Efficient routing and wavelength assignment for reconfigurable WDMnetworks

Through the use of configurable wavelength-division-multiplexing (WDM) technology including tunable optical transceivers and frequency selective switches, next-generation WDM networks will allow multiple virtual topologies to be dynamically established on a given physical topology. For N node P port networks, we determine the number of wavelengths required to support all possible virtual topologies (PN lightpaths) on a bidirectional ring physical topology. We show that if shortest path routing is used, approximately N wavelengths are needed to map N lightpaths. We then present novel adaptive lightpath routing and wavelength assignment strategies that reduce the wavelength requirements to [(N/2)] working wavelengths per port for protected networks and [(N/3)] wavelengths in each direction per port for unprotected networks. We show that this reduced wavelength requirement is optimal in the sense that it is the minimum required to support the worst case logical topology. Furthermore, we prove that a significant number of logical topologies require this minimum number of wavelengths. We also develop joint routing and wavelength assignment strategies that not only minimize the number of wavelengths required to implement the worst case logical topologies but also reduce average wavelength requirements. Finally, methods for extending these routing and wavelength assignment results to general two-connected and three-connected physical topologies are presented     

The role of switching in reducing the number of electronic ports in WDM networks

We consider the role of switching in minimizing the number of electronic ports [e.g., synchronous optical network (SONET) add/drop multiplexers] in an optical network that carries subwavelength traffic. Providing nodes with the ability to switch traffic between wavelengths, such as through the use of SONET cross-connects, can reduce the required number of electronic ports. We show that only limited switching ability is needed for significant reductions in the number of ports. First, we consider architectures where certain "hub" nodes can switch traffic between wavelengths and other nodes have no switching capability. For such architectures, we provide a lower bound on the number of electronic ports that is a function of the number of hub nodes. We show that our lower bound is relatively tight by providing routing and grooming algorithms that nearly achieve the bound. For uniform traffic, we show that the number of electronic ports is nearly minimized when the number of hub nodes used is equal to the number of wavelengths of traffic generated by each node. Next, we consider architectures where the switching ability is distributed throughout the network. Such architectures are shown to require a similar number of ports as the hub architectures, but with a significantly smaller "switching cost." We give an algorithm for designing such architectures and characterize a class of topologies, where the minimum number of ports is used. Finally, we provide a general upper bound on the amount of switching required in the network. For uniform traffic, our bound shows that as the size of the network increases, each traffic stream must be switched at most once in order to achieve the minimum port count.     

Optimal all-to-all broadcast in WDM optical networks with breakdown or power-off transceivers

All-to-all broadcast is an interesting special case of the packet transmission scheduling in which every pair of nodes has exactly one packet to be transferred. This paper considers the all-to-all broadcast problem in wavelength division multiplexed (WDM) optical star network with some breakdown or power-off transceivers. For reaching high data transmission rates, we will focus the problem on the all-optical scheduling where the traffic reaches its destination in single-hop without being converted to electronic form. Each transmitter is tunable with an associated tuning delay and each receiver is fixed-tuned to one of available wavelengths. In this model, we study two kinds of all-to-all broadcast problems depending on whether each node transmits packets to all nodes including or except itself. We identify the lower bound of the scheduling length for each kind of problems and propose single-hop scheduling algorithms to find the optimal solution in both terms of arbitrary number of wavelengths and value of tuning latency.     

On the design of node architectures and MAC protocols for optical burst-switched ring networks

Efficient multiple-token-based MAC protocols have been proposed for optical burst-switched (OBS) unidirectional ring networks using a TT-TR-based node architecture in our previous research. However, the unidirectional OBS ring network is difficult to scale to larger networks. As wavelengths accessibilities are dominated by tokens, network performance is restricted by the frequency of capturing a token. If the network is too large, it takes a long time for tokens to rotate. Thus, a destination queue may wait for a long time to be served, which results in large queuing delays and inefficiency of network resource utilization. In order to improve network efficiency and scalability for OBS ring networks using multiple tokens, this work is extended to a bidirectional ring system that uses the tunable transmitter and tunable receiver (TT-TR)-based node architecture with two pairs of transceivers, so that each queue can be served by tokens from both directions. Furthermore, two kinds of node architectures differing in sharing the two pairs of transceivers, either shared or not, are proposed. Then, two MAC protocols considering different queue scheduling algorithms are proposed for the ring network using the proposed node architectures, in order to use the network resources more efficiently. They are improved from general round-robin (GRR) and termed as half-ring round-robin (HfRR) and co-work round-robin (CoRR), respectively. The network performance of the two proposed node architectures and the two proposed MAC protocols for the networks using them as well as the network scalability are evaluated with the OPNET simulator.     

