In the last two decades, optical networks have been potentially considered as the most appropriate solutions for developing high speed backbones of past, current, and future client networks, including asynchronous transfer mode (ATM), synchronous optical networking (SONET) or synchronous digital hierarchy (SDH, SONET/SDH) networks, and internet protocol (IP) networks.
Dense Wavelength Division Multiplexing (DWDM) optical transport networks are the bulk carriage mechanism to convey data, voice, and video over the Global Internet with the potential to fulfill the ever-growing demands for bandwidth. DWDM technology provides an excellent platform to exploit the huge capacity of optical fibre by multiplexing non-overlapping wavelength channels offering multiple terabits per second (Tb/s) transmission rates. The DWDM technique is emerging as a promising technology for the next generation of high-speed communication networks.
Optical networks are prone to failures or may face attacks. When these come unexpectedly in network components, such as link and node failures at the optical layer, they can potentially lead to a catastrophic loss of data and revenue, thus producing an unacceptable deterioration in the delivered quality of service. Therefore, one of the most important optical network design issues is survivability, also called resilience or fault tolerance. This is the ability of a network to provide continuous service at an acceptable level in the presence of different failure scenarios.
Design of optical networks subject to service restoration and survivability requirements has become a crucial issue in network planning, which is known as an NP-hard problem. The integration of resilience into optical core networks is a complicated problem, which requires some redundant resources such as additional bandwidths. Typically in addition to the working lightpaths established between the original and destination node pairs, some spare lightpaths are also established in static or dynamic manners to restore traffic demands in the event of failures at working lightpaths.
There are several approaches for designing resilient optical networks, which are mainly based on protection and restoration architectures. In protection architectures, both working and spare lightpaths are established during configuration of the network for arrival requests, while in the restoration methods, the path planner uses network state variables such as link state variables and wavelength state variables to assign spare lightpaths after the occurrence of failures. Generally, the protection architecture is static, while the restoration architecture is dynamic; consequently the latter is efficient for bandwidth trading, and the former is efficient for restoration time. Network resilience approaches are assigned to different classes according to various criteria such as network cost planning, design complexity, bandwidth trading, traffic recovery time, robustness, quality of protection, scalability, type and number of failures, et cetera.
In this book we offer a collection of the latest contributions to the area of survivability in optical networks. These have been written by a number of well-established researchers in optical networks. There is also a special section which deals with the latest issues in survivability in each chapter. The book contains several chapters and is preceded by a preview and state-of-the-art introductory chapter. Each of the chapters focuses on some theoretical and practical aspects of network survivability methodologies applied to real world problems.Chapter 1
introduces the principles and state-of-the-art of survivability provisioning in optical networks, and in particular, in optical networks that employ wavelength division multiplexing (WDM). Concepts of survivability provisioning in optical networks such as protection and restoration, dedicated versus shared survivability, path-based, link-based, segment-based, cycle-based survivability, and so on, are covered to provide multiple classes of quality of protection against single failure, dual-failure, multiple simultaneous failures, or shared risk link group failures, in WDM mesh networks. Recent developments in survivable service provisioning are summarized, such as survivability provisioning that takes into account the connection holding-time, survivability in WDM light-trail networks, and optical burst switched networks. Finally, the chapter briefly examines future research directions.
Survivable routing serves as one of the most important issues in optical backbone design. Due to the high data rates enabled by the wavelength division multiplexing technology, any interruption in the service results the loss of a large amount of application data. Thus, making efforts to calculate and signal the protection resources promptly after the failure occurred would lead to an unacceptable high delay. As the main purpose of Chapter 2
, the principles of pre-planned protection approaches in mesh optical backbone networks are discussed. The Shared Risk Link Group (SRLG) concept is introduced for modeling physical and geographical dependency among seemingly unrelated link failures. Finally, methods are presented for calculating the exact end-to-end availability of a connection.
Passive Optical Networks (PON) support access network subscribers with bandwidth requirements more than 10 Mbps. Fiber and node failures in a PON network can lead to large amounts of data loss, while isolating the central office from the subscribers. Hence, high network availability is desired when a PON is used for business enterprises and for providing mobile backhaul services. To maximize network availability, several protection architectures have been proposed in literature. In Chapter 3
, the novel WDM PON protection architectures amongst those proposed in the literature are critically analyzed and compared. The comparison is performed from topology, resource utilization, and power budget perspectives. The protection mechanisms that are typically used in the architectures and their impact on restoration are also discussed.
