Green Wi-Fi Management: Implementation on Partially Overlapped Channels

Abstract

The channel assignment in the non-overlapped channel space may result in low spatial utilization and non-uniform Wi-Fi performance, which makes it difficult to cater to the rapid increase in Wi-Fi devices as well as wireless traffic demand. In this paper, we propose a Green Wi-Fi management framework for an enterprise environment in order to densely deploy the access points (APs), to appropriately turn off a portion of the APs and coordinate the APs with the partially overlapped channels (POCs). Commencing with a power monitoring experiment, we first demonstrate that energy efficiency can be achieved when clients connect the Internet with fewer APs, followed by a channel assignment experiment showing that the dense deployment of APs with POCs is capable of improving the overall channel capacity. In order to balance the energy consumption and the network capacity, the concept of AP pattern is proposed to denote the active APs and their corresponding POCs allocated including a backtracking Tabu search algorithm for yielding alternative AP patterns. Furthermore, relying on the alternative AP patterns considered, a reinforcement learning aided algorithm is presented for determining the optimal AP pattern in terms of the active interference measurement. Finally, extensive field experiments are conducted to evaluate our Green Wi-Fi framework.