Machine learning‐oriented resource allocation in C + L + S bands extended SDM–EONs

The enormous growth in the number of smartphones, internet usage, and internet of things applications require high data‐rate transport platforms. Therefore, spectrally and spatially scalable core optical networks are essential. Moreover, the current generation, field existing networks, consume the conventional (C) band of the single‐mode single‐core fibres. The recent studies evidence that the increased traffic occupies the complete C band. To overcome this, the spectral and spatial expansion employing multicore (MC), multimode (MM), and multiband (MB, C + L + S) fibre‐based spatial division multiplexing (SDM)–elastic optical network (EON) is a promising solution. However, dynamic traffic, complex networking with MC–MM–MB–SDM–EON requires dynamic network solutions. To accomplish this, the authors propose the machine learning (ML) oriented route, core, mode, band, and modulation format aware spectrum assignment. Such resource allocation (RA) reduces the impact of the physical layer impairments, inter core cross‐talk, differential mode delay, and mode dependent losses. As per the author's best knowledge, ML‐oriented RA for SDM–EON considering the aforementioned parameters is the first investigation. They propose several RA algorithms to reduce the total network cost with increased fibre capacity. Among all, priority‐based RA comes out as an optimal solution.