Rapid development of optical communication techniques has made possible large scale wavelength routing optical networks. Nowadays, these networks are deployed worldwide. Large scale networks have natural hierarchies due to either geographical reasons or administrative consideration and are therefore managed in a hierarchical way correspondingly. That is, networks are composed of several subnetworks, which are themselves composed of other even smaller subnetworks. Erlang fixed point approximation (reduced load approximation) has been extensively used in performance evaluation of end to end call blocking probability in traditional circuit switching networks, which are also called loss networks. This method has also been extended for hierarchical loss networks. However, performance evaluation of call blocking probability in hierarchical wavelength routing optical networks has not yet been presented. In this paper, a kind of approximation method is developed for analyzing end to end call blocking probability in hierarchical wavelength routing optical networks. By iteratively running processes combining a top-down load distribution and a bottom-up calculation of node-pair blocking in the network hierarchy, a steady state will be reached and approximate end to end call blocking probability in hierarchical wavelength routing optical network is obtained. Both the link independent model and the link correlation model are used in our analysis. By comparing theoretical results derived from our approximation method and results obtained from discrete event simulation, we illustrate that our method gives reasonable estimate of end to end call blocking probability in hierarchical wavelength routing optical networks, with the correlation model giving much more accurate estimate than the independent model.

In the study of dynamic routing and wavelength assignment (DRWA) algorithms and protocols in optical networks, blocking probability is a chief criterion for performance evaluation. But it does not always capture the full effect of a particular algorithm on other aspects, such as fairness problem, which refers to the variability in blocking probabilities experienced by connection requests between various node pairs. Most commercial routing protocols, such as OSPF, prefers to adopt the shortest path for a particular connection, also brings unfairness to diverse connection requests, and such an unfairness becomes more serious in optical networks without wavelength conversion due to wavelength continuity constraint. Some literature have concerned this fairness problem and proposed several fairness enhancement methods. But unfortunately, they will deteriorate the overall blocking probability. In this paper, we try to optimize the fairness and blocking probability simultaneously via the alternate routing strategy, and a novel classified alternate routing (CAR) approach will be adopted. In our investigations, we will consider the optimization of alternate path set determination strategy, wavelength assignment and path selection method, and link weight function, etc. Comparisons and collaborations with formerly proposed fairness enhancement methods will be also presented.

This paper provides an overview of recent efforts on the ASON standards for optical network control, including current status of ITU-T specifications and future work. A primary focus of the paper is also recent ASON-related testing in the industry, especially the OIF World Interop Demonstration held at SUPERCOMM 2004, where ASON signaling and routing architecture and protocols were tested across multiple vendors and carrier testing laboratories. Results were highly successful, indicating work on ASON, while not yet complete, is stable and implementable.

One of the challenges to dynamically set up a lightpath is to develop efficient control schemes in control as well as the management of the distributed-controlled optical networks. In networks with highly dynamic traffic loads, longer lightpath setup time may result in higher blocking probability and serious degradation of the network performance. In this paper, we propose a novel distributed signaling protocol designated as signaling with switching fabric status (SWFS). In this scheme, we consider the switch fabric configuration performing distributed signaling as well as routing and wavelength assignment. We forward the signaling request messages with available wavelength sets as well as the extra switching status for each wavelength at each node in the paths between the source-destination (S-D) pairs. The extra switching status indicates if the switch fabric should be reconfigured to provide internal port connection for specified wavelength at the node. When the signaling message reaches the destination, a wavelength with minimal reconfiguration times in the available wavelength set will be selected to shorten the connection establishment time. A distributed discrete event simulation platform has been developed to evaluate the performance of SWFS. The simulation results show that our proposed technology has the shorter setup time and lower blocking probability than the traditional non-optimization signaling technology.

The main purpose of this paper is to propose mobile IP network architecture based on dense wavelength-division multiplexing (DWDM), with GMPLS as a common control plane of the optical layer. A novel probability-based wavelength assignment (PWA) scheme for facilitating handoff in the proposed GMPLS-based mobile IP network is also presented. This paper studies the effects of network mobility and handoff path rerouting on the wavelength assignment scheme in a GMPLS-based mobile IP network. In order to support the same quality of service (QoS) for mobile services when handoff occurs, it is preferable to use the same wavelength in the newly established partial label switched path (LSP). In the proposed PWA scheme, when a mobile router (MR) begins to establish a new connection with the correspondent node, a wavelength is assigned based on the vacant probability of every wavelength in the new partial LSP after handoff. We use M/M/1(0) queue to model the dynamic traffic in a GMPLS-based mobile IP network and derive the close-form expression of wavelength vacant probability.

In future GMPLS-based wavelength-routed networks (WRNs), lightpaths are set up or torn down dynamically. Blocking probabilities of on-demand requests are an important performance measure. However, the network average blocking probability only indicates the global performance of the network. Due to large resource requirements and wavelength continuity constraint, the blocking probabilities of longer-hop connections are usually much higher than those of shorter-hop connections. This problem is the so-called fairness problem. In this paper, a priority-based control is proposed to improve the fairness between different connections. Connections with high blocking probabilities are assigned high priorities and they are able to preempt low priority connections with a predetermined probability. Both theoretical analyses and simulations are given. It is shown that the proposed control not only improves the fairness, but also ensures that the average blocking probability is not degraded much. Hence, the performance of the GMPLS-based optical networks will be improved greatly.

IP/DWDM optical Internet is considered to be a promising candidate for Next Generation Internet backbone. Fault-tolerant mechanisms have become critical technologies in IP/DWDM optical Internet. In this paper, based on preventive and reactive schemes, four fault-tolerant routing mechanisms in IP/DWDM optical Internet are proposed to deal with the problem of routing and wavelength assignment with dynamic single/multiple point failures under dynamic traffic demands, such as link/node/channel failures. The proposed mechanisms take load balancing into account, and a heuristic wavelength assignment algorithm is introduced to reduce the end-to-end delay. Simulation results have shown that the proposed fault-tolerant routing mechanisms are both feasible and effective.

This tutorial overviews optical-label switching technologies in light of packet, burst, and circuit switched networking applications. In particular, we will discuss architecture, protocol, and systems integration issues, and further provide possible network evolution scenarios from today’s circuit-oriented optical networks to tomorrow’s agile networks supporting various applications. Optical-label switching networks simultaneously support various applications optimized for circuit, burst, or packet switched networks on multiple optical channels. Systems applications of optical-label swapping, and optical-label switching routers have shown promising results including variable-length packet switching with 600 psec switching speed, 11-hop cascaded packet switching, 477 km field trials, and video transmission using optical-label switching core routers. Future interoperable networking with MPLS and legacy networks with GMPLS signaling are also discussed.

Optical packet switching techniques have been studied for many years and many feasibility experiments have demonstrated the viability of this technology at different areas of the network. In this paper an analysis of the optical technology potential in different network areas: metro access, metro core and backbone, is carried out. But the slow emergence of an integrated and available optical fast technology and a lack of standards is a brake for its introduction on the market place. In a society more and more user-centric approach, it becomes then urgent to raise the bottlenecks of a full electronic technologies and to propose hybrid architectures combining optics and electronics to reach all the specifications of the next generation of networks. So in this paper, after a status on optical packet switching techniques, after raising the needs in the metro and backbone areas, concrete solutions will be proposed, and more important some key catalysts for a rapid introduction will be addressed. Finally the paper will give the important steps to cross to really take this technological corner.

We have experimentally demonstrated some key techniques in label switching optical networks. These techniques include optical label generation, optical label swapping and preamble free clock recovery. The label generation using optical carrier suppression and separation (OCSS) technique has these advantages of high bit rate operation, low cross-talk and narrow bandwidth occupancy. The label swapping is realized by combining OCSS and wavelength conversion. The payload can be simultaneously regenerated after wavelength conversion; hence this technique is very useful for cascaded label swapping in label switched optical networks. We have also demonstrated how to realize preamble free clock recovery for burst mode label and payload generated by OCSS technique.

One of the major bottlenecks in conventional OBS networks is the processing delay of the control packet at each node. This is due to slow electronics of the node processors. In this paper, first we describe our proposed protocols of ultrafast optical path setup using an OC-labeled control packet. The OC-labeled control packet is sent out from the source node to destination node, in which the routing information is mapped onto OC-labels. At intermediate nodes the control packet is processed on the fly in optical domain, and available wavelengths are reserved for the transmission of the burst, which follows the control packet. Next, we will evaluate the performance of the proposed protocols, compared with conventional ones such as Just-Enough-Time (JET) protocol. Finally, we will present the experimental results, demonstrating the optical path setup using bipolar 8-chip-long OC-labeled control packet.

Optical burst switching (OBS) has been emerging as a promising technology that can effectively support the next generation IP-oriented transportation networks. JIT signaling protocol for OBS is relatively simple and easy to be implemented by hardware. This paper presented an effective scheme to implement the JIT protocol, which not only can effectively implement reservation and release of optical channels based on JIT, but also can process the failure of channel reservation and release due to loss of burst control packets. The scheme includes: (1) a BHP (burst head packet) path table is designed and built at each OBS node. It is used to guarantee the corresponding burst control packet, i.e. BHP, BEP (burst end packet) and BEP_ACK (BEP acknowledgement), to be transmitted in the same path. (2) The timed retransmission of BEP and the reversed deletion of the item in BHP path tables triggered by the corresponding BEP_ACK are combined to solve the problems caused by the loss of the signaling messages in channel reservation and release process. (3) Burst head packets and BEP_ACK are transmitted using “best-effort” method. Related signaling messages and their formats for the proposed scheme are also given.

Superstructured fiber Bragg grating (SSFBG) can be applied as the optical code (OC) encoder/decoder for optical code division access (OCDMA) and optical packet switching (OPS) networks. OC-based traffic (either OCDM traffic or OC-based packet labels) could be transmitted over the WDM networks to provide higher spectral efficiency and better scalability. The performance of this scheme is investigated for the first time in this paper, particularly the impact of the crosstalk in WDM networks with different channel spacing. These results can be used for assigning the optimal channel-spacing to get the highest spectral efficiency.

We discuss how all-optical signal processing might play a role in future all-optical packet switched networks. We describe a few approaches to optical header processing, all based on nonlinearities in a semiconductor optical amplifier. In first approach a SLALOM configuration is used. The second approach uses a Terahertz Optical Asymmetric Demultiplexer. We also describe a header pre-processor to improve the extinction ratio of the header processor output. The second functional block on which we focus is optical buffering. We show how all-optical signal processing technology can be used to route a packet into a fiber delay line and we describe a circulating optical loop based op optical technology.

Virtual Private LAN Service (VPLS) is generating considerable interest with enterprises and service providers as it offers multipoint transparent LAN service (TLS) over MPLS networks. This paper describes an effective technology - VPLS, which links virtual switch instances (VSIs) through MPLS to form an emulated Ethernet switch and build Scalable Transparent Lan Services. It first focuses on the architecture of VPLS with Ethernet bridging technique at the edge and MPLS at the core, then it tries to elucidate the data forwarding mechanism within VPLS domain, including learning and aging MAC addresses on a per LSP basis, flooding of unknown frames and replication for unknown, multicast, and broadcast frames. The loop-avoidance mechanism, known as split horizon forwarding, is also analyzed. Another important aspect of VPLS service is its basic operation, including autodiscovery and signaling, is discussed. From the perspective of efficiency and scalability the paper compares two important signaling mechanism, BGP and LDP, which are used to set up a PW between the PEs and bind the PWs to a particular VSI. With the extension of VPLS and the increase of full mesh of PWs between PE devices (n*(n-1)/2 PWs in all, a n² complete problem), VPLS instance could have a large number of remote PE associations, resulting in an inefficient use of network bandwidth and system resources as the ingress PE has to replicate each frame and append MPLS labels for remote PE. So the latter part of this paper focuses on the scalability issue: the Hierarchical VPLS. Within the architecture of HVPLS, this paper addresses two ways to cope with a possibly large number of MAC addresses, which make VPLS operate more efficiently.

As a promising solution for next-generation broadband access networks, EPON could provide full-service access such as voice, video and data applications. However, EPON’s standard IEEE 802.3ah does not specify a particular supporting mechanism to guarantee QoS and priority requirements of various services, allowing it to be vendor specific. Meanwhile, how to segregate user traffic to guarantee security, remains unsolved. This paper creatively introduced the 802.1Q VLAN (Virtual Local Area Network) technique into the EPON system to solve these problems. Firstly, a brief introduction of EPON system is given. Secondly, the VLAN solution is presented in detail. Unlike VLAN mapping according to port or MAC in Gigabit Ethernet, EPON’s VLAN mapping is based on LLID tag. At last, OLT MAC layer design is given and FPGA implementation is described in detail. Detailed simulation experiments have been conducted to study the performance and validate the effectiveness of the proposed mechanism.

Following the development of the high way, the communication system of rescue for the high way is becoming more and more important. A new PON technology for the emergent telephone system of the high way is proposed. A concrete solution for this kind of network is also illustrated. Besides, system cover-range is calculated theoretically. Our project makes full use of the advantage of PON technology, and deduces the cost of installation and system maintenance for the first time. The application of this project demonstrates that it can offer a more efficient, safer and reliable solution for high way rescue.

The next generation home network that will accommodate various services with different quality requirements is introduced. Technical study items for realizing this network are described. Current technologies are overviewed to clarify what will need further consideration.

