UPM Institutional Repository

Resource management algorithms for wavelength division multiplexing optical metro slotted ring networks


Citation

Mohammed Abdulrazak, Fuaad Hasan (2012) Resource management algorithms for wavelength division multiplexing optical metro slotted ring networks. PhD thesis, Universiti Putra Malaysia.

Abstract

The performance of the optical packet ring is highly dependent on the method of resource sharing deployed. In addition, the increase in demand of bandwidth by web applications relying on multicasting has made optical fibers a preferred solution. Therefore, fair sharing with collision avoidance have to be ensured when MAC protocols are designed. The most promising technology for the optical network development is the optical packet switched (OPS) WDM networks designed for the Metropolitan Area Networks (MANs). In addition, Hybrid Optoelectronic Ring NETwork (HORNET) and High-performance Optical-Packet-Switched Metro Area Network (HOPSMAN) ring architectures, have been studied as a next generation platform for MANs in this research. These architectures employ a separate wavelength as the control information channel. This control information enables the nodes to monitor their access to the transmission media. Each node is equipped with a tuneable transmitter and fixed/tuneable receiver for the purpose of data communication. Access nodes are equipped with a fixed transmitter and fixed receiver for the control information exchange. In this research, the management of the shared network resources has been researched extensively with the intention to enhance the performance of the network and overcome the MAC protocols related limitations. Among the pertinent issues addressed are the network under utilized resources, the best-than-average access for the upstream nodes in the network, the existence of misbehaving nodes and the partiality towards homogeneous traffic. Buffers and network bandwidth are the main resources analysed, where their consumptions have been enhanced in the aim of ensuring fairness and QoS support. This research presents novel algorithms to enhance the performance of the Wavelength Division Multiplexing (WDM) dual ring network. The proposed Enhanced Virtual Output Queues Management (EVOQM) algorithm provides an elevated fairness for unicast and multicast data in the network. In addition, it is able to provide provision which are extended for QoS support. The Dynamic Based Bandwidth Reservation (DBBR) algorithm is designed to ensure that fair bandwidth access for both upstream and downstream nodes are achieved. The proposed DBBR aims also to prevent misbehaving (i.e. malicious) nodes from affecting the network performance. The third proposed algorithm which is the Dynamic Scaling Bandwidth Reservation (DSBR) has been proposed to enhance the bandwidth utilization in HOPSMAN slotted WDM ring architecture. HOPSMAN considers a minimum of at least one node as a server node in unidirectional ring HORNET architecture with tuneable receiver at each node in the network. This proposed algorithm optimizes a collaborative resource management strategy with traffic heterogeneity to enhance the utilization of network resources and ensures the requested QoS pre-requisites are guaranteed. The realization of these strategies have been achieved and displayed by the reduction of packet loss (i.e. maximizing the total value of transferred packets) in the network. Thus, leading to the improvement of the throughput while adhering to the delay pre-requisites. The performance evaluation of the proposed algorithms has been done through a detailed and extensive discrete-event-simulation analysis. Metrics of network throughput, packet loss and delay measures, have been the main basis for the acquired results.


Download File

[img]
Preview
PDF
FSKTM 2012 14R.pdf

Download (695kB) | Preview

Additional Metadata

Item Type: Thesis (PhD)
Subject: Wavelength division multiplexing
Subject: Ring networks (Computer networks)
Subject: Computer algorithms
Call Number: FSKTM 2012 14
Chairman Supervisor: Assoc. Prof. Shamala Subramaniam, PhD
Divisions: Faculty of Computer Science and Information Technology
Depositing User: Haridan Mohd Jais
Date Deposited: 16 Jan 2015 04:03
Last Modified: 02 Feb 2015 07:06
URI: http://psasir.upm.edu.my/id/eprint/32304
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item