We propose and demonstrate a simple and efficient technique to monitor the optical paths in all-optical WDM transport networks using pilot tones. This technique could also monitor the crosstalk increases caused by the partial failures of optical switches in the optical cross-connects. In addition, unlike the other monitoring techniques based on the pilot tones, this technique was insensitive to the cross-gain modulation of EDFA. Thus, the proposed techniques could be used to monitor both the optical paths and switch failures in a large-scale all-optical WDM transport network.

We propose and demonstrate a simple technique to monitor the optical signal-to-noise ratio (OSNR) of WDM signals by analyzing the high-frequency receiver noise above the modulation frequency. The results show that the OSNR of WDM signals could be monitored with an accuracy better than 1 dB even after transmission over 640 km of single-mode fiber.

Survivability is an important issue in communication networks. Protection and restoration are two mechanisms to provide the survivability. Conventional restoration scheme is classified into link and path restoration. This paper proposed a new hybrid restoration scheme that combines link and path restoration methods using priority wavelength strategy. This new method provides better recovery blocking probability than the conventional link/path restoration method. Due to the characteristics of protection methods, which must reserve spare capacity and is stiff to variable traffic, we further proposed a dynamically adjusted protection scheme that can have more flexible usage of network resource and gives better recovery blocking probability. We simulate the performance and compare with the conventional protection/restoration schemes. The results show that the proposed methods have better recovery blocking probability than others.

The optical network essentially provides point-to-point connectivity between two end equipments that are directly attached to optical network edge in terms of high bandwidth lightpath. Recently, there are lots of proposals to enhance intelligence of optical network as well as flexibility, reliability and scalability. However, it is difficult to validate, verify and evaluate these proposed protocols because they usually employ distributed control mechanism, which requires a huge network testbed. One alternative to expensive hardware is the use of computer simulations. In this paper, we constructed a functional and fault model for the transport plane of optical networks and implemented an Optical Network Simulator prototype on the basis of our model. Finally, using the developed tool, we have evaluated our design of Optical Network Management Protocol (ONMP), which is a Network-Network-Interface (NNI) optical network protocol in control plane.

Emerging optical technologies with the rapid growth of the data traffic has made IP over WDM as next generation network infrastructure. Given the wide range of services envisioned for future IP networks, network survivability is a crucial concern. There are many research activities on how to provide resilience in IP over WDM networks, but most of them focus on either optical layer survivability or IP layer survivability. In this paper, we present an integrated protection scheme to protect IP over WDM network. Mainly, we investigate the impact of shared protection mechanisms in IP over WDM networks. The purpose of the shared protection is to improve the network utilization while assuring the reliability attributes that commit to the users' applications. We study shared protection schemes' applications in different network layers, such as optical layer protection, and IP over WDM two-layer protections. Finally, performance evaluations based on simulations are presented.

In recent years, software agent has gained great momentum in network management, and it is recognized as an effective technology to build distributed network management system. However, most researchers focus their attentions on the distributed management applications based on some specific agent architectures or platforms. Up to now, there is no a common methodology to guide the construction of the system of this paradigm. From the viewpoint of modeling, this paper adopts multiple models used in software agent and combines them together to form a novel systematical modeling approach. This modeling approach covers two stages in development: system analysis and system design. A prototype for managing network flows was also developed to testify the modeling approach.

This paper presents a Policy-based Control Plane, which allows Carrier Optical Network to dynamically configure and deploy Optical Virtual Private Networks (Optical VPNs) over multiple administration domains. A configurable policy-based control mechanism is required to regulate and control the information propagation across both UNI and NNI interfaces and the actions allowed on behalf of the users. The proposed framework extend the ITU Control Plane by adding based IP-centric policy mechanisms like admission control protocol COPS (Common Open Policy Service).

