Introducing node architecture flexibility for elastic optical networks

A large number of factors generate uncertainty on traffic demands and requirements. In order to deal with uncertainty optical nodes and networks are equipped with flexibility. In this context, we define several types of flexibility and propose a method, based on entropy maximization, to quantitatively evaluate the flexibility provided by optical node components, subsystems, and architectures. Using this method we demonstrate the equivalence, in terms of switching flexibility, of finer spectrum granularity, and faster reconfiguration rate. We also show that switching flexibility is closely related to bandwidth granularity. The proposed method is used to derive formulae for the switching flexibility of key optical node components and the switching and architectural flexibility of four elastic optical node configurations. The elastic optical nodes presented provide various degrees of flexibility and functionality that are discussed in the paper, from flexible spectrum switching to adaptive architectures that support elastic switching of frequency, time, and spatial resources plus on-demand spectrum defragmentation. We further complement this analysis by experimentally demonstrating flexible time, spectrum, and space switching plus dynamic architecture reconfiguration. The implemented architectures support continuous and subwavelength heterogeneous signals with bitrates ranging from 190 Mb/s, for a subwavelength channel, to 555 Gb/s for a multicarrier superchannel. Results show good performance and the feasibility of implementing the architecture-on-demand concept.