Reconfiguration and Dynamic Load Balancing in Broadcast WDM Networks*

In optical WDM networks, an assignment of transceivers to channels implies an allocation of the bandwidth to the various network nodes. Intuition suggests, and our recent study has confirmed, that if the traffic load is not well balanced across the available channels, the result is poor network performance. Hence, the time-varying conditions expected in this type of environment call for mechanisms that periodically adjust the bandwidth allocation to ensure that each channel carries an almost equal share of the corresponding offered load. In this paper we study the problem of dynamic load balancing in broadcast WDM networks by retuning a subset of transceivers in response to changes in the overall traffic pattern. Assuming an existing wavelength assignment and some information regarding the new traffic demands, we present two approaches to obtaining a new wavelength assignment such that (a) the new traffic load is balanced across the channels, and (b) the number of transceivers that need to be retuned is minimized. The latter objective is motivated by the fact that tunable transceivers take a non-negligible amount of time to switch between wavelengths during which parts of the network are unavailable for normal operation. Furthermore, this variation in traffic is expected to take place over larger time scales (i.e., retuning will be a relatively infrequent event), making slowly tunable devices a cost effective solution. Our main contribution is a new approximation algorithm for the load balancing problem that provides for tradeoff selection, using a single parameter, between two conflicting goals, namely, the degree of load balancing and the number of transceivers that need to be retuned. This algorithm leads to a scalable approach to reconfiguring the network since, in addition to providing guarantees in terms of load balancing, the expected number of retunings scales with the number of channels, not the number of nodes in the network.     

Reducing electronic multiplexing costs in SONET/WDM rings with dynamically changing traffic

In this paper, we consider traffic grooming in WDM/SONET ring networks when the offered traffic is characterized by a set of traffic matrices. Our objective is to minimize the cost of electronic add/drop multiplexers (ADMs) in the network, while being able to support any offered traffic matrix in a rearrangeably nonblocking manner. We provide several methods for reducing the required number of ADMs for an arbitrary class of traffic matrices. We then consider the special case where the only restriction on the offered traffic is a constraint on the number of circuits a node may source at any given time. For this case, we provide a lower bound on the number of ADMs required and give conditions that a network must satisfy in order for it to support the desired set of traffic patterns. Circuit assignment and ADM placement algorithms with performance close to this lower bound are provided. These algorithms are shown to reduce the electronic costs of a network by up to 27%. Finally, we discuss extensions of this work for supporting dynamic traffic in a wide-sense or strict sense nonblocking manner as well as the benefits of using a hub node and tunable transceivers. Much of this work relies on showing that these grooming problems can often be formulated as standard combinatorial optimization problems.     

A cost effective approach for survivable multicast traffic grooming in bidirectional multiwavelength ring networks

In this paper, we address the problem of survivable multicast traffic grooming in WDM bidirectional ring networks. The rapid growth of multicast applications such as video conferencing, distance learning, and online auction, has initiated the need for cost-effective solutions to realize multicasting in WDM optical networks. Many of these applications, being time critical and delay sensitive, demand robust and fault-tolerant means of data communication. The end user traffic demands in metro environment are in fractional bandwidth as compared to the wavelength channel capacity. Providing survivability at connection level is resource intensive. Hence cost-effective solutions that require minimum resources for realizing survivable multicasting are in great demand. In order to realize multicast traffic grooming in bidirectional ring networks, we propose a node architecture based on Bidirectional Add Drop Multiplexers (BADM) to support bidirectional add/drop functionality along with traffic duplication at each node. We also propose two traffic grooming algorithms, namely Survivable Grooming with Maximum Overlap of Sessions (SGMOS) and Survivable Grooming with Rerouting of Sessions (SGRS). Extensive simulation studies reveal that the proposed algorithms consume minimum resources measured in terms of BADM grooming ports, backup cost, and wavelengths.     