Recent years, data traffic is increasing at an unprecedented rate. Digital Subscriber Line (DSL) and cable modem (CM) networks will not afford the huge bandwidth consuming in the near future. Ethernet Passive Optical Network (EPON) is an inexpensive, simple, scalable, and capable of delivering bundled voice, data and video services to an end-user over a single network. EPON, which is a typical point to multi point topology, has become an attractive technology to both vendors and carriers. To ensure compatibility with the IEEE 802 architecture, Logical Link Identifier (LLID) tag is introduced in EPON to achieve point to point emulation. Usually one LLID tag associates with one stream or one ONU (Optical Network Unit) which is remote device of EPON. In order to simplify operations, LLID tag is added as part of preamble in Optical Line Terminal (OLT) Reconciliation Sublayer (RS) and is checked at ONU side. It’s urgent to have a powerful mechanism to achieve dynamic LLID filtering. Just in the downstream, each packet must have its own LLID according to the Destination Address field before being sent; in the upstream, LLID is learned based on the Source Address Field of the packet. Although address filtering is very popular in the L2 switch chips, it can’t be applied to EPON LLID filtering directly, owing to EPON is neither a point to point network nor a broadcast network. In this paper, the main principles and features of Logical Link Identifier (LLID) tagging system are firstly explained, and then the process of LLID both in OLT and in ONU side is described. What’s more, a novel method to realize dynamic LLID filtering is discussed in detail, involving in LLID scanning, LLID learning, LLID aging, and so on. An erasable Content Addressable Memory (CAM) structure is used to meet this demand. Not only the architecture of the concrete circuits, but the related timing sequences, operation process, service arbitration and aging state transition diagram are presented. Note that this mechanism has been verified in the actual EPON system and worked very well. At last, some conclusion remarks about using LLID filtering to enhance the performance of EPON is given.

Ethernet Passive Optical Network (EPON), which represents the convergence of low-cost Ethernet equipment and reliable fiber infrastructure, appears to be the best candidate for the next-generation broadband access network. A coin has two sides. Although EPON inherits the benefits of Ethernet, it also suffers some disadvantages. With its origin as a LAN technology, where “best effort” delivery and 99.96% uptime were acceptable, Ethernet was never considered to be a candidate for achieving “carrier-class” status, mainly because lack of powerful OAM capabilities. That’s to say OAM function is essential to carrier-class EPON. In this study, the principals of EPON and OAM, together with the necessity of OAM, are presented firstly. Besides, different ways to provide OAM in ITU-T APON, GPON are analyzed respectively. Three means to obtain EPON OAM, using MAC Control Frame, Physical Link Service Overhead Coding Sublayer (PLSOCS) and Reconciliation Sublayer (RS) Preamble are proposed in detail. Moreover, the main functions, such as remote failure indication, remote loop-back, link monitoring and variable retrieval are studied subsequently. After that, the concrete implementation of OAM in EPON is addressed, combining with hardware and software. As a result, an effective way to carry out EPON OAM can be easily drawn.

As a promising solution for the next generation broadband access network, Ethernet passive optical networks (EPONs) are being designed to deliver multiple services and applications, such as voice communications, video conferencing, real-time transactions and new evolving bandwidth-consuming applications. To support these applications with their various requirements, Quality-of-Service (QoS) management over EPON is required. In this study, we present a novel scheduling scheme, hierarchical weighted bandwidth guaranteed (HWBG) scheme, which could effectively support QoS. Simulation results validate that the proposed scheme can achieve high efficiency and improve packet delay performance.

Normal GIGA ETHERNET continuously transmits or receives 8B/10B codes including data codes, idle codes or configuration information. In ETHERNET network, one computer links one port of switch through CAT5 and that is OK. But for EPON, it is different. All ONUs share one fiber in upstream, if we inherit the GIGA ETHERNET PHY, then collision will occur. All ONUs always transmit 8B/10B codes, and the optical signal will overlay. The OLT will receive the fault information. So we need a novel EPON PHY instead of ETHERNET PHY. But its major function is compatible with ETHERNET’s. In this article, first, the function of PCS sub layer is discussed and a novel PCS module is given that can be used in not only EPON system but also in GIGA ETHERNET system. The design of PCS is based on 1000BASE-X PCS technology. And the function of 1000BASE-X PCS should be accomplished first. Next we modify the design in order to meet the requirements of EPON system. In the new design, the auto negotiation and synchronization is the same to the 1000 BASE-X technology.

Nowadays, EPON technology is regarded as one of the best candidate technology for next generation broadband access technology. In this paper, the architecture of EPON is introduced first. The upstream in EPON adopts TDMA/TDM technology, and it is necessary to know the exact RTT(Round Trip Time), the principle of RTT measurement (ranging) and RTT compensation are introduced. We use Xilinx FPGA Spartan^II chip series to realize ranging function module, the main idea of implementation is given in this paper, and this can be used for further study of hardware implementation.

With the tremendous growth of broadband IP services, the impact of bandwidth bottleneck in access network becomes more prominent and largely restricts the network carriers to provide satisfied services for subscribers. Thus, a new access network with sufficient bandwidth resources and efficient network resource management must be designed. In this paper, a Generalized MPLS (GMPLS) and EPON-based optical access network is proposed to provide end-to-end quality-of-service (QoS) guarantee for broadband IP services. The proposed network architecture is presented and the Label Switched Path (LSP) setup procedure is developed. In the proposed access network, GMPLS is used to manage network resources dynamically and to conduct traffic engineering. In addition, an efficient dynamic bandwidth allocation algorithm is designed to accommodate bursty IP traffic and improve bandwidth utilization in the proposed network. The performance analyses show that under the same network load the proposed scheme has lower average packet delay and shorter queue length than current schemes.

Class-based service differentiation is provided in DiffServ networks. However, this differentiation will be disordered under dynamic traffic loads due to the fixed weighted scheduling. An adaptive weighted scheduling scheme is proposed in this paper to achieve fair bandwidth allocation among different service classes. In this scheme, the number of active flows and the subscribed bandwidth are estimated based on the measurement of local queue metrics, then the scheduling weights of each service class are adjusted for the per-flow fairness of excess bandwidth allocation. This adaptive scheme can be combined with any weighted scheduling algorithm. Simulation results show that, comparing with fixed weighted scheduling, it effectively improve the fairness of excess bandwidth allocation.

Wavelength division multiplexing (WDM) has been considered a promising transmission technology in optical communication networks. With the continuous advance in optical technology, WDM network will play an important role in wide area backbone networks. Optical wavelength switching, compared with optical packet switching, is a more mature and more cost-effective choice for optical switching technologies. Besides, the technology of time division multiplexing in optical communication networks has been working smoothly for a long time. In the proposed research, the problem of multicast groups aggregation and multicast routing and wavelength assignment in wavelength-routed WDM network is studied. The optical cross connect switches in the problem are assumed to have limited optical multicast/splitting and TDM functionalities. Given the physical network topology and capacity, the objective is to maximize the total revenue by means of utmost merging multicast groups into larger macro-groups. The groups in the same macro-group will share a multicast tree to conduct data transmission. The problem is formulated as an optimization problem, where the objective function is to maximize the total revenue subject to capacity constraints of components in the optical network, wavelength continuity constraints, and tree topology constraints. The decision variables in the formulations include the merging results between groups, multicast tree routing assignment and wavelength assignment. The basic approach to the algorithm development for this model is Lagrangean relaxation in conjunction with a number of optimization techniques. In computational experiments, the proposed algorithms are evaluated on different network topologies and perform efficiently and effectively according to the experiment results.

This paper investigates the function modules in ASON model, and analyzes the strategic planning process and the fundamental planning process in the ASON plan. To facilitate evaluation algorithms and technologies, we develop a visual simulation system for ASON modelling and planning and give a simple example of ASON planning case.

In this paper, we evaluate link-state update triggers in wavelength-routed networks. We consider two popular triggers, namely the Absolute Threshold Trigger and the Relative Threshold Trigger. The blocking performance of the two triggers is confirmed to be highly dependent on traffic load. Moreover, it is observed that under light traffic load, blocking performance degrades significantly once the threshold is shifted away from advertising every single link state change immediately. We expect that the above observation would be of significant importance and would impose a critical challenge to the future development of link-state update methods.

A novel technique for optical labeling of payloads and labels using optical single-sideband modulation is proposed. This method can effectively reduce the bandwidth of optical signal and utilize the low frequency RF devices to generate the labels. And we also investigate the power and phase character of this scheme. Simulation results show that this scheme is robust to the phase differences of the single-sideband payload and label. After fiber transmission and label swapping, the power penalties of payload and label are both less than 1dB. Therefore this scheme is suitable for 40Gb/s optical label switching networks.

Slotted WDM, which achieves higher capacity compared with conventional WDM and SDH networks, has been discussed a lot recently. The ring network for this architecture has been demonstrated experimentally. In slotted WDM ring network, each node is equipped with a wavelength-tunable transmitter and a fixed receiver and assigned with a specific wavelength. A node can send data to every other node by tuning wavelength accordingly in a time slot. One of the important issues for it is scheduling. Scheduling of it can be reduced to input queued switch when synchronization and propagation are solved and many schemes have been proposed to solve these two issues. However, it’s proved that scheduling of such a network taking both jitter and throughput into consideration is NP hard. Greedy algorithm has been proposed to solve it before. The main contribution of this paper lies in a novel genetic algorithm to obtain optimal or near optimal value of this specific NP hard problem. We devise problem specific chromosome codes, fitness function, crossover and mutation operations. Experimental results show that our GA provides better performances in terms of throughput and jitter than a greedy heuristic.

PONs are gaining popularity in areas with high population density. Extensive deployment of fiber plant, along with the introduction of cost effective network devices promise ample of affordable bandwidth to the home and the premise. The distribution network is typically organized as tree architecture and utilizes simple optical splitters. The use of fixed splitters results, however, in a rigid topology, incapable of changes or expansion to meet the needs of demographic changes. A new breed of splitters, based on Planar Lightwave Circuit (PLC) technology, allows service providers to programmatically change the split rations. This feature facilitates a phased build-out of the PON network or changes to the network topology. This article describes the underlying technology for these programmable splitters and explains how they can add flexibility to PON networks.

The major goal of TE (traffic engineering) is to facilitate efficient and reliable network operations while simultaneously optimizing network utilization and traffic performance. Therefore it is necessary to introduce TE mechanism to ISPs’ (Internet Service Providers’) networks with the purpose of improving network performance and reducing costs. In this paper, a novel TE mechanism based on GMPLS (Generalized Multi-Protocol Label Switching) is proposed and analyzed in detail, and is demonstrated on a network with IP over ASON (automatically switched optical network) architecture. The ASON layer acts as the server of IP layer which uses traditional IP protocols, and realizes TE mechanism through the method of dynamically changing bandwidth seen by IP layer through UNI (User-Network Interface). Simulation results have shown that the model has a superior operational agility to the conventional method and lower congestion probability in certain conditions.

This paper discusses the integrated routing algorithm for IP over WDM network, and presents a new online band priority algorithm (BPA) using the hierarchical multiplexing scheme. Based on the integrated routing model, where both IP layer and optical layer share full routing information with each other, the BPA treats the waveband as the grooming unit to adapt to the hierarchical OXC architecture. The BPA introduces a waveband grooming related cost function to obtain the optimal route. Simulation results show that the proposed BPA achieves a comparatively low blocking probability for online routing.

This paper describes a way of creating the OVPN through the management plane and manages the OVPN with a centralized method in Automatic Switch Optical Network (ASON). Through the extension of resource management, signal mechanism and connection control, the architecture of OVPN creation and OVPN maintenance with a centralized method is efficiently realized, and the requirements of multi- granularity, bandwidth modification and link monitory are realized. Moreover, Service Lever Agreement (SLA) can be included in OVPN through Switched Connection (SC) and Soft permanent connection (SPC). If further study is carried through on this scenario, the optical multicast will be realized under both the mechanisms of management plane and control plane. At last, the applications area, account model and development prospect of OVPN are discussed.

In the parallel processing system, large numbers of processors are interconnected in order to improve the performance of the computer, such as the symmetric multiprocessor (SMP) architecture. When the basic node is an SMP or a computer having a single processor, the characteristics of an interconnection networks are important factors which influence the performance of the entire system. Fibre Channel (FC) has a lot advantages, such as excellent scalability; the bandwidth is large; delay time is short and fault tolerance is large. It is assumed that an SMP is used for a basic node. We construct the cluster system using FC as interconnection network, which are a fabric method and a FC Arbitrated Loop (FC-AL) method. According the method, if the number of nodes supported by the interconnection network is small, the addition of extra nodes can be added at small expense. The bandwidth of each node is large, the delay time is short, and the fault tolerance effect is large in the interconnection network. In the case of connecting to a shared disk, a large bandwidth is provided and time required for gaining access to the shared disk becomes short.

This paper presents a multicasting extension to OIF User Network Interface (UNI) standard, which allows dynamically setup/teardown/graft/prune a transport network multicasting tree. Transport network multicasting tree means a point-to-multipoint connection that originates from an ingress Transport Network Element (TNE) port and end at multiple egress TNE ports across the transport network for transporting user services. A global multicasting tree ID is defined to identify the transport network multicasting tree. Without addition of any new abstract messages, we just add some new attributes to existing UNI abstract messages to implement the signaling actions needed to support optical multicasting. Experimental results show the performance of our proposal under different strategy.

