The rapid expansion of global communications has placed increased requirements on optical fiber infrastructure and has stimulated the demand for increased system capacity. As a result, high data rate, dense wavelength-division-multiplexed (WDM) systems are being widely deployed in long-haul terrestrial and trans-oceanic links. In these systems, polarization effects can severely limit the performance of the system.
System designers commonly allocate a prescribed power penalty margin to polarization effects say 2 dB. An outage is said to occur when this margin is exceeded. A common constraint is to limit outage to a probability of 10-6. Because outages are rare events, it is difficult to quantify their probability of occurrence either theoretically or experimentally. In the past year, we have achieved a theoretical breakthrough by the use of importance sampling, which allows us to accurately quantify rare events. We both extend this technique theoretically and apply it for the first time experimentally. Our focus is on determining the causes of outage and, from that, effective mitigation techniques.
The proposed research is based on a close experimental and theoretical cooperation to study the issue of channel outage due to polarization effects in long haul, high data-rate, terrestrial WDM transmission systems. A series of experiments will be carried out and theoretically modeled to study the interactions of polarization effects with nonlinearity, chromatic dispersion, and amplified spontaneous emission noise, and to assess the impact these combined effects have on channel outage for both RZ and NRZ formatted data. We will experimentally determine the region of validity of a reduced model that only follows the Stokes parameters in a WDM system, rather than the full time-domain behavior. We have already validated this model in simulations. Other modeling tools will be developed and validated as needed to accurately and efficiently calculate the channel outage probability due to polarization effects.
