Efficient failure management in transparent optical networks (TONs) is critical due to the extremely large fiber throughput involved. Namely, fast and successful reaction and restoration mechanisms can prevent from loosing large amounts of data which can cause severe service disruption. These mechanisms rely on monitoring information exchanged via optical supervisory channels along each physical link. We propose applying common structural properties of self-organizing systems (specifically, the small-world and scale-free models) to help achieve faster and more robust system-wide communication in the context of failure management.