This paper explores the positive effects of the new multi protocol label switching (MPLS) routing platform in IP networks. In particular, novel node numbering algorithms based upon graceful numbering of trees are presented. The first part presents the application of the well-known graceful numbering of spanning caterpillars to the MPLS multicast routing problem. In the second part of the paper, the numbering algorithm is adjusted for the case of unicast routing in the framework of IP-over-WDM optical networks using MPLS, e.g., particularly lambda-labeling and multi protocol lambda switching.

The emerging IP over WDM will be a new layer structure with a help of MPLS and OXC. To use WDM's large transmission capacity effectively, we design IP over WDM networks that reconfigure IP routing and lightpaths every day or month. Using integer linear programming, we formulate a problem that maximizes the network throughput; that is, the sum of the bandwidths allocated to an end-to-end node pair. This formulation is applicable to a network that is composed of twenty nodes and eight wavelengths. In such a network, we show maximization level of the network throughput derived from our formulation.

In this paper, we propose an objective function to evaluate the performance of IP over WDM networks and a traffic engineering method which considers both IP and WDM layers. Since WDM engineering and traffic engineering used in the IP layer have similar functional goals, both functions should be harmonized to improve the network performance. We formulate an integer optimization problem to describe the tradeoff relation between control functions of both layers and find the unique optimal solution. The proposed traffic engineering method tries to minimize the total cost of both IP and WDM layer. The performance of the proposed method is analyzed and evaluated through simulation.

Protection switching interactions in wide-area networks need to interoperate with each other in order to restore a wide variety of services, provide survivability to mission-critical applications and accommodate an evolving network infrastructure. Without some coordination between restoration mechanisms, an outage duration would be lengthened as methods assigned to each layer interfere with each other or the network would be locked up in a deadlocked state that never converges to a new topology. A set of control policies can be specified to coordinate between restoration mechanisms in a network that spans multiple layers and regions. These control policies are expressed as rules, and are collectively denoted as the escalation strategy. The escalation strategy can be provisioned by a network manager and is implemented as a distributed coordination protocol between peer recovery agents in the nodes. As rules for coordinating between restoration mechanisms are formalized, a mathematical proof could be provided to prove that the network does indeed converge to a new topology.

A fundamental task of the optical layer in modern telecommunication systems consists of providing a fast protection mechanism against possible faults in the network. A particularly attractive protection technique in the optical layer is the so called shared line protection, in which network lines are protected using shared resources. A previous work of the authors formally describes the problem of minimizing the total wavelength mileage, or (lambda) -mileage, necessary in a Wavelength Routing network with arbitrary topology to provide shared line protection. This paper presents an efficient approach to solving the above wavelength mileage problem.

This paper newly proposes a self-healing architecture in all- optical WDM networks based on virtual embedded multiple rings (Virtual Multiple Self Healing Rings: VM-SHR). Focusing upon the network design aspect of the proposed architecture, this paper describes design methodologies for VM-SHR networks. For two major problems in all-optical WDM network design, that is, the connection routing and wavelength assignment problems, we first established solution models based on mathematical programming formulation, each of which can be solved by common integer programming algorithms, respectively. In addition, we also developed an efficient heuristic algorithm for the wavelength assignment problem. Their usefulness and performance are demonstrated through the extensive simulation results.

We present a new approach to WDM multifiber network design and optimization under static traffic aimed to minimize the number of fibers in the network. The heuristic optimization comprises an initial setup of all the demanded lightpaths and an iteration during which the network is progressively improved by rerouting lightpaths and consequently eliminating fibers with a high number of unused wavelengths. In our approach static traffic is actually managed with the same techniques as dynamic traffic by suitably sorting the static connection demands and offering them in sequence to the network. The employment of the multifiber layered graph as a template to represent the network states allows to compare a very wide range of routing, fiber and wavelength assignment criteria, link metrics, wavelength-conversion environments and so on. In this paper we will show and discuss the optimization of a particular case-study network and we will analyze the influence of various design parameters on the final results.