In this paper we propose a new concept of future service-aware transport networks that are built on converged wireless-wired transport networks, enable service layer convergence, and also support application layer convergence. This concept is derived from the emerging trend of business convergence among the telecommunication, cable, and entertainment segments, and technology convergence among the telecommunications, computer, and software industries. We also identify QoS as a technical backbone to future service-aware transport networks, and from the perspective of an integrated carrier having both wireless and wired assets, discuss related research opportunities and challenges, especially those involved in optical networks. We finally present an adaptive apparatus for high availability, capacity efficiency, and QoS-guaranteed protection and restoration for optical networks. This paper is intended to promote academic and industrial attention to some core technical challenges that lie ahead for the telecommunication industry and to spur strategically important research and perhaps also some standards activities.

Traditional metro network architectures comprise multiple layers of networking equipment supporting a wide array of services and packet-oriented applications. Among others, these include WDM, SDH, ATM, Ethernet and IP, each requiring its own network elements and associated management solutions to perform its own independent networking functions. While these work well individually, the combined network is cumbersome and inefficient. Recent advancements in network technologies are now changing the way metro networks are designed. Multi-functional consolidation through technology integration and the standardization of protocol inter-networking methods are leading to a converged network solution in support of a diverse set of packet-aware service offerings. This presentation will explore new technologies that are enabling convergence in the metro network, both across layers and across services.

This paper discusses the innovation in network architectures, and optical transport, that enables metropolitan networks to cost-effectively scale to hundreds Gb/s of capacity, and to hundreds km of reach, and to also meet the diverse service needs of enterprise and residential applications. A converged metro network, where Ethernet/IP services, and traditional TDM traffic operate over an intelligent WDM transport layer is increasingly becoming the most attractive architecture addressing the primary need of network operators for significantly improved capital and operational network cost. At the same time, this converged network has to leverage advanced technology, and introduce intelligence in order to significantly improve the deployment and manageability of WDM transport. The most important system advancements and the associated technology innovations that enhance the cost-effectiveness of metropolitan optical networks are being reviewed.

Optical data transmission has undergone a tremendous evolution. Starting with unrepeated point-to-point transmission in the 80s the inventions of wavelength division multiplexing (WDM) and erbium doped fiber amplifiers (EDFAs) have let to an explosion of system capacity as well as of system reach. After the steep downturn of recent years network operators have now regained the strength to upgrade their networks and to implement new services. This paper will review current and upcoming technologies in the long haul (LH) and ultra long haul (ULH) data transmission. It will further discuss the future evolution of transparent optical networks towards dynamically routed meshed optical networks with respect to operator’s technical operational and economical requirements. Upgradeability turns out as a key issue as it on the one hand side facilitates low front investments for network providers and on the other hand side enables organic and flexible network growth.

Advanced network planning concerns effective network-resource allocation for dynamic and open business environment. Planning methodologies of ASON implementation based on qualitative analysis and mathematical modeling are presented in this paper. The methodology includes method of rationalizing technology and architecture, building network and nodal models, and developing dynamic programming for multi-period deployment. The multi-layered nodal architecture proposed here can accommodate various nodal configurations for a multi-plane optical network and the network modeling presented here computes the required network elements for optimizing resource allocation.

This paper concentrates on solutions for next-generation IP (Internet Protocol) backbone networks. This is a key issue for many network operators since IP will be the dominating network layer technology on which an ever increasing number of applications with growing bandwidth requirements will be based. Without new network solutions, this trend would lead to a strong increase in number and size of IP routers while already today’s requirements make it difficult to realise large-scale IP backbone networks in a stable and cost-efficient way. This paper investigates the benefits that an appropriate transport network based on ASTN (Automatic Switched Transport Network) and OTN (Optical Transport Network) technology can bring to future IP backbone networks - providing the stable basis on which next generation IP networks can be built. Network modelling is used to show that in addition to qualitative benefits a transport network based IP backbone solution can lead to a significant reduction of network equipment cost.

In designing future networks and systems, we must consider the current paradigm changes; IP convergence and the divergence of architectures and technologies. To start with, I will discuss broadband access technologies being deployed in different countries. Harnessing the full power of light will spur the creation of new broadband and ubiquitous services networks. The key requirements of enhancing the performance and reducing the cost of future IP-based multimedia communication networks can be effectively achieved by exploiting wavelength routing. This requires, however, not only the optimization of photonic technologies but also coordination with complementary electrical technologies. MPLS will play the key role, and its status and the present perception of its inefficiency are discussed. The evolution of GMPLS and the way to be taken for further development are depicted. Finally, I discuss some of the cutting-edge photonic transport technologies that will enable us to achieve another advance in network performance in the future.

Ethernet Passive Optical Network (EPON), which combines the low cost Ethernet equipment and economic fiber infrastructure, is being considered as a promising solution for Fiber-To-The-Home (FTTH). However, since EPON is an optical shared medium network, some unique features make it more vulnerable to security attacks. In this paper, the key security threats of EPON are firstly analyzed. And then, considering some specific properties which might be utilized for security, such as the safety of transmissions in upstream direction, some novel methods are presented to solve security problems. Firstly, based on some modification about registration, the mechanism of access control is achieved. Secondly, we implement an AES-128 symmetrical encryption and decryption in the EPON system. The AES-128 algorithm can process data blocks of 128 bits, but the length of Ethernet frame is variable. How to deal with the last block, which is not up to 128 bits, is discussed in detail. Finally, key update is accomplished through a vendor specific OAM frame in order to enhance the level of security. The proposed mechanism will remain in conformance with P2MP specification defined by 802.3ah TF, and can supply a complete security solution for EPON.

In dynamic networks, the failure detection time takes a major part of the convergence time, which is an important network performance index. To detect a node or link failure in the network, traditional protocols, like Hello protocol in OSPF or RSVP, exchanges keep-alive messages between neighboring nodes to keep track of the link/node state. But by default settings, it can get a minimum detection time in the measure of dozens of seconds, which can not meet the demands of fast network convergence and failure recovery. When configuring the related parameters to reduce the detection time, there will be notable instability problems. In this paper, we analyzed the problem and designed a new failure detection algorithm to reduce the network overhead of detection signaling. Through our experiment we found it is effective to enhance the stability by implicitly acknowledge other signaling messages as keep-alive messages. We conducted our proposal and the previous approaches on the ASON test-bed. The experimental results show that our algorithm gives better performances than previous schemes in about an order magnitude reduction of both false failure alarms and queuing delay to other messages, especially under light traffic load.

In this paper, we introduce that we are developing a software platform with generalized multi-protocol label switching (GMPLS) -based Intelligence Optical Networks (ION) capacity. The survivability was the crucial character of optical transport network. The optical network is transmitting from static transport network to a more dynamic, flexible network. So the intelligent optical network (ION) has been proposed. IUT-T standardizes the automatically switched optical network (ASON) on G.8080 which defines the architecture. The ASON simulation software platform is based on GMPLS. The protection and restoration mechanisms specified in the next. Our simulation architecture is illustrated and the recovery time is analyzed.

We introduce a new survivable network concept called the “Generalized Survivable Network” (GSN), which has the special property that it remains survivable no matter how traffic is provisioned dynamically, as long as the input and output constraints at the nodes are fixed. We study a network synthesis problem called the Virtual Topology Mapping Problem (VTMP), which aims at finding the edge capacities for a given physical network topology with the I/O constraints at the nodes that will make it a GSN. Two concepts, tau-matization and sigma-tization are introduced. Tau-matization refers to the conversion of any physical network into a non-blocking network. Sigma-tization refers to the conversion of a multi-edge-connected network into a fully survivable network with the necessary edge capacities. We develop a basic framework for solving the VTMP problem of GSN. We obtain a lower bounding procedure, an upper bounding procedure and a heuristic called MMRA for solving the tau-matization problem. The lower bounding procedure has a very nice analytic form that can be solved by the simplex method in polynomial time. Some experimental results are given. We have shown that the actual cost of developing a GSN with the MMRA heuristic is within 10 - 35% from the absolute lower bound and much less than the tremendous cost that is generally thought would be needed to realize a GSN (a reduction of 85% to 95% from the upper bound). The framework is applicable to ASTN / ASON survivable network planning and bandwidth-on-demand resource allocation.

Optical packet-switched networks (OPNs), which consist of edge nodes and core nodes, are regarded as one of the most promising solutions to meet the increasing service requirements of the global Internet and to utilize efficiently the huge bandwidth of dense wavelength division multiplexing (DWDM) technology. In particular, the edge nodes perform adaptation function which interfaces between the legacy packet networks (i.e. IP networks) and OPNs. Namely, the edge nodes assembly IP packets from electronic layer into a larger optical packet. The function of packet assembly has become an important issue to be addressed in OPNs, since it not only increases the bandwidth utilization, but also reduces the self-similarity of the traffic. Studied in this paper are three kinds of typical packet assembly schemes in unslotted OPNs, i.e. the timer-based one, the length-threshold one and the hybrid algorithm. Based on a generic architecture of the edge nodes, these assembly mechanisms are examined by simulations under various self-similar traffic loads. The impacts of these algorithms on the traffic, including statistics of generated optical packets and self-similarity of optical packet traffic, are compared. The simulation results indicate that optical packet assembly can fulfill traffic shaping by improving key traffic characteristics of interest in network performance evaluation, such as the long-term correlation and so on. Moreover, the hybrid algorithm proves to be a better solution when both the assembly time and the packetisation efficiency have been also considered at the same time.

The first optical burst switching (OBS) system has been demonstrated in China, which includes three edge routers and one core-node. A kind of fast wavelength selective optical switching was used in the system. The core OBS node consists of a kind of wavelength selective optical switch we developed. It consists of two SOA switches and one wavelength selective thin film filter with centre wavelength at one wavelength. There are one input optical fiber and two output fibers, each fiber carries two wavelengths. The Dell PE2650 servers act as the edge OBS routers. The wavelength of each data channel is located in C-band and the bit rate is at 1.25Gbps. The control channel uses bit rate of 100Mbps at wavelength of 1310 nm. A novel effective scheme for Just-In-Time (JIT) protocol was proposed and implemented. OBS services, such as Video on Demand (VOD) and file transfer protocol (FTP), have been demonstrated. Assembling and scheduling methods that are capable to guarantee the QoS (quality of service) of the transported service are studied.

In this paper, we propose a novel modified architecture of Optical Burst-Switched Networks, called Reliable Optical Burst Switching (ROBS), which introduces congestion control and retransmission mechanisms of TCP to OBS layer. Design scheme of edge and core nodes for ROBS are both demonstrated. Part of the ROBS function has been implemented on ns-2 platform. Performance improvements due to burst retransmission are simulated and presented.

By eliminating optical synchronizers in optical IP routers, more complex scheduling algorithms are needed to schedule asynchronous packets. This will result in voids at the switch output thus reducing switch throughput. A novel scheduling algorithm had previously been proposed to reduce these excess losses by filling voids caused by asynchronous and variable length operation. Non-degenerate (i.e. non-uniform) buffer depth has previously been shown to improve the packet loss performance especially under bursty traffic. We investigated the performance of the void filling algorithm by combining non-degenerate and degenerate (uniform) fiber delay lines. Performance is studied for different threshold levels, i.e. the number of uniform delay lines in the feedback delay line before introducing the non-degenerate delay lines. The packet loss performance for combined non-degenerate and degenerate delay lines with void filling algorithm is presented for an optical router with a feedback delay lines buffer under self similar traffic. The recirculating delay lines buffer emulates a two-stage buffer where first stage buffers smaller packets whilst the second stage stores larger packets. This buffering mechanism is similar to SLOB (i.e. Switch With Large Optical Buffer).

Optical burst switching (OBS) provides a feasible paradigm for IP over DWDM. And one of key issues in OBS is the fast and efficient scheduling of bursts with different QoS demands. In this paper, we propose a novel window-based scheduling (WS) algorithm, in which all bursts arriving in the window are scheduled together based on priority BORA. Therefore batching scheduling can achieve an optimized BORA efficiency and then the number of contention in OBS can be reduced largely, and simultaneously priority scheduling in the window can efficiently enhance QoS. Simulation results show that the proposed WS algorithm with the optimal parameters such as window and max-delay, compared with BORA, can achieve more than one order blocking probability reduction for high priority bursts and approximately 3.5% improvement in bandwidth utilization at 0.8 load.

Optical burst switching (OBS) has been paid increasing attention as an attractive alternative for building future photonic Internet. Besides pure theoretical research, experimental prototype and proof-of-concept demonstration are also significant parts toward practical burst switched WDM optical networks. This paper describes hardware architecture, software module and experimental results of optical burst switching nodes using a just-enough-time (JET) protocol and contention resolutions. The core node forwards bursts in the optical domain with deflection routing resolution. It is designed to be transparent to the format, the protocol and the bit rate of traffic. The edge node performs burst assembly/disassembly, as well as burst forwarding, which makes it flexible to be applied to various topologies. The related control protocols and algorithms are implemented taking into account characteristics of being used commercial off-the-shelf devices. By the experiments, the basic function of optical burst switching is demonstrated, the delay measurement result is analyzed, and the effort direction is given.

Optical Burst Switching (OBS) is a promising solution for realizing IP-over-WDM. However, OBS has its disadvantages involving in high burst lost ratio and low channel utilization. In this paper, a new Hybrid Optical Burst Switching (H-OBS) is proposed, and the H-OBS with a ring is investigated with numerical simulation. Simulation results indicate that the packet lost ratio in H-OBS is less than that in OBS, and its performance depends on some parameters such as traffic volume, the packet proportion (PP), the electronic buffer size and the processing speed in edge node and core node.

