Advanced Testbeds to Validate Physical Estimators in Heterogeneous Long Haul Transparent Optical Networks

— The capacity growth driven by increasing traffic and the introduction of new devices and technologies has resulted in new functionalities in optical networks. The property of transparency enables compatibility between different system generations and the coexistence of multiple bit-rates without raising the global network cost. In a transparent network, the signals travel through many links and nodes without the need for opto-electronic regeneration, accumulating physical impairments. Therefore, standard point-to-point test-bed setups are no longer adequate to emulate signals propagating in transparent optical networks. New test-beds integrating the different heterogeneity cases due to the use of transparent cross-connects becomes necessary. In this paper, we present various test-bed realizations accounting for the following network heterogeneities: fiber link heterogeneity and neighbor changes while a signal propagates in the fiber. To be able to emulate such network heterogeneity, we present a double-loop setup. Such a setup is then used to assess the propagation results obtained by simulation and validate the quality of transmission (QoT) estimators. I. INTRODUCTION he improvement of transmission devices enables both signal propagation over longer distances and also increases in the transported bit-rate. Improvements on optical switches for which the transparent pass-through and wavelength reconfiguration of optical signals is possible, have enabled greater flexibility and automation. Key characteristics of future networks include: higher transported capacity, greater ease of network upgradeability and greater flexibility to different technology evolution paths (i.e., transmission and/or modulation format). Underpinning all of these attributes is the advantage wavelength division multiplexing (WDM) technology which allows for the propagation of many channels on the same medium using shared resources. These properties reduce the number of devices that need to be changed to accommodate network evolution and hence the total cost over