The advantages of transparent optical networks, such as high capacity and low cost, can be outweighed by their complex fault management and the high impact of the faults occurring within them. Indeed, transparent optical networks reduce unnecessary, complex, and expensive opto-electronic conversion, at the cost of having more damaging faults that affect longer distances than in opaque networks. Moreover, transparent optical networks have limited monitoring capabilities which could hinder efficient and accurate fault detection and localization. Different approaches have been proposed in the literature to perform fault localization, targeting different fault scenarios (e.g. single/multiple faults or looking at the optical/higher layers), and considering different assumptions (e.g. ideal/existence of false or lost alarms). Furthermore, fault management depends on the placement of monitoring equipment, whose optimization is studied and also presented in Chapter 4
In traditional optical networks, configured as static physical pipes, the carrier-grade network resilience is provided by means of protection and restoration capabilities. However, there is a need to develop a new generation of dynamic reconfigurable all optical networks with built in network resilience capabilities. In the next generation high-speed photonic packet switching networks, ultrafast packet header processing and packet switching are the vital building blocks. In Chapter 5
, a review of different routing schemes for high-speed photonic packet switching networks, as well as the concept of reducing the size of the look-up routing table, is presented. A novel PPM signal format is presented in order to reduce the size of the routing table so as to reduce packet switching and processing time compared to conventional routing tables. A failure self detection and a routing table reconfiguration in the optical domain is introduced, and a number of factors such as system performance, reliability, and complexity are also discussed. Chapter 6
outlines the different survivability approaches for mesh networks under static and dynamic traffic environments. It describes the different solution options and their implementations. Also included are detailed performance analyses and evaluations for the difference survivability approaches under both traffic environments. Finally, a performance comparison between the different survivability approaches is presented, and the chapter ends with some concluding remarks.
Achieving low blocking probability and connection restorability in the presence of a link failure is a major goal of network designers. Typically, fault tolerant schemes try to maintain low blocking probability by maximizing the amount of primary capacity in the network. In Chapter 7
, maximizing primary capacities in survivable networks is proposed. It is assumed that the total capacity on each link is fixed, and then it is allocated into primary or backup capacity. The distribution of primary capacity affects blocking probability for dynamic traffic. This is seen by simulating dynamic traffic with different ways to distribute capacities in a network. A Hamiltonian p-cycle is a capacity optimal way of allocating primary and backup capacity. However, different Hamiltonian p-cycles may deliver different blocking probabilities for dynamic traffic. In general, more evenly distributing the backup and primary capacity lowers the blocking probability. This chapter provides upper bounds on how much primary capacity a network can provide if it uses a link based protection strategy to guarantee survivability for one or more link failures. Using integer linear programs, it is shown that requiring pre-configuring carries a cost in terms of capacity if the solution is structured as a set of cycles.
Traffic grooming supports efficient utilization of network resources by allowing sub-wavelength granularity connections to be groomed onto a single lightpath. Chapter 8
investigates the problem of dynamic traffic grooming for WDM networks under a differentiated resilience scheme. Two differentiated resilience schemes at different grooming levels – the Differentiated Resilience at Lightpath (DRAL) level scheme and the Differentiated Resilience at Connection (DRAC) level scheme - are presented. These two explore different ways of provisioning backup paths and tradeoff between bandwidth efficiency and the number of required grooming ports. Both schemes support three resilience classes: dedicated protection, shared protection, and restoration. Simulation is carried out to evaluate and compare the two differentiated resilience schemes. Simulation results show that the DRAL scheme is relatively insensitive to the changes in the number of grooming ports, while the DRAC scheme utilizes grooming ports more aggressively as it trades grooming ports for bandwidth efficiency in routing and grooming.
The high capacity advantage of optical networks also introduces the risk of substantial data loss in case of a service interruption even if the outage lasts only a short time. Therefore, survivable and reliable design and management of optical networks is urgent. However, deployment of efficient survivability policies does not always guarantee the continuity of the service. Long failure restoration delays, multiple failures, and lack of protection resources may lead to service unavailability. Hence, connection availability arises as a design constraint, and it is defined as the probability of a connection being in the operating state at any time. Availability-constrained optical network design and availability-constrained connection provisioning are two important problems to guarantee robustness of connections in a survivable network which are discussed in Chapter 9
Originally, networks were engineered to provide only one type of service, i.e. either voice or data, so only one level of resiliency was requested. This trend has changed, and today’s approach in service provisioning is quite different. A Service Level Agreement (SLA) stipulated between users and service providers (or network operators) regulates a series of specific requirements, e.g., connection set-up times and connection availability that has to be met in order to avoid monetary fines. In recent years this has caused a paradigm shift on how to provision these services. From a “one-solution-fits-all” scenario, we witness now a more diversified set of approaches where trade-offs among different network parameters (level of protection vs. cost and/or level of protection vs. blocking probability) play an important role. Chapter 10
aims at presenting a series of network resilient methods that are specifically tailored for a dynamic provisioning with such differentiated requirements. Both optical backbone and access networks are considered. In the chapter, a number of provisioning scenarios - each one focusing on a specific Quality of Service (QoS) parameter - are considered. First, the effect of delay tolerance, defined as the amount of time a connection request can wait before being set up, on blocking probability is investigated when Shared Path Protection (SPP) is required. Then, the problem of how to assign “just-enough” resources to meet each connection availability requirement is described, and a possible solution via a SPP scheme with Differentiated Reliability is presented. Finally, a possible trade off between deployment cost and level of reliability performance in Passive Optical Networks (PONs) is investigated. The results presented highlight the importance of carefully considering each connection’s QoS parameters whilst devising a resilient provisioning strategy. By doing so, the benefits in terms of cost saving and blocking probability improvement become relevant, allowing network operators and service providers to maintain satisfied customers at reasonable capital and operational expenditure levels.Chapter 11
provides new distributed frameworks to support Quality of Service (QoS) differentiation. These frameworks provide differentiated protection services to meet the availability requirements of customers in an effective manner. The availability-analysis for connections with different protection schemes is described. Through this analysis, it is shown how connection availability is affected by resource sharing. Based on the availability analysis, the proposed framework provisions each connection, in which an appropriate level of protection is provided according to its predefined availability requirement. Networks without wavelength conversion capability are considered as well as dynamic traffic environment. In these distributed frameworks several distributed schemes to provision and manage connections cost-effectively while satisfying the existing and new connections availability requirements are proposed.Chapter 12
presents a novel approach for dealing with failures in and attacks on Transparent Optical Packet Switching (TOPS) mesh networks. The approach is composed of two phases, wherein the first one dynamically dimensions the resources in the network and the second one applies an incremental learning algorithm, which generates an intelligent policy. At each node, such a policy allows a self-healing behavior when there are failures or attacks in the network. Finally, the performance of this approach is presented as well as future research lines.