Policy-based management system provides the configuration capability for the system administrators to focus on the requirements of customers. The service level agreement performance monitoring mechanism helps system administrators to verify the correctness of policies. However, it is difficult for a device to process the policies directly because the policies are the management concept. This paper proposes a mechanism to decompose a policy into rules that can be efficiently processed by a device. Thus, the device may process the rule and collect the performance statistics information efficiently; and the policy-based management system may collect these performance statistics information and report the service-level agreement performance monitoring information to the system administrator. The proposed policy-based management system achieves both the policy configuration and service-level agreement performance monitoring requirements. A policy consists of a condition part and an action part. The condition part is a Boolean expression of a source host IP group, a destination host IP group, etc. The action part is the parameters of services. We say that an address group is compact if it only consists of a range of IP address that can be denoted by a pair of IP address and corresponding IP mask. If the condition part of a policy only consists of the compact address group, we say that the policy is a rule. Since a device can efficiently process a compact address and a system administrator prefers to define a range of IP address, the policy-based management system has to translate policy into rules and supplements the gaps between policy and rules. The proposed policy-based management system builds the relationships between VPN and policies, policy and rules. Since the system administrator wants to monitor the system performance information of VPNs and policies, the proposed policy-based management system downloads the relationships among VPNs, policies and rules to the SNMP agents. The SNMP agents build the management information base (MIB) of all VPNs, policies and rules according to the relationships obtained from the management server. Thus, the proposed policy-based management system may get all performance monitoring information of VPNs and policies from agents. The proposed policy-based manager achieves two goals: a) provide a management environment for the system administrator to configure their network only considering the policy requirement issues and b) let the device have only to process the packet and then collect the required performance information. These two things make the proposed management system satisfy both the user and device requirements.

With Internet traffic doubling every nine months, demand will only increase pressure to deliver all-optical networking (see Figure 1. Network Life Cycle (Yesterday)). Every technology lifetime follows the typical introduction, deployment, and obsolescence. Presently, optical transport technology, growing tenfold each year, is outpacing microprocessor technology. This has allowed optical technology performance to exceed Moore's Law, which assumed microprocessor speed would double every eighteen months.

Next generation optical network technology has attracted lots of attention from both academia and industry. Large optical networks are typically partitioned into metro and core sub-networks due to the different service and technology requirements. A client optical connection may pass through multiple sub-networks, originating in a metro sub-network, traversing the core network, and accessing the destination on a different metro network. This structure is similar to current ISP networks, which may be divided into multiple routing areas for scalability. However, optical networks are different from IP networks. The difference affects the routing information that needs to be exchanged in metro/core optical networks such that IP routing protocols require significant extensions if they are to be applied in optical networks. In this paper, we focus on Metro/Core routing requirements and present a proposal to extend the current IP multi-area Open Shortest Path First (OSPF) protocol to exchange routing information over metro/core sub-network interfaces. A brief discussion of our prototype implementation is given at the end of this paper.

This paper consider the design of lightpath Routing and Wavelength Assignment (RWA) problem in Wavelength Division Multiplexing (WDM) networks with or without wavelength conversion. To minimize the required network cost, one has to device as few network devices and take the least network building cost with respect to the required demand requirements. In our model, the required cost including the one to install wavelengths in the network nodes and the building cost to use a specific wavelength in a specified optical link. The problem is formulated as a binary linear programming where the objective function is the minimization of network building cost. Many literatures have pointed out that solving the formulation of this kind is very computationally demanding and heuristic algorithms and/or relaxation techniques are needed for problems with nontrivial size. In this paper, a Lagrangian relaxation based solving procedure is developed for the RWA problem. In particular, We first transfer the RWA problem into a multicommodity integer flow problem using graph transformation technique by adding some artificial network nodes and links with proper cost on them. To achieve minimum network cost, two problem solving phases are developed for networks with and without wavelength converter respectively. In the first phase, we try to optimize the cost of routing without violating the wavelength continuity constraints. If no feasible solutions are obtained in this phase, it means there are no sufficient paths to route lightpaths without wavelength converter. We then take another graph extension with wavelength converter geared to the RWA problem and then applying a shortest path based heuristic algorithm to solve the problem based on the solution obtained from first phase. Two network topologies, GTE network and NSFNET network, are used to evaluate the computational results. Examining the Lagrangian based heuristic results and the lower bounds reveal that the proposed algorithm can efficiently provide a nearly optimal solution for our problem.

The data traffic is growing dramatically in recent years, especially in Metro areas. The emerging of Data Networks and Optical Networks presents new challenges of the network design. Here we use the term of Optical Internet as it best represents the integration of the optical layer and data layer in the future networks. This paper describes the network design solutions of Metro Optical Internet. It explains the advantages and disadvantages of Data Over X technology, protection and restoration schemes for Metro Optical Internet Networks and the future smart optical network standard-G.ASON

Design, control and management of all-optical transparent islands based on fibers, optical amplifiers and transparent Optical Cross Connects (OXC) have to deal with a vast number of issues such as network topology, routing, protection, and restoration schemes. However, since in such a transparent network light-paths generally have to transit through several fiber sections, amplifiers, and OXCs, all of which add noise and degrade the signal performance, optical layer network protocols have to be aware of the end-to-end signal quality for each specific light-path. This information should be obtained without burdening these protocols with details of the physical layer. In this paper we propose a simplified network design approach based on normalized optical transmission sections where the physical links are built from a cascade of identical basic blocks. For an upper layer protocol only the number of traversed blocks has to be monitored for estimating signal quality. Using numerical simulations the maximum number of cascaded building blocks for 10 Gbit/s RZ modulation format for different basic building block-configurations is investigated and compared with the performance of NRZ based systems under similar conditions. It is shown that robust transparent islands of about 1600 km maximum link length without 3R regeneration are feasible for both modulation formats using the concept of normalized sections.

