Energy and Delay-Aware Traffic Control and Management in Large Scale Networks
Qun Yu, Taieb F. Znati · 2017
Reducing power and energy consumption in large scale networks remains a challenging problem. The increasing need to support multimedia applications in future Internets further compound this problem. This paper addresses a key issue of how to efficiently assign per-router flow delays and set per-processor execution speeds, along a routing path, to jointly minimize energy consumption and meet end-to-end delay requirements of the underlying applications. To this end, we propose a DVFS-enabled, energy-aware strategy to optimize the energy consumption of network components through scaling their processor execution speeds. The strategy takes into consideration the workload and the delay requirements of the applications. Given a flow specification, which characterizes the flow's traffic rate and QoS performance requirements, a per-router feasible minimum and maximum delay values are computed. Using these values, the energy- and delay-aware problem is modeled as a routing path energy-minimization problem to determine, for each router along the path, the processing router execution rate and the flow per-router delay budget that minimize energy without violating the flow's end-to-end delay requirement. A simulation-based analysis shows that energy minimization can be achieved without QoS violation. The results show that up to 88:33% dynamic energy saving and up to 26:76% power saving of the total power consumption can be achieved by the proposed strategy.