A Survey of Survivability Techniques for Optical WDM Networks

Abstract

In an optical wavelength division multiplexed (WDM) network, link or node failures may result in huge amounts of lost data, due to the enormous fiber throughput. This requires that optical WDM network be designed to be resilient to failures. Thus, survivability can be defined as the ability to respond gracefully to such failures. This chapter presents a comprehensive survey of various mechanisms proposed to achieve survivability. The survey considers different topologies, different failure models, implementation issues, signaling issues and quality-of-protection issues. Keywords:Survivable Optical WDM Networks, Protection and Restoration, Single and dual-link failures, Node failures, Channel failures, Implementation, GMPLSbased Signaling, Quality of Protection. IntroductionOptical fiber based networks, characterized by a bandwidth of over 50 terabits per second (Tb/s), have emerged as the transmission medium of choice for high speed communication due to their capacity, reliability, cost and scalability. Recent advances in optical technology and in particular wavelength division multiplexing (WDM), a multiplexing technique that partitions the bandwidth provided by an optic fiber into individual multi-gigabit channels, have been identified as enabling technologies that will allows us to fully and effectively utilize the fiber capacity [1, 2]. Current optical technology demonstrations have shown a feasibility of 160 channels, each operating at 10 gigabits per second (Gb/s), and future networks are expected to operate at 40 Gb/s per channel or higher.

Dynamic virtual network restoration with optimal standby virtual router selection

“…Sivakumar et al [104] presents a survey on the survivability optical networks in two common topologies: the SONET/SDH ring topology and the WDM mesh topology.…”

“…For the SONET/SDH ring topology, there are three standard protection schemes [104]: (1) autonomic protection switching (APS) (2) dual homing (3) self-healing rings.…”

“…In the dedicated-backup paradigm, the common ways are 1+1 and 1:1 [104,118]. In the shared-backup paradigm, the backup wavelength could be shared by multiple paths that do not fail simultaneously, and the paths can be either the primary paths or the backup paths.…”

“…Usually a node failure may cause multiple failures in the links that are attached to this node, and to recover these links requires multiple node disjoint backup paths to reroute the traffic traversing these links [104]. Liu et al [71] proposed a spare capacity allocation model with successive survivable routing for node failures in a mesh topology, where each traffic flow was preplanned a node-level disjoint path so any single node failure along the path can be protected.…”

“…For the optical networks, centralized approaches have been widely used for the restoration [40,104], because a central controller has a global view of the whole optical network and can quickly calculate the restoration paths for the primary path, and then conduct the restoration procedure.…”