To natively support the bursty IP datagrams over all-optical Wavelength Division Multiplexing (WDM) networks, the Optical Burst Switching (OBS) WDM network has been proposed as a suitable architecture for future optical Internet backbone networks. However, managing the OBS network will be complicated due to the scale of the networks and the correlation between different technology layers. This paper presents an information model for the OBS core node, from the network-element view, to describe the management information flows between the optical burst layer and the traditional WDM transport layer, and how to model them using various Managed Objects (MOs). We also provide the structure of Management Information Base (MIB) used in SNMP management interface for managing the parameters identified at different layers.

In this paper the influence of routing and wavelength assignment (RWA) algorithms on the blocking probability in wavelength division multiplexing (WDM) all optical network under the condition of static routing was studied. The Single Route and Reroute routing strategies, longest path first (LPF) and shortest path first (SPF) wavelength assignment strategies were compared respectively. In moderately connected network, it is proved by numeric simulation that lower blocking probability can be obtained by Reroute and SPF strategy in virtual wavelength path (VWP) mode. While in wavelength path (WP) mode, when the traffic load is low, Reroute and LPF strategy should be adopted; when the traffic load is high, Reroute and SPF should be adopted. In lowly connected network, Reroute strategy has no effect, so Single Route can be used combined with SPF strategy.

Based on an analytical model which is verified by numerical simulations, we study the influence of the dispersion map on XPM induced signal degradations in long-haul WDM systems for different fibre types. Distributed undercompensation is shown to be an optimum choice for 10 Gb/s per channel transmission. Statistical studies of map variations confirm, that moderate deviations from the optimum dispersion map further reduce the XPM limited reach. For transmission over standard single-mode fibre, XPM generated in the dispersion compensating fibers which are required for a compensation of the large dispersion of the transmission fiber may impose a stringent limitation for terrestric ultra-long haul systems.

In this work, our aim is to provide finer priority levels in the design of a general-purpose Web multimedia server with provisions of the CM services. The type of services provided include reading/writing a web page, downloading/uploading an audio/video stream, navigating the Web through browsing, and interactive video teleconferencing. The selected priority encoding levels for such operations follow the order of admin read/write, hot page CM and Web multicasting, CM read, Web read, CM write and Web write. Hot pages are the most requested CM streams (e.g., the newest movies, video clips, and HDTV channels) and Web pages (e.g., portal pages of the commercial Internet search engines). Maintaining a list of these hot Web pages and CM streams in a content addressable buffer enables a server to multicast hot streams with lower latency and higher system throughput. Cold Web pages and CM streams are treated as regular Web and CM requests. Interactive CM operations such as pause (P), resume (R), fast-forward (FF), and rewind (RW) have to be executed without allocation of extra resources. The proposed multimedia server model is a part of the distributed network with load balancing schedulers. The SM is connected to an integrated disk scheduler (IDS), which supervises an allocated disk manager. The IDS follows the same priority handling as the SM, and implements a SCAN disk-scheduling method for an improved disk access and a higher throughput. Different disks are used for the Web and CM services in order to meet the QoS requirements of CM services. The IDS ouput is forwarded to an Integrated Transmission Scheduler (ITS). The ITS creates a priority ordered buffering of the retrieved Web pages and CM data streams that are fed into an auto regressive moving average (ARMA) based traffic shaping circuitry before being transmitted through the network.

SDL is designed for transmission packet data in optical communications. To reduce the complexity of SDL framers, we present a hardware implementation that uses only one CRC detector. In addition, the DC balance operation arithmetic and hardware circuit are derived and integrated into the proposed circuit.

Provision of Quality-of-Service (QoS) guarantees is an important and challenging issue in the design of integrated-services packet networks. Call admission control is an integral part of the challenge and is closely related to other aspects of networks such as service models, scheduling disciplines, traffic characterization and QoS specification. Here we are concerned with two sub-optimal scheduling discipline [Start-Time Fair Queuing, Self-Clocked Fair Queuing] (STFQ , SCFQ) which are derived from the optimal scheduling discipline Generalized Processor Sharing (GPS). Using an easy and efficient discipline is essential for integrated services, which require specific QoS, and which ATM is mainly used to serve. An evaluation study is performed on a certain Call Admission Control (CAC) scheme. Also, a comparison between the presented disciplines is given.