The use of optical buffering (OB) and wavelength converters (WC) is recognized as essential for reducing the complexity of optical packet switches (OPS). The object of this study is to investigate the performances of optical packet switches in which the wavelength converters are shared per input channel (SPIC) and optical buffering are partially shared working in unicast and multicast ways through analyzing a traffic model. We compare the packet loss probability of the unicast case and the multicast case, as the number of converters and depth of buffering in them are fixed. We also analysis of the improved performance by increase the number of wavelength in each channel.

The 300-pin Multi Source Agreement (MSA) and other MSAs provide basic requirements from a transponder or transceiver used in 10Gb/s optical networks. These MSAs typically address a wide range of applications, including: SONET/SDH, 10GbE and 10GFC for Metro, long-haul (LH) and ultra-long-haul (ULH) networks. Nonetheless, being a basic standard, the 300-pin MSA addresses the minimal required specifications set and does not cover the whole set of requirements and applications that system vendors are interested in. For example, widely tunable and extended reach transponders are not included in the 300-pin MSA. Chromatic dispersion is one of the major reach limiting factors in optical networks. In reconfigurable optical networks, chunks of DWDM channels may travel through different routes and therefore require tunable dispersion compensation. In static ULH optical networks, the number of dispersion compensation fibers (DCFs) dictates the amount of residual chromatic dispersion. This residual chromatic dispersion differs from one DWDM channel to the other. Unless it is compensated at the receiver, it further restricts the link length and reduces the distance between one regenerator to the other. This results in shorter links and more O-E-O blocks, which dramatically increases the cost of the network. This paper discusses a specially designed optical dispersion compensation (ODC) device that is packaged in a standard butterfly package and can fit into a 300-pin MSA transponder. A transponder with the proposed ODC can still satisfy all the basic requirements that are described in the 300-pin MSA while providing improved chromatic dispersion tolerance.

For the application in Passive Optical Network (PON), a novel design of OCDMA system scheme is proposed in this paper. There are two key components included in this scheme: a new kind of OCDMA encoder/decoder system based on TCM and FFH and an improved Optical Line Terminal (OLT) receiving system with improved anti-interference performance by the use of Long Period Fiber Grating (LPFG). In the encoder/decoder system, Trellis Coded Modulation (TCM) encoder is applied in front of the FFH modulator. Original signal firstly is encoded through TCM encoder, and then the redundant code out of the TCM encoder will be mapped into one of the FFH modulation signal subsets for transmission. On the receiver (decoder) side, transmitting signal is demodulated through FFH and decoded by trellis decoder. Owing to the fact that high coding gain can be acquired by TCM without adding transmitting band and reducing transmitting speed, TCM is utilized to ameliorate bit error performance and reduce multi-user interference. In the OLT receiving system, EDFA and LPFG are placed in front of decoder to get excellent gain flatness on a large bandwidth, and Optical Hard Limiter (OHL) is also deployed to improve detection performance, through which the anti-interference performance of receiving system can be greatly enhanced. At the same time, some software is used to simulate the system performance for further analysis and authentication. The related work in this paper provides a valuable reference to the research.

Nowadays multi-protocol label switching (MPLS) technology has achieved great popularity in the world. Because MPLS adopts the scheme of “route once and switch many”, routing is the key element to ensure packets are transmitted to destinations. Though there are many designs of MPLS routing protocol for public networks, they can hardly fit the private network very well for its special characteristics. This paper presents the design of MPLS routing protocol in the private network using the specification and description language (SDL), which is an object-oriented formal language and mainly used in communications area. In our design the whole autonomous system (AS) is regarded as a single area to reduce the number of protocol packets. The SDL module of MPLS routing protocol consists of four kinds of processes, namely Management, Checking, NeigSession and LSADbase. Simulations of the SDL design are run in many scenarios, the results of which indicate that the SDL module is manageable and the overall performance is satisfying. From our design and simulation experience, we find that the software development cycle can be drastically reduced by the use of SDL, and the code is easy to maintain and reuse.

In this work we examine various survivability schemes for an agile all-photonic backbone network (AAPN). The AAPN architecture has been proposed by the telecommunication industry as a possible candidate for the ultra high speed Next Generation Optical Network (NGON) architecture, and implements an adaptive optical core and edge routers or switches in a star formation. In this paper we study the issue of network survivability in AAPN architecture. We examine different choices for edge-switch traffic restoration, including 1+1 path protection, cycle-based approaches and protection trees, and their pros and cons. We have also analyzed and compared the performance of the proposed schemes.

As an alternative of the Shared Backup Path Protection (SBPP) method, we develop a framework for dynamic provisioning of survivable services based on the use of p-cycles to form a Protected Working Capacity Envelope (PWCE) within which dynamic provisioning of protected services is greatly simplified. Based on p-cycles, the restoration speed of rings is obtained, but with the capacity efficiency of shared-mesh networks. In addition, with PWCE, arbitrarily fast dynamic service demands can be handled with much less complexity (in terms of database dependency and state update dissemination) than under SBPP. Only a simple OSPF-topology view of non-exhausted spans in the envelope is required. If a new path can be routed through the envelope, it is protected by virtue of being routable. This is in contrast to needing a full database of network state so that the end-user can set up a shared backup protection path under SBPP. In addition, dissemination of state updates occurs only on the time-scale of the non-stationary evolution of the demand statistics, not on the time-scale of individual connections. During statistically stationary periods, there is no dissemination of state updates whatsoever with an envelope that is well matched to its load. The PWCE concept thus offers some new tradeoffs between operational simplicity and spare capacity efficiency. The main contribution of this work is the detailed implementation and simulation of test networks operating under PWCE and designed with novel envelope volume maximizing formulations.

With the help of an independent control plane, ASON introduces lots of new intelligent functions into the traditional transport network. We use some special protocols such as signaling and routing protocols in control plane instead of network management protocol that only has few primitives to control the network elements in transport plane. During the service provision of ASON, the coordination of management plane (MP) and control plane (CP) is very important. This paper analyses the coordination of MP and CP of ASON. At first, it introduces the meaning of coordination between MP and CP and its classification. Secondly, it puts emphasis upon analyzing traffic engineering (TE) realized through the coordination of MP and CP. In this part, we modify the basic ASON routing architecture and introduce the coordination functions of MP. Thirdly, this paper analyses TE control processes of the coordination between those two planes during the services provision which has not been considered in ASON recommendations but is very important.

An ASTN (Automatic Switched Transport Network) enhances the traditional transport network by a control plane. The control plane implements signalling and routing mechanisms that allow the automatic configuration of end-to-end connections on client requests. Consequently the management of ASTNs requires integrated management of the control plane layer and the underlying transport network. This article describes a generic architecture for ASON management systems tuned to this requirement. The centre piece of this architecture is a generic CORBA network level interface which allows an easy integration of several layer networks and different administrative domains. In order to evaluate this architecture it will be implemented for an IP over ASON (Automatic Switched Optical Network) field trial developed within the Fiberhome Telecommunication Technologies. The ASON management system is designed for a multi-vendor, multi-technology, and multi-domain environment. This contribution gives a description of the architectural concepts used, the existing frameworks adopted, the design of the network level interface, and experiences made during implementation of the concepts for the Fiberhome Telecommunication Technologies field trial.

Hierarchical routing architecture has been adopted by ASON for scaling purpose, and corresponding changes in both network management structure and management information models result. Based on the discussion of new requirements for ASON management system to support the hierarchical routing infrastructure, an ASON management system is proposed and developed in this paper with new network management information models. Finally, the flexibility of the proposed management system is demonstrated on an ASON testbed with 200 emulated nodes.

In this paper, we investigate the static tunnel allocation problem in multigraularity optical cross-connect (MG-OXC) networks. Our goal is to allocate a set of tunnels that minimize the blocking probability for the dynamic traffic that will follow the prior-known historical traffic matrix. A heuristic Capacity-Balanced Static Tunnel Allocation (CB-STA) has been proposed, which always tries to allocate a tunnel from the node with maximal predicted traffic going out to the node with maximal predicted traffic coming in. However, the tunnel length constraint is not carefully considered while selecting node pairs. Thus, this work proposes a heuristic, Weighted Tunnel Allocation (WTA), to improve CB-STA. WTA adds some additional edges with predefined hop length, termed auxiliary links, to the original topology to form an auxiliary graph. Node pair incident to an auxiliary link means that there could be tunnels allocated for it. By routing the historical traffic on the auxiliary graph, preference of tunnel allocation for each node pair incident to the auxiliary link can be estimated. Finally the tunnels will be allocated according to the preference. The simulation results show that WTA outperforms CB-STA in all switching type combinations.

The blocking performance of wavelength routing WDM optical networks can be enhanced by applying routing and wavelength assignment. In this paper, we consider wavelength assignment problem in the optical network. Specifically, we propose a dynamic heuristic wavelength assignment algorithm, called Longest Segment (LS) algorithm, for WDM networks. In comparison to other proposed algorithms, the blocking performance of LS algorithm is better. In addition, the LS algorithm minimizes the usage of converters by chaining the minimum number of continuous segments which have at least one same available wavelength. Furthermore, the low algorithm’ complexity is an another advantage of the algorithm

This paper studies wavelength converter placement problem in wavelength-routed all-optical networks with dynamic traffic demand. We propose an analytical model to determine path usage probabilities for different routing policies and calculate the transit traffic for each node to evaluate the suitability of placing a converter at that node. With this, a heuristic algorithm, named maximum transit traffic (MTT), is proposed. Simulation results demonstrate the effectiveness of the proposed MTT algorithm.

In GMPLS-based networks, data is forwarded in manner of label switching on Label Switching Router (LSR). Ingress LSR must classify different IP flows into a set of Forwarding Equivalence Classes (FECs), which is a typical flow classification process. Ingress LSR will be the bottleneck of GMPLS-based networks if it could not provision fast flow classification. In this paper, we propose a novel fast flow classification scheme, coined Hierarchical Dividing Tree Scheme (HDTS), to improve the switching performance of ingress LSR in GMPLS-based networks. Four important advantages can be achieved by the proposed HDTS. First, the preprocess time in HDTS is reasonable. Second, fast FEC update is supported. Third, memory cost of HDTS is very low. Most important, the key factor that affects flow classification speed is not the number of FECs, but the depth of the search trees in HDTS. Theoretical analyses and simulations are conducted to evaluate performance of the proposed HDTS. Based on the analytical and experimental results, we can conclude that our HDTS improves the switching performance of ingress LSR greatly and is very practical for GMPLS-based networks due to its low cost.

All-optical wavelength division multiplexing (WDM) networks are the most promising candidates for the next generation wideband backbone networks. An approach to improve wavelength utilization is to allow multiple sessions share the bandwidth of a wavelength, so time division multiplexing wavelength routed networks are proposed. This paper studies the routing, wavelength and time slot assignment problem in such time-space switched networks with multi-rate traffic. We analyze the problem with a trunk-channel model, and propose an adaptive routing and trunk assignment algorithm. The effectiveness of the proposed algorithm is demonstrated by simulation.

To implement both fairness and priority scheduling in OLT scheduler, this paper proposes a SLA-supported packet fair queueing algorithm named R+WFQ. It employs WFQ discipline to serve high-priority sessions and Round Robin discipline to transmit low-priority sessions in free moments existed in traditional WFQ. Computer simulation results as well as theoretic analysis show that the proposed algorithm has excellent performance in terms of the implementation complexity, fairness and delay properties.

To solve the scalability and flexibility issue in current optical network testbeds, a Multi-functional Optical Switching Testbed (MOST) has been built. Benefiting from design of modular hardware and layered software, network experiments of scalability and complicated node architecture can be made based on it. The hardware and software structure of the MOST system is explicitly analyzed in the article and demonstrations of a 12-node ASON and shared wavelength conversion ASON on MOST are also reported.

An overview of a wide choice of current technologies and components suitable for the use of 160 Gbit/s optical time domain multiplexing (OTDM) systems is given. A selection of the presented components were used to implement a 160 Gbit/s network. In a field trial on BT’s network we demonstrated transmission over 550 km and a fully operating OTDM network including a time domain add-drop multiplexer (TD-OADM) at a data rate of 160 Gbit/s at a single wavelength and a tributary data rate of 10 Gbit/s.

Network coding is an emerging area in information theory and is expected to have a signification impact on future design of switching systems. Preliminary studied show that network coding may increase the achievable multicast throughput by significant amounts. Thus, deployment of network coding could help better exploit shared resources such as Internet connections or wireless bandwidth. Construction of network coding has become a hotspot in this area. After introducing the basic concepts of network coding, this paper summarizes the research work on required alphabet size and distribute coding scheme. Applications and open problems of network coding are also discussed.

IP/DWDM optical Internet is considered to be one of the most potential candidates for NGI (Next Generation Internet) backbone. However, the bottlenecks caused by the electronics still exist. A solution to this problem is to design an optimal virtual topology. In this paper, a novel virtual topology design method is presented, minimizing the sum of traffic-weighted hop count. Based on the given physical topology and the specific traffic matrix, the corresponding virtual topology is designed, and then, over which the traffic is routed. Due to its NP-hard nature, the design problem is decomposed into three subproblems: virtual topology construction, wavelength assignment and traffic routing, which are solved with heuristic algorithms respectively. To verify the feasibility and the effectiveness of the proposed method, a simulation environment is developed and simulation has been done over some actual network topologies. Simulation results have shown that the proposed method is both feasible and effective.

