An energy‐efficient algorithm based on sleep‐scheduling in IP backbone networks

Abstract

SummaryThis paper relates to the problem of energy conservation in Internet backbone networks. We propose a new Distributed Green Algorithm based on Sleep‐scheduling (DGAS) to switch off some network links and nodes resulting in less energy consumption. Our solution does not depend on any centralized controller and uses a link‐state protocol like Open Shortest Path First to share required information. In addition, DGAS does not require any knowledge of traffic matrix and switches off network nodes and links in 2 phases. In the first phase, some core nodes are switched off on the basis of their usage in network paths. In the second phase, DGAS starts with a minimal topology on the basis of resulted topology from the first phase and adds some links to it to minimize network paths length. Results, obtained by several real network scenarios, show that DGAS can switch off up to 27% of core nodes and up to 80% of links compared to the maximum number of switchable links.