Exploring the Impact of Topological Features on Ordering Routing and Spectrum Assignment in Elastic Optical Networks

We explore the influence of topological features on a hybrid routing and spectrum assignment (HRSA) algorithm in elastic optical networks that prioritizes either routing or spectrum assignment for network source-destination node pairs. We use logistic regression to extract information about how the HRSA makes its decision, whether it is about prioritizing proper routing or spectrum assignment choices based on efficient solutions returned by the HRSA. Results show that the proposed model achieved 81% of accuracy and provided insights on how topological features impact the decision-making criteria of HRSA.

I. Introduction

In optical networks, a connection request requires a route and a spectrum slice available in order to be established. Wavelength division multiplexed optical networks (WDM) use a frequency grid with a fixed size for any transmission. In turn, elastic optical networks (EONs) use a flexible frequency grid to adapt and serve connections with different transmission rates, enabling more efficient use compared to WDM networks [1]. In both cases, the problem of routing and allocating the portion of the spectrum is complex and essential to network performance, mainly under dynamic traffic.