High-speed railway that has been progressing very quickly is one of the greatest techniques in present time because of its high speed, economy, comfort, environment benefits and other advantages. And among all of these, safe is the backbone and lifeline, so the chief task in developing high-speed railway is to establish safety guarantee system. Also in this safety guarantee system, train control is one of the key techniques to guarantee safe train operation and to advance ability of transportation, so operation safe is located in the hardcore position. That is to say, it is imperative to set up a safe, real-time and reliable automatic train control system. And we can easily find out that this kind of system is now developed and applied in many countries. Important information related to train control, such as the received and transmitted information of track-sided equipment, is called safe information, because it deals with train operation's safe, reliability, and even directly with people's life and wealth. It is so important that if there were some kind of fault with its making, transmission, or processing, fatal accident would occur. So to some degree, it is impossible to transmit and process this information through present railway communication network because of the former's extreme importance and the latter's no safe capability. Therefore, a specific communication network that mainly considers about safe transmission and management should be established in order to realize the specific function for this specific information. High-speed railway safe information transmission optical LAN, which adopts optical fiber as transmission media and transmits safe information, is a kind of LAN designed for the request for safe, real-time and highly reliable automatic train control system in the process of our country's high-speed railway construction and commonly train speed. In this paper, after analyzing the characteristics of automatic train control system and the structure, status and function of the safe information transmission optical LAN, we analyze the main factors which affect the network's safe transmission performance based on the principle of safe design. Then, on the foundation of the elementary principle of the concept, fail-safe, in railway signal field, we discuss the real network's safety and realization of the software and hardware. The final experiment indicates that the real network works robustly, performs reliably, and achieves the designed aim. When some kind of faults occur which produced by hardware, software or transmission media, the network can easily and reliably deal with them and transfer to safety mode that will ensure the train's operation. The analysis and design of this network is a meaningful step in constructing our country's railway special communication network and also the network can be used in other similar conditions that demand extreme safe.

Recent advances in optical devices greatly enhance the feasibility of optical code-division multiplexing/wavelength-division multiplexing (CDM/WDM) network. In this paper, the performance in blocking probability in a two-dimension code/wavelength routing network is investigated. Facilitated by the optical code and wavelength conversion, the capability in reconfigurability, restoration and protection of the optical layer is enhanced. The network blocking performances under various conversion capabilities are investigated and the blocking probabilities are given in closed-form expressions. Under certain conditions, the analysis of the two-dimension routing network can be extended to N-dimension routing network by iterative decomposition procedure. Numerical results are obtained to show the performance improvement from the optical code conversion in terms of code conversion gain. With wavelength conversion provided in the network, the addition of the code conversion achieves a conversion gain as high as 10⁴ even when the code conversion is provided with a low placement density of 0.2.

ANDP (Automatic Neighbor Discovery Protocol) provides an automatic discovery and monitoring of the channel connections between two neighboring OXCs. Based on each channel's status and attributes, OXC can automatically discover its local topology and resource information. ANDP can also be used to detect the failed channels, which do not carry any service traffic. ANDP performs its function by the hardware interrupts from the physical layer, such as DOS. A test access module that is capable of cross-connecting any type of optical signals is necessary for ANDP.

Network convergence, including service, timing and management convergence, is a trend of future telecommunication networks. Because network convergence can provide carriers with cost reduction, highly integrated applications as well as greater flexibility and functionality, new technologies and standards have driven this convergence tide. However, network management convergence, managing disparate networks with a unified platform, has been a challenging task in sophisticated telecommunication network environments. Administrators are faced with the task of managing various devices with several different applications, without an effective tool set to provide visibility across the network. In general, multiple transmission networks such as SDH, PON, HDSL and digital microwave are adopted in GSM infrastructure to transport mobile traffic between BTS and BSC. Traditional method of managing these devices is that GSM network (MSC, BSC and BTS), SDH, PON, HDSL and digital microwave are managed independently. In this paper, a converged network management platform, named OMConvergence, is proposed and demonstrated. The platform aims at managing the whole GSM network covering SDH, PON, HDSL or digital microwave sub-networks within it. The OMConvergence comprises of remote access methods of OAM message, as well as processing of multiple network management protocol such as ECC (Embedded Control Channel), Q3 and simple network management protocol (SNMP). The management and maintenance message of various devices physically converges to E1 timeslots at the sides of BTS, and then convert to unified IP packages before it is terminated at the side of BSC or network administration center. In addition, extended applications of OMConvergence in image monitoring of BTS surroundings are also demonstrated.