A novel two home-wavelengths Round-Robin (THWRR) wavelength assignment scheme is proposed to achieve data delivery fairness for a slotted IP/WDM metro ring. According to our simulation, THWRR works well on achieving traffic delivery fairness across all access nodes in the ring. Furthermore, when working together with an additional fixed-tuned transmitter, THWRR is able to gain significant improvement both on fairness and delivery capacity.

The metro network has experienced three generations periods. First-generation one is a traditional SDH-based network, the second-generation metro network is SDH-based and Ethernet integrated multi-service transport platforms (MSTPs) to transmit voice and data service, the third-generation SDH-based multi-service transport platforms (MSTPs) with MPLS and RPR combined to guarantee the QoS. However traditional SDH network has many disadvantages so the 4G MSTP appears, which is favored for intelligent function to realize the end-to-end service provision., intelligent control plane with GMPLS, Virtual concatenation, LCAS traffic engineer, Grooming and Switching etc. It was developed with the objective of dynamic bandwidth adjustment, protection and restoration, high network utility rate, low time delay to transport the data traffic and improve the capability of the bandwidth management, SLA (Service Level Agreement). At the same time, it can introduce the new service to make the carriers competitive. In this paper, first we present the carrier network situation, then we give the network architecture with intelligent function. At last, we analysis the performance enhancement the GMPLS MSTP bring to.

To compare with the 2G mobile communication, 3G technologies can supply the perfect service scope and performance. 3G is the trend of the mobile communication. So now to build the transmission network, it is needed to consider how the transmission network to support the 3G applications. For the 3G network architecture, it include the 2 part: Utran access network and core network. So the metro optical network should consider how to build the network to adapt the 3G applications. Include the metro core and access layer. In the metro core, we should consider the network should evolved towards the Mesh architecture with ASON function to realize the fast protection and restoration, quick end-to-end service provision, and high capacity cross-connect matrix etc. In the access layer, the network should have the ability to access the 3G services such as ATM interface with IMA function. In addition, the traffic grooming should be provided to improve the bandwidth utility. In this paper, first we present the MCC network situation, the network planning model will be introduced. Then we present the topology architecture, node capacity and traffic forecast. At last, based on our analysis, we will give a total solution to MCC to build their metro optical network toward to the mesh network with the consideration of 3G services.

P-cycle based protection has been proved to be an effective way for WDM networks’ survivability design, summed up by the notion of “ring like speed with mesh-like efficiency”. However, current p-cycles optimal design studies all ignore the protection switching cost. In fact, the cost of protection switching is a significant factor in the total cost to provide network survivability. Switching in aggregates (fiber or waveband) can significantly reduce the number of spare switch ports. In this paper, we consider that network protection is implemented at the whole-fiber level not at the wavelength level to reduce protection switching cost. Further, we address the issue about the joint optimization design of the working path routes with fiber-level p-cycle placement, and develop an improved ILP formulation. While to reduce the complexity of joint optimization problem, the simple process of pre-selecting candidate p-cycles by the efficiency and length metric and pre-selecting candidate working path routes by the hop metric is used. For a case study of COST239 network, we compare the fiber-level p-cycles optimal design with the wavelength-level p-cycles optimal design in terms of the switching cost and the required spare capacity.

This paper provides an overview of prototyping research activities of automatically switched optical networks and transport networks (ASONs/ASTNs) in China. In recent years, China has recognized the importance and benefits of the emerging ASON/ASTN techniques. During the period of 2001 and 2002, the national 863 Program of China started the preliminary ASON research projects with the main objectives to build preliminary ASON testbeds, develop control plane protocols and test their performance in the testbeds. During the period of 2003 and 2004, the 863 program started ASTN prototyping equipment projects for more practical applications. Totally 12 ASTN equipments are being developed by three groups led by Chinese venders: ZTE with Beijing University of Posts and Telecommunications (BUPT), Wuhan Research Institute of Posts and Telecommunication (WRI) with Shanghai Jiao Tong University (SJTU), and Huawei Inc. Meanwhile, as the ASTN is maturing, some of the China’s carries are participating in the OIF’s World Interoperability Demonstration, carrying out ASTN test, or deploying ASTN backbone networks. Finally, several ASTN backbone networks being tested or deployed now will be operated by the carries in 2005. The 863 Program will carry out an ASTN field trail in Yangtse River Delta, and finally deploy the 3TNET. 3TNET stands for Tbps transmission, Tbps switching, and Tbps routing, as well as a network integrating the above techniques. A task force under the “863” program is responsible for ASTN equipment specifications and interoperation agreements, technical coordination among all the participants, schedule of the whole project during the project undergoing, and organization of internetworking of all the equipments in the laboratories and field trials.

The ASON is a typical client-server network model. Via explicit interfaces between client network and server network, the client users, including UNI-C and NMS, can create an Optical Layer Connection with three schemes. This may arise the problems such as concurrency due to simultaneously accessing the same resources from UNI and NMI. This paper proposes an approach of a middleware component using CORBA-based product. The middleware component performs to define the ownership and access rules for network resources and prove global network topology and status of the network resources to the client users. It screens the difference between NMI-T, NMI-A and UNI and shields the detail of the ASON network protocols. The standardized IDL interfaces ensure that ASON products from different vendors or employing different technologies can interoperate at the common intermediate plane level and make the implementation of the management functions more simply and efficiently.

While designing ASON/ASTN systems, the bandwidth requirements for signalling networks should be made clear in advance. An analysis model based on queueing theory is proposed, and queueing delay is used as the performance metric. A simulation model and an analytical model are presented. These models can be used to guide the design of ASON/ASTN signalling networks. We analyze the bandwidth requirements for ASON/ASTN signalling networks with RSVP-TE protocol. The analysis results show that, the main signalling network bandwidth consuming packets are RSVP-TE soft state refresh packets and establishing/releasing connection packets. The upper bound of bandwidth requirement of soft state refresh can be determined directly. But the establishing/releasing connection process is a stochastic process, and its bandwidth requirement depends mainly on mean arrival rate of the signalling packets at the rush hour, or the number of concurrent requests.

This paper presents a novel optimized distributed dynamic wavelength assignment algorithm for the wavelength routing optical network(WRON) based on the revised result of reference[5,6]: Fast Distributed Dynamic Lightpath Setup Algorithm(FDDLSA). By adopting two kinds of new optical link release mechanisms (ahead and timeout release) in the process of networks nodes setting up a link, the new algorithm can reduce the blocking probability significantly and release the burden of a router. The effectiveness of the algorithms was verified through the simulation on NSF net (Nature and Science Foundation) of USA: it can reduce the blocking probability 8% less than reference [5]’s algorithm, at the same time it can reduce the optical path setup time greatly.

In the ASON architecture, optical network is viewed as a separately deployed transport network providing transparent transport service to IP/MPLS domain networks. As ASON lack of scalability and efficient traffic grooming, GMPLS enabled multilayer network is a promising alternative for deployment of IP over WDM network. The multilayer network composed of network node that is based on the hybrid architecture integrating IP/MPLS layer and optical switch with a unified GMPLS control instance can bypass lightpath at optical layer or terminate a lightpath locally at IP/MPLS layer. This paper illustrated the architecture of G-LSR, and addressed the essential aspects of network operation mechanisms including multilayer routing, lightpath provisioning and LSP hierarchy provisioning. Current IP network can gradually evolve to IP/optical integrated multi-layer network through updating network node of IP router to G-LSR.

Generalized Multiprotocol Label Switching (GMPLS), which is developed to support common control of packet, TDM, wavelength, and fiber services, is the key enabler of the new network model. The survivability of IP over WDM networks gains importance as network traffic keeps growing. In this paper, we propose an integrated provisioning scheme to dynamically allocate Label Switched Paths (LSPs) in IP over WDM networks. This scheme takes advantage of information sharing between layers (e.g., link state information, bandwidth usage, and protection capability) to eliminate redundancies and inefficiencies in the traditional layer-independent service provisioning. The integration of information is facilitated by GMPLS signaling. The proposed scheme also uses GMPLS capabilities to provide end-to-end survivability against network failures. The ability to provision across all network layers ensures efficient bandwidth usage. We propose two integrated routing algorithms: availability-based integrated routing algorithm and joint availability-based integrated routing algorithm. The simulation is made to evaluate the performance of our proposed integrated provisioning mechanism. As a result, network performance can be optimized over all layers. This could lead to significant cost savings for service providers.

In this paper, we investigate the wavelength assignment issue in wavelength-routed all-optical WDM networks with Sparse-Partial Wavelength Conversion. Traditionally, wavelength assignment is only designed for the purpose of decreasing the blocking probability. In Sparse-Partial Wavelength Conversion, wavelength assignment can be used to decrease the number of wavelength converters if keeping the same level of blocking probability. We propose three different wavelength assignment algorithms, namely Segment-First First Fit (SFFF), Path-First First-Fit (PFFF), and Minimum Converter Allocation (MCA). Our results show that MCA is the best one among the three, and it can achieve very close performance to Sparse Wavelength Conversion by using a very small number of wavelength converters. The performance of PFFF is very close to MCA, and it is very simple to be implemented.

This paper presents a novel Policy-based Flow Switching (PFS) method used in packet networks. Unlike traditional per-packet forwarding method, PFS only performs extensive processing on the first packet of a flow, associates this flow with a policy and applies the result of this processing on subsequent packets in the flow. Thus, PFS has the capability of flexible services offering and can significantly improve data forwarding rate, without introducing any complex signals. In this paper, we present the architecture overview of PFS and introduce some of the implementation issues. Besides, this paper also demonstrates how PFS is implemented on a 128Gbps NPU-based high-end router.

With the large-scale application of metro MSTP in recent two years, a lot of problems have emerged inevitably. Some more advanced technologies or mechanisms are needed to deal with the application problems of current version MSTP. RPR and MPLS become new focus of enhanced MSTP’s features and the combination of them will boost maturity and further development of enhanced metro MSTP. The MSTP embedded with RPR and MPLS not only find a solution for perfect transport of TDM service, but also improve the support ability for data traffic and bandwidth utilization. With the help of enhanced MSTP, it’s easy for network carriers to provide a reliable, cost-effective, flexible and competitive transport infrastructure delivering various advanced services. RPR firstly solves the conflict between the transport efficiency and QoS of packet service via some mechanisms such as dual counter-rotating ringlet, statistical multiplexing, bandwidth sharing, topology discovery, fairness control etc. Secondly, RPR guarantees the QoS of subscriber by strict service classification and priority. Thirdly, RPR can provide the carrier-class protection within 50ms. On the other hand, MPLS can solve the lack of VLAN ID by LSP labeling and in addition it offers VPN service for different customers then enhances the SLA. Most of all, MPLS provide cross-networking ability, end-to-end service configuration and QoS guarantee, traffic engineering and network optimization. Enhanced MSTP will be applied in convergence or access layer of metro network first. It could have the function of RPR and MPLS at the same time, the reconciliation sub-layer should adopts GFP or LAPS encapsulation, the mapping granularity can be based on VC-12/VC3/VC4-Xc/v and the LCAS should be supported when the transmission channels are configured by virtual concatenation. Enhanced MSTP can be used to build either a single ring network or multi-ring network for improving the ability and quality of Ethernet transportation and providing new services such as MPLS L2 VPN.

With the proliferation of IP networking and wireless access, local exchange carriers (LEC) are under tremendous pressure to evolve into broadband service providers. On-demand video distribution to the mass appears to be one of the best evolution paths for the LEC. Various approaches for providing on-demand content/video distribution to massive viewers are compared. In particular, a novel low-cost network architecture called DINA for distributing video content to massive viewers is described. DINA employs hybrid multicast-unicast streaming to achieve scalability. Furthermore, it offers full interactive functions to the viewers and readily deployable over existing broadband network infrastructures. An innovative technique to offer low-cost on-demand video services to massive subscribers called multi-stream multicasting is described. It employs fixed but irregular stream intervals that can significantly reduce the startup latency. Many fault protection mechanisms are built into the DINA architecture to improve the robustness. DINA also offers additional protection against copyright infringement and has been demonstrated commercially.

In this paper, we consider congestion control in input queued crossbar switch environment where each input port with finite buffer space while TCP protocol is employed for end-to-end congestion control. We find that it is impossible to achieve efficiency and fairness among TCP flows at the same time only by queue management. Then we propose a scheme of hFS&rEDF, which combine heuristic fair switch arbitration (hFS) and queue management policy of early drop front randomly (rEDF). In our proposed scheme, switch arbitration strategy of hFS unevenly allows input ports to transfer packets to output ports while packets at head of any other input ports involved in conflicts have to be dropped by the policy of rEDF with a probability. Simulation results prove that our proposed scheme can achieve better tradeoff between throughput and fairness.

In this paper, The Next Generation Optical Network Service Platform(NGOSP) based on the architecture of ASON was studied. The paper presented the layered concept of service function around the telecommunication service requirement base on the analysis for the evolution trend of service optical network and modes of service provision, and definitized the reference model and constructed the function framework of service optical network. The general function components also were abstracted from service perspective and a thought of mapping function of service layer to network technical layers was presented.

This article presents an FPGA-based Dynamic Multicasting emulator for Wide-Area Transport Networks. We setup overlay prototype network environment based on the emulator and compare the performance of pure IP multicasting and Layer 2 multicasting.

