An Evolutionary Based Dynamic Energy Management Framework for IP-over-DWDM Core Networks

Abstract

The energy consumption of the Internet accounts for approximately 1% of the world's total electricity usage, which may become the main constraint for its future growth of the Internet. In response, we propose a novel dynamic energy management framework that reduces the overall energy consumption without degrading network performance. The main concept is to combine infrastructure sleeping with virtual router migration. During off-peak hours, the virtual routers are moved onto fewer physical platforms and the unused physical platforms are placed in a sleep mode to save energy. The sleeping physical platforms are awakened again during busy periods. In our energy management framework, an important question that is considered is where to move the virtual routers to. To this end, we develop an evolutionary algorithm to solve the destination physical platform selection problem.