ABSTRACT: This proposal is devoted to enable Future Wavelength-Routed-Optical Networks (WRONs) exploiting Optical-Hybrid Amplifiers (HAs) as key devices to use multiple optical-transmission bands. We analyze and improve HAs performances under specific scenarios by simulation and experimental testbeds.

Current WRONs must increase their data traffic at rates that can go up to 50% per year, because of emerging applications such as video streaming, cloud computing, social networks, among others. The ever-growing need for higher data-transmission capacities will soon consume the full capacity of the already installed WRONs based on Wavelength-Division Multiplexing (WDM). Mainly three approaches are being explored to face this problem: 1. Band-Division multiplexing (BDM); 2. Space-Division Multiplexing (SDM); and 3. the migration towards systems managing optical resources more efficiently. The combination of the first and third approaches avoids deploying new fiber infrastructure. HAs, defined as the concatenation of rare-earth-Doped-Fiber Amplifiers (DFAs) and Distributed-Fiber-Raman Amplifiers (DFRAs), allow to extend the gain bandwidth of the already installed networks. They are key devices to address the first approach by allowing BDM WRONs (using optical-transmission bands E, S, C, L).

In order to cover long distances, WRONs include optical amplifiers which are nonlinear components. Their gain is very sensitive to optical-input power (total power of all wavelengths carrying data, defined by the WRON). As a consequence, any variation of load (number of WDM channels) when adding/removing channels at intermediate nodes will modify the gain, inducing a degradation of the signal-transmission quality. This can happen according to the routing and can change due to a dynamic reconfiguration of the network in case of link failure; or an increase of the network’s capacity. The fact of dealing with a variable load, leads to power transients (or power-gain variations). On one hand, when dropping channels, optical amplifiers are less saturated and their gain increases, so the output power of the surviving channels may exceed the threshold where fiber nonlinearities appear. On the other hand, when adding channels, optical amplifiers get more saturated and their gain decreases, so the output power of the surviving channels shall decrease and may fall below the receiver sensitivity. Both scenarios can significantly damage the Optical-Signal-to-Noise Ratio (OSNR) and thus the allowed bit rate. Therefore, to maintain the overall network performance, the power of the surviving channels must be kept within acceptable limits.

To fully overcome these issues, we define a new global strategy based on three features: (1) Taking into account a real-case scenario for which classical techniques might underestimate the gain variation and poorly take into account the cascade effect over an optical route (a lightpath passing through several nodes where the load changes occur). As a matter of fact, literature related to power-transient effects in optical amplifiers usually focuses on analyzing one link over the route, this means they add/drop channels only once, neglecting phenomena worth to study. (2) The use of more powerful control techniques. Until now, most of control techniques for reducing power transients are based on Proportional-Integrative (PI) controllers. Therefore, we propose the implementation of the Minimum-Variance Controller (MVC) to be adapted to the link-control technique, since it may better adapt to the scenario described above. 3. An experimental validation, which implies the development of a specific testbed. All features to be assessed in terms of optical power and power excursions of the surviving channels. Thus, the novelty of this proposal covers (1) The analysis of transient effects of HAs under specific scenarios of extremely-high-load variation through a whole optical route by means simulation. (2) The development of a new control technique specifically adapted to these scenarios. (3) The experimental validation of this technique, exploiting a specifically-designed testbed.

This proposal is supported by 2 contributing institutions and 8 experienced collaborating researchers. It also includes diffusion activities to disseminate research results and to impact positively the community.