Fragmentation-Aware Algorithm with Bordering Super-Channels in Spectrally/Spatially-Flexible Optical Networks

Spectrally/spatially flexible optical networks (SS-FONs) are a promising solution to future traffic requirements in optical backbone networks. SS-FONs exploit the spatial dimension to increase the network capacity, while preserving the resource management flexibility, as they still operate within a flex-grid composed of small frequency slots (slices). A flex-grid makes it possible to realize transmission using super-channels (SChs) that comprise a set of contiguous slots. In this paper, we focus on spectral SChs, i.e., a SCh realized only on one spatial mode on each fiber. In SS-FONs, setting up and tearing down multiple lightpath requests within a flex-grid may result in spectrum fragmentation and, in turn, blocking of requests. In this work, we investigate several fragmentation metrics in a SS-FON. The problem of identifying appropriate metrics to measure fragmentation has been investigated in single-core elastic optical networks, but to the best of our knowledge, there is no such research available for SS-FONs. Therefore, we propose several fragmentation metrics for SS-FONs. We introduce the concept of bordering super-channels (B-SChs), i.e., SChs whose spectrum is allocated at the border of already allocated spectrum slots, which show promising results with regard to minimizing fragmentation. The investigation of all candidate B-SChs allows us to find the one that minimizes the network fragmentation and, in turn, spectrum waste. Hence, we propose a fragmentation-Aware (FA) algorithm with bordering SChs that assigns optical resources to dynamic requests by utilizing information from the proposed fragmentation metrics and the set of candidate B-SChs. Experiments on a representative network topology show that the investigation of multiple B-SChs in the FA algorithm reduces the blocking probability when compared to the reference FA algorithms. Finally, we analyze the impact of a spatial continuity constraint on the network fragmentation.