In the modern world of telecommunications, the concept of wireless global coverage is of the utmost importance. However, real global coverage can only be achieved by satellite systems. Satellites communication is the most important mean of the communication network. The traditional satellites communication and inter-satellites links are built by microwave. In recent years, laser links for inter-satellites communication are becoming more and more important. Laser communication systems operate in a frequency range above the regulated spectrum. Laser provides many advantages for using in point-to-point ISLs (inter-satellites links) such as for links between satellites and spacecraft in deep space. Such advantages include: high speed; high bandwidth; small antenna size; narrow field of view; and narrow antenna beam. These advantages combined with the advantages in fiber optic components (optical preamplifiers, multiplexers, detectors, etc) have made laser attractive for laser links. Now we can bring WDM (wavelength-division multiplexing) to emerging broadband satellite communication systems. By using the common antenna system and ATP (Aiming, Tracking and Pointing) system, the satellites will get more capacity. In the inter-satellites laser communication, the important performances of the systems such as BER and BL both have direct relation with the optical systems. The optical systems have the function of ATP. The optical antenna is the most important component of the optical system. So the optical antenna is an important key technology to the inter-satellites laser communication. In this paper, we mainly study the optical system in the inter-satellites laser communication. we compare with three kinds of optical antennas: refractor and reflector and Catadioptrics of the passive optical system; we also analyze the effect of bandwidth to the WDM communication systems; we use the correlative software, simulate the curves of the performance of the optical antennas. These analyses will be the base of the system of WDM laser inter-satellites communication.

With the advent of many new multimedia applications with Quality of Service (QoS), the issue of multicast routing has become more and more important. Some previous researches in this area mainly take bounded end-to-end delay into account and neglect the ability of router in supporting multicasting, therefore a new multicast routing algorithm based on genetic algorithm (GA) is proposed to construct Degree-Delay-Constrained least-cost multicast routing Tree (DDCT problem). Our proposed algorithms adopt pruefer coding and decoding technique which is very suitable for crossover and mutation operation and can save runtime; Otherwise because pruning and grafting techniques are used to adjust node degree and repair tree structure, this algorithms may also be employed when taking dynamic adjusting tree or keeping load balance into consideration. Simulations were conducted on a random graph to evaluate the performance of the proposed algorithm compared with that of the other algorithms such as CSPT and Wang_GA. The experimental result shows that the proposed algorithm performs well.

With the emergence of new distributed multimedia applications upon secure group communication, Security becomes more important increasingly. Now most investigations about multicast security focus on key management problem. In this paper, an improved version of CBT is proposed for data communication and key management and named as Dynamic CBT(D-CBT). This approach accepts the advantages of Iolus and CBT approaches, its main characteristic is that D-CBT changes dynamically with group member join/removal frequently when cumulating influence of dynamic membership reaches certain threshold so that the bandwidth/cost of multicast is as least as possible, D-CBT is also very close to realistic application case. Performance analysis will be conducted compared with other methods.

Currently Metropolitan Area Optical Networks are still based on SDH/SONET technologies. The disadvantages of such a rigid SDH/SONET rate hierarchy, especially when data applications such as Ethernet are considered, were soon realized. At the same time fast, link setup and release, and dynamic increase and decrease of the link bandwidth will be the scene of the future network. In order to meet the new requirements, the future network will combine with ASON/GMPLS, which is just the protocol to provide the basic function of topology and resource discovery and fast connection provision with intelligence. In this paper we discuss the new technologies and their combination in metropolitan area networks.

The performance of normalized throughput in a hybrid WDMA-OCDMA system is studied. Effect of optical beat noise and dispersion on the normalized throughput is analyzed. The scheme of BCH channel coding is proposed to improve the performance of a hybrid WDMA-OCDMA system. Analysis result shows that the normalized throughput can be improved greatly when traffic load is medium. Furthermore, the maximum normalized throughput will increase for the BCH-coded hybrid WDMA-OCDMA system

Internet only provides users with best effort services, today, with the development of next generation Internet, next generation Internet will provides users with tremendous services, how to manage these tremendous services and provides users with differentiated services automatically, must be solved for service providers and network providers. A two planes’ service management architecture is proposed, the one is application service management based on application plane, and the other is transport service management based on transport plane. In this paper, mapping application service into transport service and issues for QoS are studied, primarily. It is well known that different application services correspond to different QoS requirement during they are transported in network, likewise, the different users need different QoS requirement for the same application service. Therefore, Firstly, application services are classified, and QoS requirement for different application service is studied; secondly, the NPLs(Network Performance Level) are studied, and draw a conclusion, if NPL array corresponding to application service is assured in network, the application service of quality can be assured also. Thirdly, as an interface between application service and transport service can be expressed in policy, thus, mapping application service into transport service can be implemented by configuring the policy. Finally, WFQ (Weighted-Fair-Queue) algorithm is researched primarily based on DiffServ, and several simulation models based on NS2 are designed and different algorithms are simulated, the simulation results show that different application service of quality can be assured by corresponding to scheduling algorithm in DiffServ mechanism model. By means of policy idea, the complicated issues for mapping application services into transport service can be solved, and different scheduling algorithm in DiffServ are fit for different application services’ QoS assurance.

Even though the foundation of iSCSI (Internet SCSI) protocol is quite different from that of Fibre Channel, and maybe the original idea of replacing FC-SAN with truly IP-based iSCSI-SAN is hard to be fulfilled in short term, nobody can ignore the broadly existence and the bright future of the iSCSI technology. We present an introduction to build several iSCSI appliances in this paper from a single stand-alone iSCSI Target, an iManager-based iSCSI appliance, and our embedded iSCSI Target implementation to a RAIN (Redundant Array of Independent Nodes) -based iSCSI appliance. Not only do we offer some experiment results of our standalone iSCSI Target and our embedded iSCSI Target, our in-band and out-of-band iManager-based iSCSI Target designing strategy and performance comparison, but also give an explanation of our RAIN-based iSCSI Target implementation concerning to the topology of our RAIN Target, the node data distribution method and our ways to deal with the capacity-expansion-on-demand problem.

All-optical Networks (AONs) are unique in that they support many optical wavelengths , which use optical amplifiers, optical add/drop multiplexers and optical switches or cross-connects. They contain only optical components and are intrinsically different from optical networks currently being used. Note that many of the security problems present in traditional electronic or electro-optic networks are still to be found in AONs. Moreover, for AONs’s service disruption is most easily achieved in optical amplifies, while the tapping or eavesdropping is most easily achieved through nonlinearities in transmission fiber. These types of attacks on AON will thwart data security. In order to assure information in AONs which have a minimum privacy and quality of service (QoS), this paper presents an analysis of attacks in AONs and propose a conceptual system framework for information assurance in AONs.

Ethernet over SDH/SONET (EOS), which connects different Ethernets through the existing SDH/SONET infrastructure, is a promising data transmission technology in today’s networks, for it successfully combines the simplicity and affordability of Ethernet with the resilience and scalability of SDH/SONET. Virtual Concatenation (VCAT) is one of the key technologies in EOS, which provides the capability to transmit and receive over several noncontiguous parallel Virtual Container (VC) fragments as a single flow and drastically improves efficiency of over 33 percent against standard concatenations. Link capacity adjustment scheme (LCAS) over VCAT signaling scheme further enhances VCAT to tune bandwidth dynamically at the requests of network management system without disturbing the existing traffic. In addition, the scheme will automatically decrease the capacity if some member of VCAT experiences a failure in the network, and increase the capacity when the network fault is repaired. In this paper, our design of LCAS over VCAT within an EOS chip is provided, which supports four Virtual Concatenation Groups (VCG) to adjust their bandwidth simultaneously. The block diagrams of the total design for LCAS over VCAT are also provided and several open problems that we encountered during implementation and their corresponding solutions are discussed in focus. Thoroughly functional simulations and FPGA verifications have been done to the design to prove its validity. Finally, we have the design synthesized with Synopsys’s Design Compiler, which reveals that the whole design is realizable in ASICs.

Multicast services have been increasingly used by many multimedia applications. As one of the key techniques to support multimedia applications, the rational and effective multicast routing algorithms are very important to networks performance. When switch nodes in networks have different multicast capability, multicast routing problem is modeled as the degree-constrained Steiner problem. We presented two heuristic algorithms, named BMSTA and BSPTA, for the degree-constrained case in multimedia communications. Both algorithms are used to generate degree-constrained multicast trees with bandwidth and end to end delay bound. Simulations over random networks were carried out to compare the performance of the two proposed algorithms. Experimental results show that the proposed algorithms have advantages in traffic load balancing, which can avoid link blocking and enhance networks performance efficiently. BMSTA has better ability in finding unsaturated links and (or) unsaturated nodes to generate multicast trees than BSPTA. The performance of BMSTA is affected by the variation of degree constraints.

To solve the problem of optical path creation in multi-granularity optical network, a multi-granular hierarchy label is proposed in this paper. A parallel signaling scheme using hierarchy label is discussed in detail. Simulation result shows that the new proposal has a better time performance than traditional hierarchy LSP signaling scheme.

Open shortest path first (OSPF) is the most widely used routing protocol in today’s IP networks, and its excellent performance has been proved in wired environments. However, when it is executed under bad channel conditions such as in wireless networks or areas with severe signal interference, links may flap frequently and some terrible problems will appear. This paper proposes a novel algorithm called OSPF link flap damping algorithm (OLFDA). The objective of OLFDA is to reduce the events of link state advertisement (LSA) update and damp the link flap with the precondition that the overall network performance is satisfying. To accomplish this, we can define criteria to identify and dynamically suppress the poorly behaved links. Information of the suppressed links won’t be advertised in OSPF domain and used in calculation of the routing tables. In addition, we can control the maximal number of links suppressed simultaneously by a router to ensure the network connectivity. OLFDA are simulated in many scenarios, and the results indicate that the algorithm has an excellent performance.

Based on the ring topology and the applied technology at present, we propose an access and control protocol for optical packet switched metropolitan area DWDM (Dense Wavelength Division Multiplexing) network. In this paper, we have given an access and control protocol first, and explained the reservation of resource, the setting of optical transmitter and receiver, the control message, and then studied the properties of waiting delay through mathematic analysis and computer simulation.

In this paper, the problem of connection provisioning with QoS guarantees is studied in DiffServ/MPLS over optical networks. In most previous studies, the objective is to provide bandwidth guaranteed LSPs, without considering service differentiation. In practice, to guarantee QoS for some class of services in a DiffServ scenario, other constraints besides bandwidth should be enforced on the LSP and the network. We discuss these constraints and take them into account in the routing procedure. The simulation results show that QoS is guaranteed at the cost of higher bandwidth blocking probability, and an appropriate preemption mechanism is indispensable in the DiffServ/MPLS over optical networks.

Due to its low cost and packet data efficiency, Ethernet has been one of the most influential technologies for Local Area Networks (LAN). Moreover, the 10-Gigabit Ethernet has begun to move Ethernet from the LAN out to encompass the metro area network. The technology features, protocol architecture and the frame format are introduced. In order to realize the logical boundary between the physical layer and the link layer, the protocol of SPI (System Packet Interface) level4 is adopted. A method of realizing the SPI interface of 10-Gigabit Ethernet is put forward and the function block diagram is presented. In order to reduce the power waste of the chip, the parallel algorithm is chosen in the design.

The significance of Generalized Multi-protocol Label Switching (GMPLS) has been proven by its significant adoption in ASON. In recent years GMPLS has become the most global solution to provide intelligent optical resource configuration through not only devices that perform packet switching, but also devices that perform switching in the time, wavelength, and space domain. This new technology can be used to support services for applications that requiring high QoS performances in any network which using the Internet Protocol (IP), including the Passive Optical Network (PON). In this paper, we will propose a new PON technology that uses GMPLS as layer2 protocol. The focus of this paper is mainly on the issues and key characters that will encounter in the combination of GMPLS technology and PON technology. Compared to other PON technologies, such as APON and EPON, GMPLS-based PON (GMPLS PON) can provide more powerful and efficient service in access network, and it can also provide better QoS support to the video and audio applications. The most significant is GMPLS PON can also provide common control and data and administrative planes between the backbone network and the access network, so it will be more convenient and efficient to operate and manage.

Both protection and restoration are fundamental consideration in designing robust optical networks, especially in challenging GMPLS/ASON architecture. Fast Rerouting technology as one important method of MPLS-Traffic Engineering is an efficient and powerful technology for protecting MPLS-TE LSPs from link and node failures. By supporting locally reparation of the affected LSPs at the point of failure, link and node protection are achieved quickly. The goal of real-time IP applications over native IP unicast/multicast networks or LDP based MPLS networks is to limit the IP packet loss duration in the network to 10s of milliseconds in the event of link/node failures. RSVP signaled LSP is used with explicitly routed path as the re-direct tunnel, while the protected traffic can be either MPLS traffic engineered LSPs, LDP based LSPs, IP unicast, IP multicast traffic or the mix of them. This mechanism can be applied to both point-to-point links and multi-access links in the cases of the link protection and node protection. Besides the above resilience research on IP/MPLS single layer, tests or experiments on multilayer mechanism, i.e. optical and IP/MPLS network layer composing of multi-vendor routers and multi-domain, are carried out by a number of famous organizations and companies where optical dedicated/shared mesh protection/restoration and MPLS fast rerouting survivable techniques are combined to guard against both dual link failures and combined link and IP/MPLS hardware and software router failures. Those techniques used in next-generation backbone network design result in that not only familiar failures such as fiber cuts or amplifier damage but also system failures such as router failures or multiple concurrent failures can be recovered fast and effectively at corresponding layer. In this paper, on the basis of analysis of IETF's internet drafts about fast rerouting algorithm and standards in MPLS-TE and GMPLS, new fast rerouting algorithm and corresponding simulation results are proposed step by step. At the same time, resilient performance in a GMPLS/ASON design especially under mesh optical protection is studied carefully while a few coming function requirements in OXC and router equipments obtained from the above analysis are discussed to some extent. Figures and tables are presented for explaining the enhanced fast rerouting technology reducing confusion of packets and achieving falling delay time of rerouting. Finally, some issues for further work are to identify the importance where to place appropriate resilient mechanism in the next network design.

