter, flexibility, and easy install, specially its high bandwidth make POF a very attractive candidate for transmission media in access network based on OCDMA technology. In conventional OCDMA system only allows finite units to transmit and access simultaneously according to the number of channels which be restricted by BER. To resolve this problem a novel protocol also is proposed in this paper. The protocol can efficiently support variable-sized messages, and any new unit can join the network at any time without requiring network initialization. To implement the demonstration, each optical network unit is equipped with a fixed optical encoder/decoder and tunable optical encoder/decoder. The optical encoder/decoder employing planar holographic optical processors (HOPs) in this system maybe low cost fabrication. The network throughput and average delay using various system parameters has been investigated by numerical analysis and simulation experiments. It is showed that the dynamic control protocol in this POF access network based on OCDMA technology is valid and efficient.

In this paper, the relationship between source extinction ratio and performance of optical fast frequency-hop CDMA systems was deeply investigated, and the effect of extinction ratio on BER of systems was analyzed. A new concept of ideal decision, i.e. adjusting decision threshold of receiver according to extinction ratio, was presented. Sensitivity of system performance to extinction ratio degradation and its evaluation was studied. Based on these analyses, the sensitivity of FFH-OCDMA system was compared to that of DS-OCDMA system. Finally, a modified scheme of FFH-OCDMA system was brought forward.

Dense wavelength division multiplexing (DWDM) SDH rings have been extensively deployed in core networks and metropolitan area networks (MAN). Furthermore, advances in optical ADM (OADM), optical DXC (OXC) and Erbium-doped fiber amplifiers (EDFA) have resulted in evolving core networks and MAN to all-optical networks. Due to its low cost, flexibility and low loss, OADM based on fiber grating and optical circulator has exhibited potential in applying to DWDM SDH ring networks[1]. Although fiber-grating based OADM has the feature of low loss, EDFA is always deployed to compensate the loss and further improve the network performances. However, amplified spontaneous emission (ASE) noise will limit the improvement of the network performances[2]. Fiber grating is a stable and reliable filter, and this high performance filter is ideal for simultaneously achieving wide-band signal transmission and a high-level of ASE noise suppression in amplified DWDM networks. In this article, performance analysis of amplified DWDM SDH ring with OADM based on fiber grating and optical circulator has been performed with an explicit method - Signal Flow Graph. The results show that the throughput and the power penalty of amplified DWDM SDH ring with fiber-grating based OADM are effectively improved, considering ASE suppression by fiber gratings. Additionally, the network performances of pre-amplifying, post-amplifying and in-line amplifying schemes are analyzed, implying that the network performance optimization of amplified DWDM SDH ring with fiber-grating based OADM could be achieved with proper scheme.

A protection scheme that chooses protection routes in advance in All-Optical Mesh network is proposed in this paper. Two rules, minimum relativity among routes and minimum the number of hops, are given and analyzed in detail. In order to perform protection quickly and correctly, the compromise between two principles must be considered when choosing protection routes. The protection method that appointing ring networks in mesh networks is proposed too. In addition, some key technologies such as avoiding oscillation, line protection and misconnect squelched are also proposed in this paper.

One FBG-MZI-based component and one mediator are proposed in this paper. The proposed OADM building module consists of one FBG-based MZI module and one reconfigurable module, and it can add/drop or pass-by the specified wavelength. OADM created by concatenating the proposed modules, owns wavelength modularity, and can add-drop and pass-by any specified wavelengths. The mediator is one simplified 4 x 4 optical switch matrix. It can be used to interconnect multiple OADM modules to build one OXC with multiple ports.

Homodyne crosstalk can be divided into two kinds: coherent and incoherent crosstalk in WDM optical network. A general analyzing model of crosstalks in WDM optical network is developed in this paper. According to the model, two different kinds of crosstalks, viz. coherent crosstalk and inconherent crosstalk, are distinguished from each other. The experimental results, 1dB power penalty for about 20dB coherent and incoherent crosstalk signal well agree with theoretical one, could be important to the WDM optical network designing.

Optical add/drop multiplexers (OADM) can play an important role in WDM optical transport network systems. In this paper, we focus on the performance of using OADM integrated with pre-EDFA in multi-wavelength optical transmission networks called OADM subsystem. We calculate the receiver sensitivity incurred in an OADM subsystem with pre-EDFA. We theoretically investigate the effects of the spontaneous emission factor, and pre-amplifier gain, spontaneous emission factor, on the receiver sensitivity. The theoretical results agree well with the experimental results we have done.

One of the advantages for Discrete Multitone(DMT) modulation systems compared to single carrier systems is that the power and bits may be allocated to each subchannel in some way for optimal performance. Some algorithms proposed before have high complexity and use flat power distribution which is not the optimal distribution. Based on the detail analysis of Water-Pouring theorem, a new practical algorithm is presented which match approximately matches the Water-Pouring distribution and can also the discrete bit loading problem. The simulation results show this algorithm has better performance and lower complexity.

After the end of this year, most of main carriers in China will complete their backbone optical network construction, which will be up to 320Gb/s or higher. After the hundreds times increase of the bandwidth in backbone network, what is the next focus for bandwidth increase? It is local and metropolitan optical network, which will provide multi-service to the customers such as TDM voice, leased line, Ethernet interconnection for big enterprises, Internet surfing and so on. How to build the optical bandwidth service network in metropolitan or local area? For different carriers, there may be different solutions. Three general solutions will be introduced and discussed.