Hotspot clustering for ad hoc network

The adaptive clustering algorithm proposed by Gerla is purely based on geographical location freedom degree and plus time freedom degree, so it's two dimensional. In this paper, three dimensional clustering based on data flow is proposed. Three freedom degrees, data flow, the diameter of clusters, and time, are considered in the course of cluster establishment to solve hotspot phenomena. Hotspots are highly congested regions in ad hoc networks that can result in packet loss, end to end delay and out of order packets delivery. Through constructing clusters around hotspots, data flow coming to hotspots can be distributed to other nodes in the cluster with alternative route to the destination. Then it can mitigate congestion of hotspots and improve performance of wireless ad hoc network. Results from simulations on the OMNET++ platform show that various QoS indicators of the network get improved by use of hotspot clustering algorithm. And resource distribution of the whole network emerges as equilibrium effects.