Elastic optical network (EON) efficiently utilize spectral resources for optical fiber communication by allocating the minimum necessary bandwidth to client demands. On the other hand, network traffic has been continuously increasing due to the wide penetration of video streaming services, so the efficient and cost-effective use of available bandwidth plays an important role in improving service provisioning. In this work, we formulate and solve an optimization problem to perform routing and spectrum assignment (RSA) in EON with focus on video streaming. In this formulation, EON and video constraints such as spectrum fragmentation and received video quality are considered jointly. In this way, we utilize a machine learning (ML) technique to estimate the video quality versus channel state. The proposed algorithm is evaluated over two benchmarks fiber-optic network, namely NSFNET and US-backbone using numerical simulations based on random traffic models. The results reveal that the mean optical signal-to-noise ratio (OSNR) for video content data in the receiver is remarkably higher than in non-video data. This is while the blocking ratio is the same for both data types.
Download