Spectrum engineering in flexible grid data center optical networks

Data centers provide a volume of computation and storage resources for cloud-based services, and generate very huge traffic in data center networks. Usually, data centers are connected by ultra-long-haul WDM optical transport networks due to its advantages, such as high bandwidth, low latency, and low energy consumption. However, since the rigid bandwidth and coarse granularity, it shows inefficient spectrum utilization and inflexible accommodation of various types of traffic. Based on OFDM, a novel architecture named flexible grid optical network has been proposed, and becomes a promising technology in data center interconnections. In flexible grid optical networks, the assignment and management of spectrum resources are more flexible, and agile spectrum control and management strategies are needed. In this paper, we introduce the concept of Spectrum Engineering, which could be used to maximize spectral efficiency in flexible grid optical networks. Spectrum Defragmentation, as one of the most important aspect in Spectrum Engineering, is demonstrated by OpenFlow in flexible grid optical networks. Experimental results are reported and verify the feasibility of Spectrum Engineering.