Hotspot Mitigation With Measured Node Throughput in Mobile Ad Hoc Networks

A network node or region is said to be experiencing hotspot when it is subjected to a traffic load substantially larger than the design load, creating a situation wherein demand for network resources exceeds that available for use. Due to the dynamics and lack of central control in mobile ad hoc wireless network (MANET), it is common to have imbalance in the distribution of network traffic load. This issue compounded by the utilization of shortest path route selection metrics makes hotspot intrinsic to MANET. Under hotspot conditions a network would experience high contention and congestion that may in turn result in increased packet loss, end-to-end delay, transmission errors and faster battery power depletion. This phenomenon generally degrades the overall performance of the network. At a node, hotspot can be detected through high buffer occupancy, packet loss, and excessive MAC contention delay. No single one of these metrics alone can accurately detect hotspot. Monitoring these parameters independently and combining them to accurately detect hotspot would result in a complex scheme. In this paper, we are proposing a simple and effective hotspot mitigation scheme ´HUNT´ based on the throughput of network nodes. Through simulation experimentation, we prove that HUNT can effectively mitigate hotspot in best effort on-demand routing protocols such as DSR and yields improved efficiency, reliability and robustness