New algorithm for distributed frequency assignments in IEEE 802.11

Abstract

The continuous growth of IEEE 802.11 wireless local area networks (WLAN) brings the benefit of a high speed wireless access to packet networks, such as Internet. But it also entails the successive appearance of new unresolved problems. One of these problems consists in the degradation of the performance observed by the users when there is a great number of radio networks coexisting in the same area. In densely populated zones it is not strange to find WLANs of different nature (private, public, etc.) sharing a scarce resource as it is the radio spectrum. This problem can be mitigated with an appropriate channel allocation. In this paper we present an algorithm for the frequency assignment problem (FAP) in IEEE 802.11, based on classical graph colouring but adding the objective of reducing interferences among overlapping cells, and thus improving the global throughput performance. An accurate evaluation of interferences is obtained not only measuring power levels, but also being aware of traffic load. Important performance improvements have been observed on a real scenario.