Congestion Control in the Internet by Employing a Ratio dependent Plant Herbivore Carnivorous Model

Abstract

The demand for Internet based services has exploded over the last decade. Many organizations use the Internet and particularly the World Wide Web as their primary medium for communication and business. This phenomenal growth has dramatically increased the performance requirements for the Internet. To have a high performance Internet, a good congestion control system is essential for it. The current work proposes that the congestion control in the Internet can be inspired from the population control tactics of the nature. Toward this idea, each flow (W) in the network is viewed as a species whose population size is congestion window size of the flow. By this assumption, congestion control problem is redefined as population control of flow species. This paper defines a three trophic food chain analogy in congestion control area, and gives a ratio dependent model to control population size of W species within this plant herbivore carnivorous food chain. Simulation results show that this model achieves fair bandwidth allocation, high utilization and small queue size. It does not maintain any per flow state in routers and have few computational loads per packet, which makes it scalable.