Atomic functional architecture model has been accepted by ITU-T to describe the functional architecture of transport networks as well as equipment specification in an implementation independent way. With the existing functional architecture components four different solutions have been proposed to model the complex optical channel (OCh) layer network. One approach is to assign 3-R to the OMS/OCh adaptation function; the second approach is to introduce dedicated transparent layers or to split single OCh layer into several sub-layers; the third is to be deployed by using an abstract characteristic information (CI) definition; the last is using super layer architecture component. This paper analyzes these four methods and presents a new approach by introducing virtual layer network and virtual CI. Virtual adaptation, virtual connection point and virtual link connection are also introduced in order to cover the different transparent connections within a layer network. In addition, examples of using virtual layer component to model the OADM with transparent pass-through and to provide the necessary transparent connections for optical network elements are presented in this paper.

It is important to resolve the questions of how many BIP bytes are required and which EDC should be adopted for performance monitoring in Optical Transport Networks (OTN). Based on the Poisson Error Model and Burst Error Model, we obtain the mathematical expressions for the probabilities of occurrence of a block error and of detecting a block error, as well as the detection efficiency. Analyses and simulations to determine the maximum BER that can be detected (i.e., the saturation BER) and the minimum detection efficiency are given. Conclusions are drawn as follows: one BIP byte per tandem connection monitoring field is adequate for performance monitoring; 2*BIP-4 EDC should be adopted to meet the requirement that the detection efficiency be no less than 90%; 8*BIP-1 EDC is acceptable if the real-time BER measurement estimate is not required.

ATM Passive optical network (ATM-PON) is a kind of low-cost access network for resident and small-business users while Usage Parameter Control (UPC) is an important traffic control function in such networks. The basic traffic control function in an ATM network is the combination of UPC, network parameter control (NPC) and connection access control (CAC). As the entrance of an ATM network, an ATM- PON must have the function of UPC to check it the cell streams entering the network comply with the traffic contract negotiated between the user and the network during the connection setup. In this paper the UPC parameters negotiation methods in ATM-PONs are introduced, and the control algorithms and their realization are also discussed with the necessary simulation results.

The SDH network self-healing mechanism is a very important content in research of SDH networks survivability and is also key to ensure network reliability. Based on the analyzing of traditional SDH self-healing ring, we proposed a new four fibers bi-directional intelligent path protection self-healing ring (BSHR/4IP) mechanism. By computer simulation, we estimate the performance of the proposed mechanism and prove that it has a high performance at restoration time, restoration ratio, utilization and complexity.

The more and more telecommunication networks are constructed with multi-techniques. This paper looks at several aspects of survivability ATM over SDH networks. Considered the specialty of ATM over SDH networks architecture, and focused on meshed networks, we proposes a new ATM over SDH network survivability scheme based on failures for the multi-layer networks: using failure-case-oriented restoration/protection implementations in order to meet the demands for the survivability of networks.

With rapid development of computer network and a wide application of Internet, IP telephony has been realized. Currently, one of the most important research areas of IP telephony system is to develop the IP telephony gateway. In this paper, a new idea of design on an Integrated IP Telephony Gateway is proposed, and the hardware of Integrated IP Telephony is discussed in detail. The software architecture of the Integrated IP telephone gateway is studied, and this paper focus on the voice compression module and SPOX schedule module. In addition, Stochastic Petri Net Model of an Integrated IP Telephony Gateway is set up, and its performance analyzed through the Stochastic Petri Net model. Array of Petri Net Model is computed, and educes the mathematical relationship of CPU efficiency and structure parameters of the IP telephony gateway.

One novel scheme for routing optical supervisory channels adhering to optical cross-connect (OXC) is developed in this paper. Beginning with the definition of the new frame to carry the optical channel (OCH) layer overheads, the proposed scheme processes the overhead associated the individual optical channel, which is carried by optical supervisory channel (OSC), in the similar way as the message-oriented data communication channel (DCC) in OSC. Therefore, all OCH-specified information carried by OSC can be routed by a uniform manner.

In GALENOS (Generic Advanced Low-cost trans-European Network Over Satellite) a trans-European competence network via satellite, dedicated to telemedical applications, has been realized. Using off-the-shelf hardware components and a specially developed high-end software communication system (WinVicos) various telemedical applications, like teleteaching, intraoperative teleconsultation, have been realized.