Abstract

Ethernet technologies have evolved through enormous standardization efforts over the past two decades to achieve carrier-grade functionalities, leading to carrier Ethernet. Carrier Ethernet is expected to dominate next generation backbone networks due to its low-cost and simplicity. Ethernet's ability to provide carrier-grade Layer-2 protection switching with SONET/SDH-like fast restoration time is achieved by a new protection switching protocol, Ethernet Ring Protection (ERP). In this thesis, we address two important design aspects of carrier Ethernet networks, namely, survivable design of ERP-based Ethernet transport networks together with energy efficient network design. For the former, we address the problem of optimal resource allocation while designing logical ERP for deployment and model the combinatorially complex problem of joint Ring Protection Link (RPL) placements and ring hierarchies selection as an optimization problem. We develop several Mixed Integer Linear Programming (MILP) model to solve the problem optimally considering both single link failure and concurrent dual link failure scenarios. We also present a traffic engineering based ERP design approach and develop corresponding MILP design models for configuring either single or multiple logical ERP instances over one underlying physical ring. For the latter, we propose two novel architectures of energy efficient Ethernet switches using passive optical correlators for optical bypassing as well as using energy efficient Ethernet (EEE) ports for traffic aggregation and forwarding. We develop an optimal frame scheduling model for EEE ports to ensure minimal energy consumption by using packet coalescing and efficient scheduling.