Performance Analysis Of Multi-Hop Optical Networks Based On Wavelength Division Multiplexing

This dissertation comprises a major and minor study on the investigation of the combined effects of noises and crosstalk existed at each node to the multi-hop optical network. Computed results show that the packet error rate (PER) and the bit error rate (BER) performance are limited by a signal spon...

Full description

Saved in:
Bibliographic Details
Main Author: Chien, Su Fong
Format: Thesis
Published: 2001
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This dissertation comprises a major and minor study on the investigation of the combined effects of noises and crosstalk existed at each node to the multi-hop optical network. Computed results show that the packet error rate (PER) and the bit error rate (BER) performance are limited by a signal spontaneous beat noise and crosstalk. The convolutional coding is proposed to be implemented in the multi-hop optical networks in order to improve the BER/PER performances. Results with convolutional coding show that there is a considerable improvement in network performance due to an increase in the number of traversable hops from a given transmitter power at a given BER/PER. This improvement is observed in the multi-hop optical network with and without wavelength conversion. A new hybrid model is later proposed to the multi-hop optical network by adding a fiber delay-line optical buffer to the node equipped with wavelength converters. This model has two major advantages, that is it will reduce the deflection probability as well as the required number of wavelength converters in each node in full wavelength convertibility multi-hop optical network. The effect of wavelength converters on the network BER/PER perfprmance is studied. Results show that the wavelength conversion efficiency plays a major role in optimizing the BER/PER performance. As a related study to multi-hop optical network , two existed topologies( Extra-stage ShuffleNets and Manhattan Street Networks) were compared fairly in terms of number of nodes. An algorithm to find the same number of nodes for both network architectures has been developed.