Improved Approaches for Cost-Effective Traffic Grooming in WDM Ring Networks: Uniform-Traffic Case*

To fully utilize the capabilities of a SONET/ADM network, traffic grooming is needed to multiplex a number of lower-rate traffic streams into a higher-rate stream, and vice versa. Although the capacity of a SONET ring network can be upgraded by operating it over multiple wavelengths, the corresponding network design may be costly if it employs a large number of ADMs. A cost-effective design attempts to minimize the total number of ADMs used in the network while carrying the offered traffic. We introduce and evaluate the performance characteristics of two new traffic-grooming approaches for WDM ring networks, called single-hop and multihop. Our single-hop implementation uses the simulated-annealing heuristic. After placing all the traffic on virtual circles, we group the circles in order to reduce the number of ADMs in the network. Our multihop implementation places an ADM at each node based on the requested traffic in the traffic-demand matrix; then, it tries to groom the wavelengths which can be groomed. We select one of the nodes to be the hub node which has an ADM for each wavelength. The hub node, therefore, can bridge traffic between all of the wavelengths. Each algorithm is specified and illustrated by a simple example. Our results demonstrate that it is beneficial to use a single-hop approach based on simulated annealing for a small grooming ratio, but for a large grooming ratio and node number, we advocate the use of the multihop approach.     

Optimal dimensioning of the WDM unidirectional ECOFRAME optical packet ring

To efficiently support the high rate and the high dynamicity of the traffic in metro networks, an optical packet-switched WDM ring, named ECOFRAME, is proposed. The key features of the proposed ring are optical transparency and statistical multiplexing of optical packets on parallel WDM channels. Such features can be exploited by properly allocating wavelengths and receivers. This paper aims to optimally dimension the unidirectional ECOFRAME rings. The dimensioning at minimum cost (i.e., for wavelengths and receivers) is modeled with an mixed-integer linear programming formulation. An heuristic algorithm is also proposed, and its performance is compared against the optimal solutions and bounds. When considering the receiver and wavelength cost, results indicate that trading the wavelengths for receivers allows cost saving of up to 75% with respect to WDM optical packet rings with a single dedicated wavelength per node (i.e., single receiver at each node).     

Uniform waveband assignment in optical mesh networks

The cost of an optical network in wavelength division multiplexing (WDM) networks can be reduced using optical reconfigurable optical add/drop multiplexers (ROADMs), which allow traffic to pass through without the need for an expensive optical-electro-optical (O-E-O) conversion. Waveband switching (WBS) is another technique to reduce the network cost by grouping consecutive wavelengths and switching them together using a single port per waveband. WBS has attracted the attention of researchers for its efficiency in reducing switching complexity and therefore cost in WDM optical networks. In this paper, we consider the problem of switching wavelengths as non-overlapping uniform wavebands, per link, for mesh networks using the minimum number of wavebands. Given a fixed band size b _s , we give integer linear programming formulations and present a heuristic solution to minimize the number of ROADMs (number of wavebands) in mesh networks that support a given traffic pattern. We show that the number of ROADMs (or number of ports in band-switching cross-connects) can be reduced significantly in mesh networks with WBS compared to wavelength switching using either the ILP or the heuristic algorithm. We also examine the performance of our band assignment algorithms under dynamic traffic.     

A slotted access control protocol for metropolitan WDM ring networks

In this study we focus on the serious scalability problems that many access protocols for WDM ring networks introduce due to the use of a dedicated wavelength per access node for either transmission or reception. We propose an efficient slotted MAC protocol suitable for WDM ring metropolitan area networks. The proposed network architecture employs a separate wavelength for control information exchange prior to the data packet transmission. Each access node is equipped with a pair of tunable transceivers for data communication and a pair of fixed tuned transceivers for control information exchange. Also, each access node includes a set of fixed delay lines for synchronization reasons; to keep the data packets, while the control information is processed. An efficient access algorithm is applied to avoid both the data wavelengths and the receiver collisions. In our protocol, each access node is capable of transmitting and receiving over any of the data wavelengths, facing the scalability issues. Two different slot reuse schemes are assumed: the source and the destination stripping schemes. For both schemes, performance measures evaluation is provided via an analytic model. The analytical results are validated by a discrete event simulation model that uses Poisson traffic sources. Simulation results show that the proposed protocol manages efficient bandwidth utilization, especially under high load. Also, comparative simulation results prove that our protocol achieves significant performance improvement as compared with other WDMA protocols which restrict transmission over a dedicated data wavelength. Finally, performance measures evaluation is explored for diverse numbers of buffer size, access nodes and data wavelengths.     

