Improving the Capacity of Terrestrial and Submarine Systems via Channel Power Optimization

We review techniques for channel power optimization in terrestrial and submarine systems. In terrestrial systems, including mesh networks and point-to-point links, we can formulate the maximization of capacity as a convex optimization problem. We model Kerr nonlinearities using a discretized version of the Gaussian noise model. In ultra-long submarine links, capacity is limited also by the electrical power available to the undersea amplifiers. Recent works have focused on maximizing capacity by optimizing the number of spatial dimensions. We show, however, that by also optimizing channel power and data allocation over wavelength to take account of amplifier physics, the capacity per fiber may be further doubled with respect to recently published results.