Under the constraints of the number of transceivers per node and wavelength continuity, the fairness issue of connections with different bandwidth granularities in survivable traffic-grooming WDM mesh networks is investigated. An extended wavelength-plane graph (EWG) model and two dynamic grooming path protection algorithms with considering blocking fairness are proposed. Based on dynamic traffic with different load, the performance of our schemes has been evaluated via simulations. The results show that they provide good blocking fairness while at the same time ensure a small increase in the overall traffic blocking probability.

The wide use of network improves security risks, but the traditional network security tools are single-functional and they are difficult to extend and manage. These security tools can’t fulfill the need of users. The paper presents a network security oriented language: NSL (Network Security Language). The language provides a common abstract layer of all kinds of security application. It can be used to construct multifunctional, distributed and extendable security applications, satisfying various and emergent security requirements. The paper describes its network oriented structures, features and three security mechanisms. The NSL interpreter implementation is also analyzed by the performance test. At last, the paper puts forward improving work which will be done in the future.

In this paper, we propose a new shared-path protection approach, called joint routing algorithm (JRA), under shared-risk link group (SRLG) constraints for survivable WDM mesh networks. JRA differs from previous algorithms that are socalled separated routing algorithm (SRA), and can find K path pairs and select an optimal path pair as the result, while SRA can only find a path pair that may be not an optimal routing pair. So, JRA can perform better than SRA. We also study the relationship between the protection switching time and the resource utilization, and suggest a new joint cost function to compute the least-cost path pairs. Under dynamic traffics with different load, the simulation results show that JRA not only has better performances than SRA but also can determine the appropriate tradeoffs between the resource utilization ratio (or blocking ratio) and the protection switching time.

Ethernet Passive Optical Network (EPON) has been considered as the best candidate of the next generation access network, because of it’s particular advantages, such as inexpensive, simple, scalable, and capable of delivering bundled voice, data and video services. However, the standard isn’t approved, many questions are still discussing. Dynamic Bandwidth Allocation (DBA) is one of the hop topics. If we use static bandwidth allocation in EPON system, slots may not be filled to capacity in the case when the OLT grants to an ONU a slot smaller than the ONU requested based on its queue size. Then the packet will have to wait for the next slot, so there is an unused remainder at the end of the slot. As a result variable-length packets don’t fill the static slot completely. Conventional E-PON uses a slot-size based DBA algorithm. Packets of all priorities are put in the same slot. Due to the burstness of traffic and packet length variation, the starting point of each slot is push-pulled from frame to frame, as a result delay variation of packet is not under control. We can not support appropriate Qos to real-time traffic. In our design, we divide Ethernet data into three parts, one is the steady part that has N time slots for voice traffic, corresponding to N ONUs. In this case, there are M ONUs those has video traffic for transmit, the second part having M time slots will be established named Quasi-dynamic. And the other is the dynamic part that is basically one giant slot for data traffic, shared by all ONUs. Since the steady part of the frame can be approximated by constant flows, it will make influence upon voice. Establish or remove Quasi-dynamic channels will take less than 50ms delay variation. So this arithmetic can match the video’s character of less than 100-ms maximum jitter. Usually, for the low priority class traffic, there is no requirement on any QoS parameters; therefore, optimizing throughput is the goal. So we use IPACT algorithm in the third part, making utilization almost 100%.

With the need of more and more high quality services, EPON system is widely favored by most people with its advanced technology of Gigabit and PON, which will replace the traditional techniques of copper and MC gradually. We can realize the MPCP protocol defined in IEEE802.3ah by the hardware scheme, such as FPGA or ASIC. Using SNMP protocol to achieve network management is the popular way. SNMP network manager can perform the long-distance configuration of the parameters in EPON system by sending out SET message; on the other hand, it can research the information by sending out GET message. Consequently, the Nios embedded processor acts as a transmission channel or a bridge between SNMP agent and hardware system. Now SOPC is a popular design method, which processes flexible design mode, reducible, expansible, upgradeable, and have the programmable function between hardware and software synchronously in a single chip. Integrated with the advantages of SOC, PLD, and FPGA, SOPC is provided with the following basic characteristics: an embedded processor core; on-chip high speed RAM resources with small capability; processor debug interface and FPGA programmable interface, etc. The Nios embedded processor is a soft core CPU optimized for programmable logic and SOPC (System-on-a-programmable-chip) designs, which accomplishes the data collection and configuration between SNMP agent and hardware system, the report of registration and alarm information, also the fulfillment of DBA which can be operated with all kind of algorithms. SOPC builder is a tool employed as turning out a system based on bus, thereby many components are included in this design, for instance, CPU, memory interface, peripherals interface etc. Developing applications using the Nios embedded processor is slightly different from the traditional processors, since the designer can configure the processor architecture and specify the peripheral content. That is, a designer can build a microcontroller according to system design requirement, as opposed to selecting a pre-built microcontroller with a fixed set of peripherals, on-chip memory, and external interfaces. In this paper, we introduce an EPON system by way of FPGA, and a SOPC design on the basis of Nios platform is given, in which data transmission channel and DBA functions are achieved. According to the need of EPON system, CPU, memory, and peripheral interfaces are selected from the library of SOPC builder. Once the Nios system is created, it may optionally be combined with other used-defined logic. Designer writes the source code, compiles the application software, and debugs the code to meet the need of EPON system. Keywords: EPON, Nios, SNMP, SOPC, DBA

RRP has two means to transmit data: store-and-forward and cut-through. In this paper, the high and low priority packet transfer delay of the N nodes Resilient Packet Rings (RPR) in store-and-forward architecture is analyzed based on the queuing theory. According to queuing theory, we set up the nodes model and analyzed the factors that influenced the packet transfer delay in a constrained condition. By calculation and simulation, the result indicates that both high priority and low priority packets’ delay increase with the node number N of the RPR rings. The high priority traffic has less packet delay than the low priority traffic at the same node number N. The increase of the low priority transfer delay is much larger than the high priority traffic with the increase of the node number.

To keep pace with the trend of service diversity in communication, providing differentiated reliability services become a focus topic of WDM networking. While the wavelength has transmission capacity at gigabit per second rates, the network may be required to support traffic connections at much lower rates than the full wavelength capacity. In addition, for networks of practical size, the number of available wavelengths is still much lower than that of source to destination connections requests. Thus the solution lies in efficiently grooming the low rate traffics onto the wavelength channels. Traffic grooming can aggregate low-rate connections onto high-capacity wavelength channels, also called lightpaths, to make efficient use of the wavelength capacity. Considering both the survivability for differentiated reliability services and traffic grooming, a differentiated shared protection algorithm called PSPTG, which means Partial Shared-path Protection algorithm supporting Traffic Grooming, is proposed. In this algorithm, a shared protection scheme of partial lightpath based on link reliability is presented to meet the requirements of reliability and bandwidth of connections. Simulation results show that the algorithm is efficient in terms of resources utilization by routing more connections.

As the dynamic data traffic became the dominant traffic pattern in optical networks, adaptive routing strategy has already become the focus of recent studies on the routing and wavelength assignment (RWA) problem, which will determine the candidate path or path set adaptively and dynamically according to the real-time wavelength utilization status. Without any doubt, adaptive routing strategy can significantly reduce the overall call blocking probability (CBP), but it also remarkably increases the computational complexity of the RWA algorithm, especially in large-scale networks. This paper aims to dig the potential of fixed routing strategy for better overall CBP feature as well as its dramatic advantage over its adaptive counterpart in terms of much lower computational complexity. Both single and alternate routing scenarios will be considered in our study: In the investigation on single routing scenario, a heuristic method for determining optimized paths route for all node pairs will be proposed, which is developed from an approximate analysis model based on path decomposition approach; turn to alternate routing scenario, fairness among diverse connection requests will be introduced as another criterion, and our novel fixed classified alternate routing (F_CAR) approach also show its competitive performance.

Interconnection network plays an important role in Massively Parallel Processing (MPP) systems. The 2-dimensional mesh (Mesh for short) has been one of the most common interconnection networks. In this paper, we propose a new 2-dimensional network topology: Hexagon Interconnection Networks. Theoretical network characteristics such as the diameter, the bisection width are discussed in detail. It is shown that Hexagon can connect more nodes than Mesh when they have same network diameter. Moreover, a store-forward routing algorithm and a deflection routing algorithm are designed for Hexagon topology. Hexagon can also be used in optical interconnection network.

Ensuring minimum quality-of-service levels to traffic flows and groups of flows is an important challenge for future packet networks. Admission control is an important measure for QoS provision. However, the problem of multicast admission control was much less investigated. In this paper, a reservation-based edge multicast admission control scheme is investigated, which targets high QoS level and scalability. Using a new concept, virtual link, a virtual-link-oriented quota-based resource management mechanism is proposed. Besides, a hierarchical distributed bandwidth broker architecture is also employed. In this scheme, a central bandwidth broker manages the physical link bandwidth, and allocates/reclaims the bandwidth to virtual links; edge bandwidth brokers manage the bandwidth of virtual links, make admission decisions and reserve bandwidth for each flow. Simulation results show that the scheme achieves high scalability. Furthermore, the impact of quota size on the performance and complexity is investigated. A bandwidth adaptive quota size scheme is proposed to further improve the performance and reduce the complexity.

Both hardware and software are optimized simultaneously to improve the bus throughput of node computer in cluster computing network. A cost effective optical interconnection ring network is established using off-the-shelf PCs with windows operating system. A network adapter card ONIC is developed. Using ETDM and field program technique, the full bandwidth of the 32bit×33MHz PCI bus is reached in physical layer. In order to overcome the hardware and software data bottlenecks induced by using of general purpose platform such as PCs and Windows 2000 system, several methods are adopted to maximum the bus throughput of node computer. A user-level interface is designed to blur the kernel/user mode boundary of the software system, and bypass the OS overhead. Zero-copy DMA is realized. The bandwidth in application level using PIO and DMA transfer mode is improved 17 and 26 times respectively. The sustained bandwidth in application layer can reach 437.6Mbit/s. It’s also shown theoretically and experimentally that the selection of main board chipset, using memory space map, reasonable using of PIO and DMA transfer mode, using polling and large MTU will also improve the bus throughput of node computer considerably.

A new type of CATV network management system developed by universal MCU, which supports SNMP, is proposed in this paper. From the point of view in both hardware and software, the function and method of every modules inside the system, which include communications in the physical layer, protocol process, data process, and etc, are analyzed. In our design, the management system takes IP MAN as data transmission channel and every controlled object in the management structure has a SNMP agent. In the SNMP agent developed, there are four function modules, including physical layer communication module, protocol process module, internal data process module and MIB management module. In the paper, the structure and function of every module are designed and demonstrated while the related hardware circuit, software flow as well as the experimental results are tested. Furthermore, by introducing RTOS into the software programming, the universal MCU procedure can conducts such multi-thread management as fast Ethernet controller driving, TCP/IP process, serial port signal monitoring and so on, which greatly improves efficiency of CPU.

A novel restoration scheme, Parted Path Shared Restoration (PPSR), was proposed in this paper. The major idea of PPSR is the strategy of 'parted disposal'. PPSR keeps the advantage of Path Based Shared Restoration (PBSR) in utilization of capacity. In addition, restoration time of PPSR is much less than that of PBSR. Furthermore, a satisfaction function was proposed to estimate the performance of PPSR. This function takes the utilization of capacity and restoration time into a harmonious and uniform frame. Through theoretical analysis and computer simulation, the performance of PPSR was demonstrated.

In this paper we consider the problem of designing a network of optical cross-connects(OXCs) to provide end-to-end lightpath services to label switched routers (LSRs). Like some previous work, we select the number of OXCs as our objective. Compared with the previous studies, we take into account the fault-tolerant characteristic of logical topology. First of all, using a Prufer number randomly generated, we generate a tree. By adding some edges to the tree, we can obtain a physical topology which consists of a certain number of OXCs and fiber links connecting OXCs. It is notable that we for the first time limit the number of layers of the tree produced according to the method mentioned above. Then we design the logical topologies based on the physical topologies mentioned above. In principle, we will select the shortest path in addition to some consideration on the load balancing of links and the limitation owing to the SRLG. Notably, we implement the routing algorithm for the nodes in increasing order of the degree of the nodes. With regarding to the problem of the wavelength assignment, we adopt the heuristic algorithm of the graph coloring commonly used. It is clear our problem is computationally intractable especially when the scale of the network is large. We adopt the taboo search algorithm to find the near optimal solution to our objective. We present numerical results for up to 1000 LSRs and for a wide range of system parameters such as the number of wavelengths supported by each fiber link and traffic. The results indicate that it is possible to build large-scale optical networks with rich connectivity in a cost-effective manner, using relatively few but properly dimensioned OXCs.