Cost-effective traffic grooming in WDM rings

We provide network designs for optical add-drop wavelength-division-multiplexed (OADM) rings that minimize overall network cost, rather than just the number of wavelengths needed. The network cost includes the cost of the transceivers required at the nodes as well as the number of wavelengths. The transceiver cost includes the cost of terminating equipment as well as higher-layer electronic processing equipment, which in practice can dominate over the cost of the number of wavelengths in the network. The networks support dynamic (i.e., time-varying) traffic streams that are at lower rates (e.g., OC-3, 155 Mb/s) than the lightpath capacities (e.g., OC-48, 2.5 Gb/s). A simple OADM ring is the point-to-point ring, where traffic is transported on WDM links optically, but switched through nodes electronically. Although the network is efficient in using link bandwidth, it has high electronic and opto-electronic processing costs. Two OADM ring networks are given that have similar performance but are less expensive. Two other OADM ring networks are considered that are nonblocking, where one has a wide-sense nonblocking property and the other has a rearrangeably nonblocking property. All the networks are compared using the cost criteria of number of wavelengths and number of transceivers     

Probabilistic Analysis of Cyclic Packet Transmission Scheduling in WDM Optical Networks

We study the packet transmission scheduling problem with tuning delay in wavelength-division multiplexed (WDM) optical communication networks with tunable transmitters and fixed-tuned receivers. By treating the numbers of packets as random variables, we conduct probabilistic analysis of the average-case performance ratio for the cyclic packet transmission scheduling algorithm. Our numerical data as well as simulation results demonstrate that the average-case performance ratio of cyclic schedules is very close to one for reasonable system configurations and probability distributions of the numbers of packets. In particular, when the number of receivers that share a channel and/or the granularity of packet transmission are large, the average-case performance ratio approaches one. Better performance can be achieved by overlapping tuning delays with packet transmission. We derive a bound for the normalized tuning delay Δ such that tuning delay can be completely masked with high probability. Our study implies that by using currently available tunable optical transceivers, it is possible to build single-hop WDM networks that efficiently utilize all the wavelengths.     

Design of grooming architectures for optical WDM mesh networks: limited grooming with electronic wavelength conversion

In this article, we consider the problem of traffic grooming in optical wavelength division multiplexed (WDM) mesh networks under static traffic conditions. The objective of this work is to minimize the network cost and in particular, the electronic port costs incurred for meeting a given performance objective. In earlier work, we have shown the benefits of limited grooming switch architectures, where only a subset of wavelengths in a network are equipped with expensive SONET Add Drop Multiplexers (SADM) that provide the grooming functionality. In this work, we also consider the wavelength conversion capability of such groomers. This can be achieved using a digital cross-connect (DCS) in the grooming switch to switch low-speed connections between the SADMs (and hence, between wavelengths). The grooming switch thus avoids the need for expensive optical wavelength converters. Based on these observations, we propose a limited conversion-based grooming architecture for optical WDM mesh networks. The local ports at every node in this architecture can be one of three types: an add-drop port, a grooming port that allows wavelength conversion or a grooming port that does not allow wavelength conversion. The problem studied is: given a static traffic model, where should the different ports be placed in a network? We formulate this as an optimization problem using an Integer Linear Programing (ILP) and present numerical results for the same. We also present a heuristic-based approach to solve the problem for larger networks.     

Performance Comparison of Overlay and Peer Models in IP/MPLS over Optical Networks

We study the connection establishment of label switched paths (LSPs), and compare the LSP blocking performance of the overlay and peer models in IP/MPLS over optical networks. We consider two dynamic routing algorithms for the establishment of LSP connections, of which one is for the overlay model and the other is for the peer model. Our investigations on two typical network topologies, namely NSFNET and ARPA2 networks, show that the number of add/drop ports (or transceivers) on optical cross-connects (OXCs) has a significant impact on the LSP blocking performance for both models. We show by computer simulation that in each case, there is a threshold value for the add/drop ratio, which can achieve almost the best blocking performance. This threshold value remains virtually unchanged as the traffic load varies, but it does depend on the network topology and the number of wavelengths per fiber. This will then indicate the number of add/drop ports to be used so that one can achieve a near optimal blocking performance without incurring unnecessarily excessive network costs. Our investigations reveal that the peer model achieves a much better blocking performance than the overlay model when the number of add/drop ports is relatively high, but that this is not always true when the number of add/drop ports is small.     