The bandwidth requirements of the Internet are increasing every day and there are newer and more bandwidth-thirsty applications emerging on the horizon. Wavelength division multiplexing (WDM) is the next step towards leveraging the capabilities of the optical fiber, especially for wide-area backbone networks. Virtual private networks are seen as the killer applications of tomorrow, and a great amount of research and development effort is being channeled into finding efficient ways for implementing them. In our work, we introduce the notion of VPNs operating over optical wavelength-routed WDM backbone networks. We formulate the problem of VPN design over optical WDM networks as an integer linear problem. A new method to deal with traffic from diverse VPNs is introduced and its performance analyzed.

In this paper, we study the problem of employing virtual private network (VPN) over wavelength division multiplexing networks to facilitate the guarantee of diverse quality of service requirements of different VPNs. A wavelength routed backbone network is considered. A VPN is specified by the desired logical topology and an a priori traffic matrix.

In this paper, we propose path accommodation methods for unidirectional rings based on an optical compression TDM technology. We first derive a theoretical lower bound on the numbers of slots and frames, in order to allocate all paths among nodes. Three path accommodation algorithms for the all-optical access are next proposed to achieve the lower bound as close as possible. Path splitting is next considered to improve the traffic accommodation. Finally, we analyze the packet delay time for given numbers of slots/frames, which are decided by our proposed algorithms. Numerical examples are also shown to examine the effectiveness of our proposed algorithms including path accommodation and path-splitting methods.

Current optical networks are migrating towards WDM-based transport between traditional electronic multiplexers/demultiplexers, routers and switches. The ONRAMP Program addresses technologies, architectures and designs of future high performance data networks. This paper focuses on WDM aware IP networks with emphasis on access networks. The access network can be divided into the feeder network and the distribution network. Generally bandwidths of fiber closer to the end user is less precious. Thus in the distribution network, we propose the use of all passive optical components and remotely pumped amplifiers trading bandwidth assignment efficiencies for lower costs. We will examine the design of the distribution network from the view of services to be supported by the network and its desirable properties. Generic physical topologies will be considered and implementation optics presented.

As all-optical networks emerge as the high-speed networking standard, the performance of optical networks for mixed traffic requires investigation. Currently, traffic on optical networks varies from monomedia or one type of session only, to multimedia, in which multiple monomedia sessions are presented as one multimedia session. Today's internet offers elementary examples of this, with modest video and audio clips accompanying data. The performance of multimedia sessions over optical networks has not met with user expectations, primarily due to network latency as a result of the underlying protocols in the network infrastructure. This protocol incompatibility has the most significant impact on multimedia IP traffic on optical networks. In this paper, a new structure for the transport of dissimilar packet and non-packet information over optical networks is presented.

This paper presents and discuses three main points on the design and planning of transport networks. The first one is the pure backbone design mainly based on WDM technology. Several methods are discussed and some comparative results are given based on real-test cases. The second point is related to the protection issues in a multi-layer transport network. We will examine the different constraints, which have to be taken into account and we will discuss the main parameter as the hold-off time we can used to face interaction between SDH/SONET and WDM layers. The final point is introducing the management of the Optical Channel layer with the description of a possible attracting and flexible framing structure.

We consider a class of WDM (Wavelength Division Multiplexing) all-optical networks, the well known broadcast-and-select networks, where each transmitter can broadcast its message on a different wavelength, and the receivers can be tuned to receive the desired message. We consider fast (resp. slow) tunability on the transmitters (resp. receivers). It has been shown in that in such a context whenever the traffic load is not well balanced among the different channels then the network performance is poor. Here, we study the performance of some variants of the well known LPT (Longest Processing Time) algorithm with lookahead for the reconfiguration of networks of this class in response to traffic changes.

IN this paper,l we present a reliable WDM (Wavelength-Division Multiplexing) multicast protocol in optical burst-switched (OBS) networks. Since the burst dropping (loss) probability may be potentially high in a heavily loaded OBS backbone network, reliable multicast protocols that have developed for IP networks at the transport (or application) layer may incur heavy overheads such as a large number of duplicate retransmissions. In addition, it may take a longer time for an end host to detect and then recover from burst dropping (loss) occurred at the WDM layer. For efficiency reasons, we propose burst loss recovery within the OBS backbone (i.e., at the WDM link layer). The proposed protocol requires two additional functions to be performed by the WDM switch controller: subcasting and maintaining burst states, when the WDM switch has more than one downstream on the WDM multicast tree. We show that these additional functions are simple to implement and the overhead associated with them is manageable.

