Spectrum-efficient and agile CO-OFDM optical transport networks: architecture, design, and operation

Traditional ITU-T fixed frequency grid-based optical transport networks suffer several drawbacks such as low fiber spectral efficiency, difficulty in supporting large-bandwidth super-channels, and inflexibility in network bandwidth reconfiguration and modification. To overcome these drawbacks, a new-generation optical transport network based on the concepts of agile spectrum operation and elastic bandwidth allocation has been recently proposed and is receiving increasing attention. This new-generation network is called coherent optical orthogonal frequency-division multiplexing (CO-OFDM) optical transport network. It employs the promising CO-OFDM transmission technique and the new-generation bandwidth-variable ROADMs that use the coherent detection capability of the CO-OFDM transmission for optical channel filtering. The CO-OFDM optical transport network is characterized by arbitrarily assigning center frequency and bandwidth of an optical channel, thereby providing flexibility in network design and operation and achieving efficient fiber spectrum utilization. Despite the increasing attention and considerable progress, there are still many outstanding issues regarding the implementation of CO-OFDM optical transport networks. This article reviews the literature on the architectures of the CO-OFDM optical transport network and discusses key issues, particularly involving network control plane, lightpath routing and spectrum assignment, impact of channel modulation format and optical reach, subwavelength traffic grooming, network survivability, and network reconfiguration.

Introduction

With the increased demand and related business opportunities of bandwidth-intensive Internet applications, especially video-based services, it is important to find cost-effective means to keep on increasing the capacity of the optical transport network. In general, optical network capacity can be increased by laying more fibers, or improving capacity utilization (or spectral efficiency) of each fiber. Comparing these two alternatives, the first is normally more expensive because it is labor-intensive and often incurs further cost associated with interruption of people's daily lives and transportation in metropolitan areas. Thus, the second alternative (i.e., increasing fiber spectral efficiency) is more attractive. Fiber spectral efficiency is referred to as the bitrate that can be transmitted over a given optical bandwidth (spectrum) in a fiber-optic transmission system. In theory, a standard single-mode fiber can carry up to . However, today's most advanced industrial products achieve at most around 10-Tb/s capacity (e.g., 100 Gb/s per wavelength and 100 wavelengths per fiber). This gives rise to the opportunity of exploiting additional 20–40-Tb/s capacity for the deployed fibers without laying new fibers.