Context aware self-configuring wi-fi network

Abstract

Wireless networks are increasingly popular among end users primarily to provide Internet access services. For this reason it is becoming easier to find highly congested zones with multiple wireless networks. Many of these networks use default settings which causes, in some cases, that some of communication channels in the 2.4GHz band are overused. This document presents the work done to design a context aware self-configuring system which is intended to improve the performance of independent WI-FI networks by avoiding the interference generated by neighboring wireless networks. For the realization of this system were carried out many experiments which show that wireless networks working on the same channels suffer interference among them that reduces the capacity of each network. In addition to these experiments multiple metrics were proposed in an attempt to predict the capacity of a channel. The number of interfering radios and the number of rogue data packets were the two proposed metric that better fit the needs for capacity forecast. From the obtained results a prototype that implements context aware self-configuring system was developed. The proposed algorithm uses a quality value which is calculated from the weighted sum of the two metrics explained before. This prototype was tested against two algorithms (random and static). The results show that the proposed system improves the mean Throughput of the wireless network under realistic conditions.