Recent developments in all-optical networking and wavelength division multiplexing technologies allow for the support of optical multicasting, a missing feature towards the optical Internet. In this paper we propose a protocol to construct source-rooted WDM multicast trees. The protocol works under dense mode multicasting routing IP protocols and supports network nodes with different degrees of light splitting, wavelength conversion, and add/drop capabilities.

Future broadband networks must support integrated services and offer flexible bandwidth usage. In our previous work, we explore the optical link control layer on the top of optical layer that enables the possibility of bandwidth on-demand service directly over wavelength division multiplexed (WDM) networks. Today, more and more applications and services such as video-conferencing software and Virtual LAN service require multicast support over the underlying networks. Currently, it is difficult to provide wavelength multicast over the optical switches without optical/electronic conversions although the conversion takes extra cost. In this paper, based on the proposed wavelength router architecture (equipped with ATM switches to offer O/E and E/O conversions when necessary), a dynamic multicast routing algorithm is proposed to furnish multicast services over WDM networks. The goal is to joint a new group member into the multicast tree so that the cost, including the link cost and the optical/electronic conversion cost, is kept as less as possible. The effectiveness of the proposed wavelength router architecture as well as the dynamic multicast algorithm is evaluated by simulation.

This paper addresses the issue of capacity fairness in WDM networks with traffic grooming capabilities, supporting lower- rate circuit-switched traffic streams. Traffic grooming in WDM networks, is defined as the act of multiplexing, demultiplexing and switching lower rate traffic streams onto higher capacity lightpaths. In such a network, in addition to add/drop and full wavelength switching features, some or all of the network nodes can be provided with the capability to switch lower-rate traffic streams from one wavelength on an input port to another wavelength on an output port. Call requests arrive randomly and can request a lower-rate traffic connection to be established between the node pair. The call requests that ask for capacity nearer to the full wavelength capacity are bound to experience higher blocking than those that ask for a smaller fraction. This difference in loss performance is more pronounced as the traffic switching capability of the network is increased. In this paper, we study the capacity fairness of existing dynamic wavelength assignment algorithms.

This paper considers a problem concerning the blocking analysis of all-optical networks with arbitrary topologies and unconstrained routing. That is, we consider designs where the underlying routing and wavelength assignment algorithm selects a route from the set of all available routes for connection setup. Unconstrained routing is a natural strategy to consider for harnessing all-optical bandwidth; its strength has been demonstrated by recent simulation studies. Unfortunately, exact computations of blocking for unconstrained routing is computationally intensive due to the presence of a potentially exponential number of paths in mesh networks. In fact, exact computation of the source-destination blocking probability for unconstrained routing is #P-complete, even with the simplifying assumption that the per-link blocking distribution can be estimated a priori.

Dynamic wavelength assignment has been known to provide significant improvement of wavelength utilization in a wavelength routed all-optical network. This paper considers the dynamic wavelength assignment problem and presents a distributed algorithm to solve it. The presented algorithm shows a significant improvement over existing schemes by reducing the blocking probability often by an order of magnitude. The algorithm is based on the assumption of the availability of adequate processing power at the routers and lightpaths are constructed dynamically when two end-nodes need to communicate. The effectiveness of the algorithm is compared to the schemes in terms of blocking probability and set-up time through simulations.

Increasing the capacity and intelligence of the next-generation Internet requires the application of optical technologies to switching nodes as well as transmission lines, and the development of advanced network architectures with end-to-end connection setup processing at the source node and autonomous routing at intermediate nodes. In this paper, one configuration of a photonic IP switching network is presented, where a set of multiple encoding CDM codes is utilized as routing information. Also, the node configurations of this network are proposed. In addition, we show characteristics between the BER and the number of nodes in each layer of the network.

This paper presents a comparative study of contention-resolution schemes based on wavelength, time, and space domains in an unslotted optical packet-switched network with a large irregular mesh topology consisted of 15 nodes. For the first time, to the best of our knowledge, we investigated the effect of selective deflection and limited wavelength conversion. Features and performances of different combinational schemes are listed and compared. While simulation results show the effectiveness of wavelength conversion for resolving contentions over optical buffering and space deflection, physical explanations of the different effectiveness in resolving contentions of these schemes are also discussed.

