We describe a stochastically-based master-slave dual system configuration approach to photonic networking design enabling intrinsic energy-efficiencies to dynamically approach comparable single-device systems that are fully elastic with respect to presented traffic loads. We show analytically that energy-efficiency improvements of more than 90% are statistically possible by adopting a master-slave design configuration. In addition to enabling a lower carbon footprint, our approach allows system switching-speed tolerances to be relaxed, and also improves system reliability as compared with a fully elastic, single-device approach.