Degree Type

Dissertation

Date of Award

2002

Degree Name

Doctor of Philosophy

Department

Electrical and Computer Engineering

First Advisor

Arun K. Somani

Abstract

Wavelength division multiplexed (WDM) optical networks have emerged as the viable solution to providing high-speed networking solutions that meet the increasing bandwidth demands of the Internet. WDM networks divide the fiber bandwidth into multiple WDM channels called wavelengths. The current transmission speed on a wavelength is 40 Gigabits per second. The user requirement are still of sub-wavelength capacity. Such a mis-match between the user requirements and the minimum granularity of bandwidth offered by the network has necessitated the use of efficient wavelength sharing mechanisms or traffic grooming.;In this dissertation, we propose a novel conceptual network model called Trunk Switched Network (TSN) to model WDM grooming networks. The links in a network are viewed as a set of channels. The channels are combined at a node to form groups called trunks. Nodes can switch channels across different links that fall within the same trunk. We model WDM grooming networks with different grooming capabilities as TSNs. We develop a framework for connection establishment, called MICRON (Methodology for Information Collection and Routing in Optical Networks). We develop an analytical model based on fixed-path routing mechanism includes three main features: (1) incorporates heterogeneous switching architectures; (2) considers multirate traffic streams; and (3) employed to evaluate blocking performance of multicast tree establishment. The analytical models are shown to closely approximate simulation values. Through extensive simulations, we demonstrate that routing requests based on the request characteristics and splitting higher capacity requests into multiple requests of lower capacity can be employed as a cost-effective alternative to improving network performance as compared to the high-cost solution of increasing grooming capability at the nodes in the network. We believe that the proposed network model and the results presented in this dissertation will have a significant impact in the design and operation of future high-speed backbone networks.

DOI

https://doi.org/10.31274/rtd-180813-9834

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu

Copyright Owner

Srinivasan Ramasubramanian

Language

en

Proquest ID

AAI3143531

File Format

application/pdf

File Size

141 pages

Share

COinS