The idea of Chapter 13
is to give an overview of fiber-optic communication systems. The most important devices for fiber-optic transmission systems are presented and their properties discussed. In particular, there is consideration of such systems working with those basic components necessary to explain the principle of operation. Among them is the optical transmitter, consisting of a light source, typically a high speed driven laser diode. Furthermore, the optical receiver has to be mentioned; it consists of a photodiode and a low noise high bit rate front-end amplifier. Nevertheless, in the focus of the considerations the optical fiber is found as the dominant element in optical communication systems. Different fiber types are presented and their properties explained. The joint action of these three basic components leads to a fiber- optic systems, mainly applied for data communication. The systems operate as transmission links with bit rates up to 40 Gbit/s. Furthermore, optical communication systems have also been used for recent application areas in the MBit/s region, e.g. in aviation, automobile, and maritime industry. Therefore, besides pure glass fibers, polymer optical fibers (POF) and polymer-cladded silica (PCS) fibers have to be taken into account.
Thus, this collection provides a wide ranging overview of many of the salient issues in modern optical networks from the point of view of resilience. It is intended to appeal to a wide range of readers, and for this reason, it includes both material of a tutorial nature and reports on state of the art methods. It is essential that the key differences and similarities between optical and other networks are appreciated, particularly with regard to all optical networks. Moreover, given the importance of WDM in the provision of multiple channels, it receives substantial coverage. Furthermore, optical networks make it necessary to take a broad view of resilience, and this collection has done so, including several layers of the OSI model. For readers new to many aspects of the topic, there is coverage of many of the essentials fundamentals, particularly in the first few chapters. Needless to say, these can be skipped over quickly by those more familiar with the subject area.
When considering transparent optical networks with their advantages such as high capacity, it was also part of the aim to bring to the attention of the reader their complex fault management and the high impact of the faults occurring within them. Thus, the collection here considers cutting-edge aspects of this problem such as incremental learning and the optimum placement of monitoring equipment.
Resilience is not just of significance within the core network, and in this collection, high speed access networks are not forgotten, with a particular emphasis on Passive Optical Networks (PON) supporting relatively demanding end users. Given the tree structure implicit in PONs, failures can lead to large data losses and subscriber isolation, so protection is of paramount importance. Moreover, the differentiated services offered throughout modern optical networks (including the access part) mean that differentiated resilience must also be offered. For the future, it will be necessary to provide an appropriate level of protection across the same network for users from impoverished students tackling homework assignments to major corporations handling voluminous and confidential data processing tasks. Whereas in the past, networks were often designed for just one type of service, modern service level provision is via a totally different model. In the networks of today and tomorrow, each user may have a different level of agreed service, leading to inevitable tradeoffs that may include price as well as purely technical factors. In this respect, distributed frameworks are germane to the delivery of the differentiated QoS required. Moreover, traffic grooming offers a route to the efficient of network resource utilization, and thus, features significantly in the book in the context of differentiated resilience.
Achieving low blocking probability and connection restorability when failures occur also figure prominently in the work. Within these topics, methods for bounding the capacities available and for offering rapid restoration are covered. In addition, the traffic changes with time in modern high speed networks, so the increasing prevalence of dynamic traffic environments figures within the collection
Research and development engineers, graduate students studying optical networks, and senior undergraduate students with a background in algorithms and networking will find this book interesting and useful. This work may also be used as supplemental readings for graduate courses on internetworking, routing, survivability, and network planning algorithms.Yousef S. Kavian
Shahid Chamran University, IranMark S. Leeson
University of Warwick, UK