Scaling SDM optical networks using full-spectrum spatial switching

In today's optical networks, the capacity supported by a single optical fiber is far higher than the demand between source-destination pairs. Thus, to take advantage of the installed capacity, lightpaths allocated between distinct source-destination pairs share the capacity provided by an optical fiber. This sharing is only possible if these lightpaths have different wavelengths. With the exponential evolution of traffic, the demand between source destination pairs in a network can approach or exceed the capacity of a single fiber. In this scenario, spacedivision multiplexing (SDM) networks appear as a viable option, and new network-sharing technologies become relevant. In this paper, we study different switching approaches for SDM networks with uncoupled spatial superchannels-such as networks with multicore fibers or bundles of single-mode fibers-which result in different network-sharing strategies. We start by comparing three switching architectures considering the evolution of traffic over time: (1) full-spectrum spatial switching (full-spectrum SS); (2) wavelength switching (WS) of uncoupled spatial superchannels; and (3) independent switching (IS) of spatial and wavelength channels. IS provides high network flexibility at the cost of complex switching nodes. On the other hand, full-spectrum SS and WS simplify network nodes but reduce flexibility by switching superchannels. Simulation results indicate that WS offers poor utilization in the long term. Full-spectrum SS, in contrast, exhibits low utilization in the short term, but outperforms IS in the long term. We also propose alternative hybrid switching strategies that start with IS and change to full-spectrum SS after a certain number of spatial channels are activated. The simulations suggest that well-designed strategies for migration from IS to full-spectrum SS delay premature activation investments while saving on switching costs10129911004CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO - CNPQFUNDAÇÃO DE AMPARO À PESQUISA DO ESTADO DE SÃO PAULO - FAPESPnão tem2015/04382-0; 2015/24341-