Performance Evaluation of Survivability Approaches in Optical Networks

Performance Evaluation of Survivability Approaches in Optical Networks

Abdelhamid Eshoul (University of Ottawa, Canada) and Hussein T. Mouftah (University of Ottawa, Canada)
DOI: 10.4018/978-1-61350-426-0.ch006
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The chapter 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, we present a performance comparison between the different survivability approaches and end the chapter with some concluding remarks.
Chapter Preview
Top

Introduction

Network survivability requires reserving enough spare capacity during the connection setup and utilizing the reserved spare capacity upon the occurrence of a network failure. The objectives of any survivability scheme are to allocate network resources efficiently and minimize restoration time during a network failure. Achieving both objectives simultaneously, especially under wavelength continuity constraint, poses a major challenge in survivable WDM networks.

Survivability approaches are classified based on the protection ranges as link, segment and path based protection (Eshoul and Mouftah, 2009). In link-based protection, the traffic is rerouted around the end nodes of the failed link; whereas, in path-based protection, a backup path is pre-determined between the source and the destination nodes. On the other hand, Segment-based protection is a trade-off between the link-based and path-based protection schemes. Protection schemes are also classified based on the possibility of resource sharing as dedicated and shared protection. Dedicated protection schemes have fast restoration times at the expense of higher resource redundancy. In contrast, shared protection schemes reduce resource redundancy significantly at the expense of increased restoration time. There are two different implementations to shared protection in mesh networks: the diverse routing approach (Dongvun and Subramaniam, 2000) and the p-cycle approach (Grover and Stamatelakis, 1998).

Diverse Routing Approach

The diverse routing approach is a path-based protection scheme, where a working path (primary path) and a backup path are set up during the RWA process. The routes of the working and the backup paths must be link-disjoint in order to protect against link cut, or node-disjoint in order to protect against node and link failures. In the shared protection scheme, a number of backup paths can share the same resources as long as their working paths are not under the same risk of failure. The concept of Shared Risk Trunk Group (SRTG) is introduced to check when different working paths can share the same backup resource. The SRTG of a trunk t consists of the working resources that pass through it. All working paths that are part of STRG t fail when t fails. As a result, the working paths of any SRTG cannot share any backup resources. Stating the idea differently, any backup resource in a trunk t can only protect one working resource, of the same size or smaller, in any other trunk.

P-Cycle Approach

Unlike the diverse routing approach, the p-cycle approach is a link based approach where one or more pre-configured protection cycles, which may overlap with each other, are formed. The major advantages of p-cycle protection schemes over the diverse routing protection schemes are their ability to achieve both good resource efficiency and fast restoration times simultaneously. The p-cycle approach can achieve fast restoration time due to its switching over mechanism, during times of failure, which is similar to that of the link-based protection scheme, where switching over to the backup path involves only the end nodes of the failed link. Moreover, p-cycle protection schemes can reach good resource redundancy compatible to that of conventional protection schemes used in mesh networks.

Top

Applied Traffic

The traffic applied to wavelength-routed WDM networks is mainly confined to two types: static traffic and dynamic traffic. Under the static traffic environment, the objective is to set up a given set of demands while minimizing network resources. Depending on the complexity of the static traffic problem, Integer Linear Programming (ILP) techniques can effectively be used to optimally solve small to moderate size problems for both diverse routing and p-cycle approaches. Several heuristic algorithms have been proposed for larger problems to solve the survivable Routing and Wavelength Assignment (RWA) problem. In this chapter we mainly use ILP techniques to solve the static survivable RWA.

Complete Chapter List

Search this Book:
Reset