In order to provide better QoS service, optical networks with survivability require complete procedures to provide routing and protection for the traffic carried on the light paths. Thus, the challenging problem of disjoint path routing and protection in WDM networks is pivotal to realize the end-to-end QoS guarantee. In this paper, we propose the novel routing algorithm, CHRTG (Constrained Heuristic Routing with Traffic Grooming). CHRTG with multiple QoS constraints is different from previous routing algorithms, which are based on the traditional single constraint (bottleneck or additive). We describe QoS constraints with different weights and corresponding metrics can be produced for path computation. Better performance can be realized through analyzing the complexity degree of time and space. The traffic load from each node can be expressed in self-similar traffic and several on/off source models. Simulation results show that CHRTG can achieve maximum sharing protection paths. The results of simulation also show that the algorithm is scalable, efficient and survivable.

In this paper, we describe and investigate the protection design for survivable WDM mesh networks, and propose our heuristic shared-path protection algorithm, called dual-link failures protection (DLFP). We also present a backup resources assignment scheme to completely protect the dual-link failures, and suggest a joint selection method, which consider the integrated performance of resource utilization and protection recovery time, to compute an optimal working and backup path pair. The simulation results show that DLFP can not only perform better than previous three-step algorithm (TSA) for dual-link failures but also determine the appropriate tradeoffs between the resource utilization ratio (or blocking ratio) and the protection recovery time.

With the explosive growth of streaming services, users are becoming more and more sensitive to its quality of service. To handle these problems, the research community focuses of the application of caching and replication techniques. But most approaches try to find specific strategies of caching of replication that suit for streaming service characteristics and to design some kind of universal policy to deal with all streaming objects. This paper explores the combination of caching and replication for improving streaming service quality and demonstrates that it makes sense to incorporate two technologies. It provides a system model and discusses some related issues of how to determining a refreshable streaming object and which refreshment policies a refreshable object should use.

This paper describes our work on the key techniques in ASON control plane and introduces a practicable software simulation-based ASON network framework and node structure. The implementation methods of the control channel and the composition of the control protocol traffic flow are analyzed. Then simulation of the control channel capacity requirements in ASON under different conditions is realized. Relationship between the control channel traffic rate and the network traffic load of the ASON network is introduced. The capacity requirements of the control channel under different connection setup time limits are also analyzed. In the end the approximate equations to describe these relations are developed.

With the explosive growth of the Internet and dramatic increase in wireless access, users expect high quality streaming media service over wireless network. Due to the high bandwidth requirements and the long-lived nature of streaming objects, it is a challenge to provide efficient multimedia streaming service in highly dynamic wireless environment. This paper discusses the ways to improve wireless network streaming media service quality using proxy caching. We present a novel streaming media caching model for cell-based wireless networks. Its architecture is compatible and interoperable with existing wireless network infrastructure. The model utilizes some well known caching techniques such as segmentation, prefix caching, and cooperative caching. We also describe the implementation details and explain how this model works. Our evaluations show that this model is effective in the high quality streaming service for wireless networks. The model not only reduces the initial delay of applications, but also minimizes the application startup delay and wireless network bandwidth utilization.

A major difficulty of current intrusion detection model is the attack set cannot be separated from normal set thoroughly. On the basis of paraconsistent logic, an improved intrusion detection model is proposed to solve this problem. We give a proof that the detection model is trivial and discuss the reason of false alerts. A parallel paraconsistent detection algorithm is presented to develop the detection technology based on our model. An experiment using network connection data, which is usually used to evaluate the intrusion detection methods, is given to illustrate the performance of this model. We use one-class supported vector machine (SVM) to train our profiles and use supported vector-clustering (SVC) algorithm to update our detection profiles. Results of the experiment indicate that the detection system based on our model can deal with the uncertain events and reduce the false alerts.

The foreseeable lightpath demands (FLD) include scheduled lightpath demands (SLD) and predictable lightpath demands (PLD, lightpath demands that can be approximatively forecasted through the history traffic record). Since the time disjointness each FLD, the network planner can design the capacity of the network more compactly. We formulate the routing and wavelength assignment (RWA) problem when network dimensioning is coupled as the Integer linear programming (ILP) which includes the link formulation and path formulation. To reduce the computing complexity, a new routing algorithm based on the consideration of backup path is proposed. The software experiment is run and shows that the gain is about 20% when the backup path needs to be setup. At last we propose a new RWA algorithm to adapt the FLDs and dynamic traffic into the wavelength routing network.

Algorithm for updating key forms the basis for any group key management (GKM) system. Existing hierarchical key tree based GKMs are complex, having high computational and communication overhead complexities up to O(logn). In this paper an improved multicast key management (MKM) system is presented, in which only few keys used for encrypting group key are renewed with both backward and forward secrecy being perfectly guaranteed. Furthermore, a mechanism named OWFDL (one-way function and double-lock encryption based) was introduced to reduce the complexities, in which an efficient algorithm of rekeying the group when many members join and leave in a batch interval is proposed. The results show that the OWFDL scheme can reduce both to O(1), which assures perfect multicast secrecy, that is not possible with the previous key management schemes.

In this paper we consider the problem of reconfiguration of virtual topology on wavelength-routed optical ring networks with dynamic changes in traffic. Unlike previous studies on this problem, we select the traffic needing rerouting and the most congested link load as our objectives. In addition, we consider the traffic with both rapid and slow changes. When the traffic changes slowly, we will select the routing strategy based on the shortest path, taking into account the cost caused by the reconfiguration and choose a heuristic similar to generalized least processing time (GLPT) as the wavelength assignment algorithm. Our consideration does make sense in that the strategies mentioned above enable the virtual topology to be updated to follow the change in traffic closely, while considering some certain cost of the reconfiguration. On the other hand, if the traffic changes dramatically, we choose the strategy of adding lightpaths or deleting lightpaths(LAD). It is clear that this strategy will result in a smaller change in the virtual topology. This implies the reconfiguration will lead to a lower cost. Numerical results show that our approach outperforms the previous ones and that we can set up a network with no change in our physical topology, which can sustain various patterns of traffic.

EPON is an emerging local subscriber access network that consists of low-cost point-to-multipoint fiber infrastructure with Ethernet. Because several ONUs use a single shared medium in EPON, it is important to control upstream traffic. An EPON allocates upstream bandwidth to ONUs using a request/permit mechanism. In this paper, we proposed a new DBA algorithm supporting multiple priority queues and evaluated its performance. The key performance of proposed algorithm through simulation is described and the simulation results show the algorithm is fair and feasible.

Ethernet passive optical network (EPON) is a promising local subscribe access architecture that combines low cost point-to-multipoint passive optical network with Ethernet. One of the key problems in EPON is how to allocate the upstream bandwidth effectively. In the paper we present a dynamic bandwidth allocation scheme that will incorporates the priority operation into medium access control (MAC) protocol design to achieve quality of service (QoS) for subscribers. The scheme minimizes packet delay and delay variation for high priority traffic to ensure Qos through introducing a fixed cycle time. Simulation result proves that.

To reduce the searching cost is the key of improving the performance of P2P network. We propose a search algorithm for unstructured P2P network, which is based on query agent. The algorithm can adaptively control the scale of query by satisfactory degree. We carry out simulations under two query-replication modes and the simulation results show that the algorithm can effectively decrease the redundant cost. The mobility of query agent assures that the repeat messages are trivial.

In this paper, a new type of p-cycles, multi-granularity flow p-cycles (MFP), is presented. A two-step approach is developed for the design of MFP with span failure recovery. Our simulation results show that MFP are more capacity-efficient than conventional span-protecting p-cycles (SP) and flow p-cycles (FP). The impacts of network topology on performance of SP, FP and MFP are evaluated. In this paper, we also propose a new p-cycle pre-selection strategy. Results show that our strategy is more efficient than two existing pre-selection strategy in p-cycle networks.

Resilient Packet Ring (RPR) is a new MAC protocol within metropolitan area network, and Operation, Administration and Maintenance (OAM) function can improve network reliability, simplify network operation and reduce network operating costs. This paper discusses that by sending and receiving OAM control frames OAM function can be implemented in the RPR network. It puts forwards a new hybrid performance management scheme for monitoring the performance on the ring. First an OAM frame is transmitted through the whole ring periodically. If the degradation has happened, the source station can choose serial checking, parallel checking or parallel checking with multicast method to locate the site of the degradation, and consequently QoS is guaranteed. The results of simulations show how to balance bandwidth utilization and station overhead with different methods.

This paper describes the design considerations and architecture of a Generalized Multi-Protocol Label Switching (GMPLS)-based scalable control plane that we are prototyping for optical packet switched (OPS) networks. Functional components of the control plane include a user network interface (UNI), optical label coding, multi-layer routing/traffic engineering algorithm and integrated signaling protocol. Initial implementation and experimentation has demonstrated the feasibility of our prototype as a testbed for various control schemes for OPS networks. One key element of the architecture proposed is the use of external MPLS labeling controlled by the UNI. This proposal reduces the load on the OPS domain header processing while having little impact on the MPLS domain.

Multicast applications have attracted more and more attention due to more efficient bandwidth usage and the increasing popularity of the point-multipoint multimedia applications. Supporting multicast directly at physical layer, namely optical multicasting, can provide improved performance over electronic multicasting. In this paper, a new Mixed Integer Linear Programming algorithm is proposed to resolve the light-tree routing and wavelength assignment problem with wavelength continuity constraints. The node in our system is limited branching and power-efficient multicast capable OXC. The node constraints are considered in our model formulation. Numerical results are given by using representative examples and discussed to show the efficiency of our algorithm.

An algorithm is presented in this paper to implement Dynamic Bandwidth Allocation of Gigabit-capable Passive Optical Network(GPON) which is one of the key technologies in GPON, and simulation result is shown to demonstrate the algorithm’s performance.

Disruptive all-optical switching and ultra-long reach WDM transmission technologies enable all-optical networks without OEO conversion along paths that can span thousands of kilometers. Coupled with advances in automatic discovery, distributed routing and signaling mechanisms, the next generation all-optical intelligent networks (AOIN) has been one of the most promising technology innovations. Based on the particular PXC structure, this paper addresses in detail the process of AND in AOIN, which includes three major periods. As far as the simulation is concerned, a GMPLS-based software platform built by us can demonstrate the AOIN performance, and a meaningful result is derived when comparing the speeds of AND under different monitoring schemes.

Optical Burst Switching (OBS) has been proposed to be a competitive switching technology for DWDM networks. However, the abrupt burst for large traffic loads will result in unbearable performance degradation in the traditional JET network. To combat with the load sensitivity, a wavelength-routed OBS (WR-OBS) network architecture was confirmed effective. In this paper, we have proposed a distributed control WR-OBS network architecture combining JET protocol with a two-way reservation method. Based on this architecture, two new resource reservation protocols are compared by evaluating their performance. Then based on the Reduced Load Approximation, which is often used in the analysis of the circuit switch network, a novel analysis method was derived to evaluate the blocking performance of the network. And then we have developed a simulation model to evaluate the performance of the network and validate the proposed analysis method. Analytical and simulation results have indicated that the proposed architecture outperforms the traditional JET OBS network in burst blocking probability with a good trade-off between efficiency and complexity.

The combination of optical fiber and coaxial cable is referred to as Hybrid Fiber Coax (HFC) network. The HFC CATV network is considered a promising infrastructure and the most general access path to connect the residential with the Internet. The advantages of the HFC network include using a less number of signal amplifiers, furnishing more bandwidth, and reducing signal attenuation. In this paper, we reported the realization of the Broadband Information System Based on HFC CATV Networks. In such a system, the characteristics of upstream channel noise are the most important. We studied the causes of the noise and interference, and discussed the solution by using Synchronous Code Division Multiple Access (S-CDMA), trellis coded modulation and Interleaver. We used the cable modem equipment based on S-CDMA for the access method, adopted 1550nm and 1310nm fiber optic cables to combined a HFC CATV network and provided Internet services. This Scheme has been used in several districts in Tianjin. It is welcomed for its high-performance and cheaper in prices.

The increasing switching capacity brings the optical node with considerable complexity. Due to the limitation in cost and technology, an optical node is often designed with partial switching capability and partial resource sharing. It means that the node is of blocking to some extent, for example multi-granularity switching node, which in fact is a structure using pass wavelength to reduce the dimension of OXC, and partial sharing wavelength converter (WC) OXC. It is conceivable that these blocking nodes will have great effects on the problem of routing and wavelength assignment. Some previous works studied the blocking case, partial WC OXC, using complicated wavelength assignment algorithm. But the complexities of these schemes decide them to be not in practice in real networks. In this paper, we propose a new scheme based on the node blocking state advertisement to reduce the retry or rerouting probability and improve the efficiency of routing in the networks with blocking nodes. In the scheme, node blocking state are advertised to the other nodes in networks, which will be used for subsequent route calculation to find a path with lowest blocking probability. The performance of the scheme is evaluated using discrete event model in 14-node NSFNET, all the nodes of which employ a kind of partial sharing WC OXC structure. In the simulation, a simple First-Fit wavelength assignment algorithm is used. The simulation results demonstrate that the new scheme considerably reduces the retry or rerouting probability in routing process.

A novel method called data plane shared ring scheme (DPSRS) is proposed in IP over automatic switched transport networks (ASTN) overlay network. Based on the novel scheme, explicit analysis on related network architecture, edge node structure and operation process are made. For decision of bandwidth restriction parameters two algorithms are also proposed.