Modelling Network Energy Reductions Through Dynamic Wavelength Funtionality

Dynamic wavelength capability is well known as a technique with the potential to increase the utilization of network resources and has recently been shown to lead to greater energy efficiency [1]. We examine specifically the energy savings due to dynamic wavelength capabilities that are optimized around particular services and identify key enablers to achieve deep energy savings. The emergence of high bandwidth and/or persistent data flows in packet networks for which the routing energy is strongly determined by the header processing leads to inefficiencies due to a granularity mismatch between the flow size and the packet size. Breaking large flows into many small packets, each with its own unique header, increases the energy consumption. In the extreme example in which the service requires a full, continuous and persistent circuit, the benefits of packet transport are lost while still paying the higher cost in energy. Indeed, circuit or flow based services can be better implemented using end-to-end transparent paths through a network, thereby realizing greater efficiency. Many services, however, have a complex set of needs or performance requirements that must be taken into account to understand the opportunities for using transparent networking capabilities such as dynamic wavelength functionality. In this work, we consider specific service cases related to content distribution networks and examine the potential for energy efficiency gains using dynamic wavelength functionality.