Local Traffic-aware Green Algorithm based on Sleep-scheduling in autonomous networks

Over recent years, green communication has been investigated to reduce the economic costsand environmental pollution caused by network infrastructures. There exist many approachesto conserve energy in autonomous networks while according to over-provisioning of networkingresources for peak traffic periods, sleep-scheduling based approach is one of effective methodsbetween them. In this regards, in this paper, a new Local Traffic-aware Green Algorithmbased on Sleep-scheduling (LTGAS) is proposed in autonomous wired networks in order toreduce energy consumption during low traffic periods. This proposal is a dynamic solutionthat considers online and local traffic knowledge to switch off some underutilized networknodes and links during low traffic periods and re-switch on some sleep devices in congestionsituations. In addition, LTGAS is executed in fully distributed mode without any centralizedcontroller and uses only information provided by link-state protocols like OSPF. Moreover, thissolution does not have to gather global traffic knowledge of network during a sleep-schedulingdecision, resulting in low complexity and low network overhead. The results obtained by thereal network scenario show that LTGAS switches off up to 38% of network nodes and up to 96%of the maximum number of switchable links during low traffic periods. In addition, LTGAS hasa low effect on network performance metrics, including end-to-end delay, packet delivery ratioand average link utilization, compared with similar proposals while switches off more numberof network devices rather than other approaches