A high-speed optical packet switching node using the WDM all optical switch structure is studied in this paper. Tele-traffic performance of an all-optical IP packet router is simulated under the self-similar (bursty) traffic condition. Four different control algorithms are investigated for the performance and complexity. A simple round-robin algorithm cannot attain an acceptable performance. Finding minimum buffer occupancy and sorting the packets by length are methods used to improve the IP router performance.

Given the significant progress made and the continuing advances expected in the optical networking technology, it becomes attractive to build a future Optical Internet that natively supports bursty IP datagrams. Burst switching WDM optical networks are touted as suitable network architectures for future Optical Internet backbones. However, the lack of optical processing capabilities results in increased burst blocking probability, which in turn leads to very limited network performance. Efficient contention resolution method is therefore necessary. Based on discussions of the state of the art of recent optical technologies, a burst optical deflection routing protocol is proposed. The idea of this approach is to use idle optical links as fiber delay lines for contention resolution. Simulation results show that the proposed protocol is available solution for effectively reducing the blocking probability and increasing the performance of burst switching WDM optical networks.

This paper considers capacity provisioning for service and restoration in a WDM optical network which provides lightpaths to higher layer networks. We develop a framework for jointly assigning wavelengths to service and restoration paths for all failures. For various types of path-based and link-based restoration methods, the problem is reduced to a vertex coloring problem in a single graph. Using this framework, we present routing and wavelength assignment algorithms for service and restoration with varying capacity/restoration time tradeoffs, and evaluate their performance through simulations. We also evaluate the difference in wavelength requirement under on-line incremental traffic and off-line traffic based on the properties of the conflict graph.

IP (Internet Protocol) over WDM networks where IP packets are directly carried on the WDM network is expected to offer an infrastructure for the next generation Internet. For IP over WDM networks, a WDM protection mechanism is expected to provide a highly reliable network (i.e., robustness against the link/node failures). However, conventional IP also provides a reliability mechanism by its routing function. We thus need to treat functional partitioning or functional integration for IP over WDM networks with high reliability. In this paper, we first formulate an optimization problem for designing IP over WDM networks with protection functionalities of WDM networks, by which we can obtain IP over WDM networks with high reliability. Our formulation results in a mixed integer linear problem (MILP). However, it is known that MILP can be solved only for a small number of variables, in our case, nodes and/or wavelengths. We therefore propose two heuristic algorithms, min-hop-first and largest-traffic-first approaches in order to assign the wavelength for backup lightpath. Our results show that the min- hop-first approach takes fewer wavelengths to construct the reliable network, that is, all of lightpaths can be protected using the WDM protection mechanism. However, our largest-traffic- first approach is also a good choice in the sense that the approach can be saved the traffic volume increased at the IP router by the link failure.

Service availability is an indispensable requirement for many current and future applications over the Internet and hence has to be addressed as part of the optical QoS service model. Network service providers can offer varying classes of services based on the choice of protection employed which can vary from full protection to no protection. Based on the service classes, traffic in the network falls into one of the three classes viz., full protection, no protection and best-effort. The network typically relies on the best-effort traffic for maximizing revenue. We consider two variations on the best-effort class, (1) all connections are accepted and network tries to protect as many as possible and (2) a mix of protected and unprotected connections and the goal is to maximize revenue. In this paper, we present a mathematical formulation, that captures service differentiation based on lightpath protection, for revenue maximization in a wavelength routed backbone networks. Our approach also captures the service disruption aspect into the problem formulation, as there may be a penalty for disrupting currently working connections.

The complexity of multi-layer resilience schemes and the high availability parameters provided by these methods require efficient availability analysis techniques, as well. In the paper the effectiveness of stratified sampling in availability analysis is demonstrated by the investigation of different protection strategies in SDH/SONET over WDM networks. The numerical results show that the application of stratified sampling results in sufficiently accurate estimation of the performance degradation with some hundreds of sample states even in the case of very efficient resilience schemes.