On service provisioning using light-trails in WDM optical networks with waveband switching

As the number of wavelengths in a single optical fiber increases, so does the number of ports needed for wavelength switching in optical cross-connects (OXCs), which may significantly increase the cost and difficulty associated with controlling large OXCs. Waveband switching (WBS) treats several wavelengths as a bundle that is switched through a single port if they share the same switch route, so that the number of ports needed can be reduced. On the other hand, light-trails in wavelength division multiplexing (WDM) optical networks allow intermediate nodes on established optical paths to access the data paths whereas light-paths only allow two end nodes to access the data paths. Therefore, light-trails offer significantly better flexibility for service provisioning and traffic grooming. In this article, we study service provisioning using light-trails in WDM optical networks with the WBS capability under a static traffic model. For comparison, integer linear programs are formulated for establishing light-trails with and without WBS. Numerical studies show that in certain cases, service provisioning with WBS in light-trail networks can reduce the number of ports needed while providing a more flexible sub-wavelength service provisioning capability. However, contrary to intuition, in most cases applying the WBS technique requires more ports in OXCs in light-trail networks. This study provides insights into the network design problem that applies the WBS technology to light-trail based optical networks.     

Design and Dimensioning of a WDM Mesh Network to Groom Dynamically Varying Traffic

In this paper, we address the traffic grooming problem in WDM mesh networks when the offered traffic is characterized by a set of traffic matrices—a variant of dynamically changing traffic. We justify the need to address this problem in mesh networks and also argue for the validity of our approach to solve this problem. Our primary objective is to design the network in terms of the number of wavelengths and transceivers required to support any offered traffic matrix. We provide a simple and generic framework to minimize the number of transceivers needed in the network. Simulation results have been presented in contrast with a possible approach, to enable comparison with our solution strategy. An ILP formulation of our approach is also presented.     

On the All-to-All Broadcast Problem in Optical Networks1

This paper considers the transmission of uniform deterministic traffic in an optical broadcast-star network using Wavelength Division Multiplexing. Lower bounds are established on the minimum time to exchange information between every node pair in such a network with tunable transmitters and fixed-tuned receivers. Three different scheduling algorithms are developed that are strictly optimal in three regimes of system parameters. The results are applicable to arbitrary tuning delays and arbitrary numbers of wavelength channels, and indicate the existence of a well-defined transition regime from tuning-limited operation to bandwidth-limited operation. Finally, the problem of computing the optimal number of wavelengths is addressed to achieve a schedule with minimum schedule length, and exact solutions are given.     

Light trees: optical multicasting for improved performance inwavelength routed networks

We introduce the concept of a light-tree in a wavelength-routed optical network. A light-tree is a point-to-multipoint generalization of a lightpath. A lightpath is a point-to-point all-optical wavelength channel connecting a transmitter at a source node to a receiver at a destination node. Lightpath communication can significantly reduce the number of hops (or lightpaths) a packet has to traverse; and this reduction can, in turn, significantly improve the network's throughput. We extend the lightpath concept by incorporating an optical multicasting capability at the routing nodes in order to increase the logical connectivity of the network and further decrease its hop distance. We refer to such a point-to-multipoint extension as a light-tree. Light-trees can not only provide improved performance for unicast traffic, but they naturally can better support multicast traffic and broadcast traffic. In this study, we shall concentrate on the application and advantages of light-trees to unicast and broadcast traffic. We formulate the light-tree-based virtual topology design problem as an optimization problem with one of two possible objective functions: for a given traffic matrix, (i) minimize the network-wide average packet hop distance, or (ii) minimize the total number of transceivers in the network. We demonstrate that an optimum light-tree-based virtual topology has clear advantages over an optimum lightpath-based virtual topology with respect to the above two objectives