An Adaptive Spectrum Time Assignment in OFDM Datacenter Optical Networks

Data center optical networks need to periodically reconfigure the networks to deal with traffic changes. Prior to network reconfiguration, tenant requests are known. Adjusting tenant spectrum demand not only predicts application performance, but also improves network throughput. This paper investigates an adaptive spectrum time assignment (ASTA) problem in orthogonal frequency division multiplexing data center optical networks, which is modeled by Integer Linear Programming (ILP). However, in large-scale networks, the ILP model cannot provide an optimal solution within an acceptable time. To address this problem, the most data volume first with an adaptive 90 rotation MDVFA90 R algorithm is proposed in this paper, which is a heuristic algorithm with an adaptive 90 rotation and can find a sub-optimal solution within a shorter time. Simulation results verify that the ILP model can provide an optimal solution for small-scale networks, while the MDVFA $90 R algorithm can provide a better sub-optimal solution for large-scale networks.