| In this thesis, we focus on two important research problems in next generation optical networks with wavelength-division multiplexing (WDM) circuit switching (flow switching) technologies: (1) scalability of network management and control, and (2) resilience/reliability of networks upon faults and attacks. Our main technical approaches are decision theory and probabilistic graphical models. As optical networks grow in size and complexity, there is a need for inter-domain light-path assessment using partial management information.; Therefore, to understand scalable network management of flow switching, we investigate the scalability of network management information for inter-domain light-path assessment. Using the framework of decision theory and probabilistic graphical models, we formulate the lightpath assessment as a decision problem. We show that partial information available can indeed provide the desired performance, i.e., a small percentage of erroneous decisions can be traded off to achieve a large saving in the amount of management information. A second consequence of the large network size and great network complexity is that: networks face an increasingly adverse environment with more frequent faults and malicious attacks.; To understand network resilience under malicious attacks, we study the resilience of all-optical networks under in-band crosstalk attacks using probabilistic graphical models. Graphical models provide an explicit view of the spatial dependencies and interactions between the physical layer and the network layer, as well as computationally efficient approaches, e.g., sum-product algorithm, for studying network resilience. Based on the proposed cross-layer model of attack propagation, we investigate key factors that affect the resilience of the network under in-band cross-talk attacks from both the physical layer and the network layer. In addition, we obtain analytical results on network resilience for typical topologies including ring, star, and mesh-torus networks.; To understand network performance upon failures, we systematically investigate traffic-based network reliability. We first adopt a uniform deterministic traffic at the network layer. This allows us to focus on the impacts of network topology, failure dependency, and failure protection on network reliability, and to obtain analytical results on the network reliabilities of typical network topologies. We then apply a random network layer traffic model with Poisson arrivals to further investigate the effect of network layer traffic distributions on network reliability. We study the interaction between the network reliability and the connection arrival rate, and obtain asymptotic results of network reliability metrics with respect to arrival rate for typical network topologies under heavy load regime.; The main contributions of this thesis include: (1) fundamental understandings of scalable management and resilience of next-generation optical networks with WDM flow switching; and (2) the innovative application of probabilistic graphical models, an emerging approach in machine learning, to the research of communication networks. |
