Efficient Bandwidth Allocation Techniques in Survivable WDM-PON Architectures

The aim of this project is to properly analyze and characterize the limitations of current Ethernet Passive Optical Network (EPON) architectures and propose survivable and efficient architectures for Wavelength Division Multiplexing (WDM)-PON systems. This project proposes to develop efficient bandwidth allocation techniques as well as effective survivability schemes. It is also the aim of the proposed project to devel op a Discrete Event Simulator using C++ to simulate the proposed WDM-PON architectures and analyze the performance of the proposed algorithms and approaches. The Discrete Event Simulator will be essential to conduct research in WDM-PON and it will facilitate the start of an undergraduate research group to support a full-scale research investigation to solve the access bottleneck/gap problem.

Efficient Localized Energy Routing in Mobile Ad Hoc Networks

I am working on the problem of localized energy aware routing in mobile ad hoc networks. Specifically, this research project focuses on localized routing algorithms, where each node forwards a message based on the position of itself, its neighbors and the destination. The objective of energy aware routing algorithms is to minimize the total power for routing a message from source to destination or to maximize the total number of routing tasks that a node can perform before its battery power depletes. I will propose new localized energy aware routing algorithms to provide high packet delivery rates with low packet forwarding and low battery power consumption. In addition, the algorithms will ensure good energy distribution among the nodes. The new algorithms will be an effort to optimize more than one metrics at the same time (i.e., minimizing the energy consumption to deliver a packet, maximizing packet delivery rates, ensure good energy distribution), which it has not been addressed in the literature before. The performance of our algorithms will be compared with other existing energy and non-energy aware localized algorithms through computer simulations.

Dynamic Provisioning of Survivable Heterogeneous Traffic in WDM Networks

The problem of protecting purely multicast connections in WDM mesh network has recently started to receive some attention in the literature. In fact, WDM traffic is heterogeneous in nature and only part of the traffic is multicast and the rest is unicast, therefore we expect the presence of both unicast and multicast traffic in future optical networks. In this research project I studied the problem of dynamic provisioning of survivable heterogeneous unicast and multicast traffic in WDM networks. Specifically, I proposed new protection schemes to provision and protect unicast and multicast connection requests against single-link failures in WDM-mesh networks. The simulation results of the proposed protection schemes are compared with each others and with those found in the literature. The results showed that our proposed scheme TP-OSPT outperforms all other schemes for moderate and large multicast group size. On the other hand, my proposed scheme OCR performs the best for small multicast group size.

ACCOMPLISHED RESEARCH WORK

Real-Time Provisioning

My initial work focused on investigating the RWA problem that is inherent in real-time provisioning in WDM Networks. The investigation included both static and dynamic routing and assignment techniques and was also extended to multi-fiber environments. Since the problem is NP-hard, heuristic algorithms were developed and it was shown that considering the network status (in a dynamic and/or adaptive way) always improves the admission rate making the network more cost-effective.

Integrating Internet Protocol (IP) and WDM architectures (IP-over-WDM interconnection model)

One of the critical issues facing the implementation of this model is how to provide the desired features of rapid provisioning and restoration between the optical layer and the client. The work involved proposing a novel, simple and scalable optical networking paradigm where most of the networking functionalities and intelligence has been shifted to the optical layer including supporting selective provisioning/restoration of diverse traffic granularity entirely on the optical layer’s terms. Rather than utilizing the conventional IP/MPLS-based optical control plane that only manages connections among optical switches within an optical “cloud”, a GMPLS-based unified control plane was devised that manages connections across many layers.  Utilizing the proposed model, an integrated dynamic routing approach that take into account the combined topology and resource usage information at both the IP and optical layers was developed to traffic-engineer the connection requests of the IP over WDM-based networks.

Dynamic Multicast Traffic Grooming in WDM-based Networks

This problem is equivalent to designing a light-tree based logical topology for multicast streams. Most studies of traffic grooming instances have assumed only unicast traffic, although it is expected that a part of the future traffic of high performance optical networks will be multicast traffic. Therefore, efficient use of optical channel bandwidth is required when routing these multicast demands and many such sub-rate connections should be appropriately groomed together onto one optical channel. This problem consists of four parts:  (1) Routing (2) Wavelength assignment, (3) Design of a light-tree based logical topology, and (4) Traffic-grooming. Different routing schemes were developed to efficiently groom low-speed connections on the light-tree based logical topology and efficiently utilize the network resources. Numerical results prevailed that the proposed approaches use the network resources more efficiently compared to the non-grooming approach and the approach of serving the multicast requests as separate unicast requests.

 Analyzing connection blocking probabilities in Optical Networks

Given a set of connections, the problem of setting up lightpaths by routing and assigning wavelength to each connection is called the routing and wavelength assignment problem (RWA). A new analytical model was proposed to solve the problem of RWA and compute approximate blocking probabilities using the layered-graph approach. Specifically, the Equivalent Random Method (ERM) was used to model the overflow traffic between different wavelength layers and between alternate routes. The blocking performance of fixed routing (FR), fixed-alternate routing (FAR) and fixed alternate routing with trunk reservation (FARwTR) was evaluated using this model. The performance of the analytical model was compared with those of the conventional models that use heuristic algorithms.