A Control Theory Approach for Congestion Control in Intranetwork

Abstract

This paper presents a congestion control technique in an intranetwork gateway by using a control theory approach. Such an approach is very much necessary to determine how essentially information regarding congestion can be sent to the hosts as a precautionary measure in time. As a result those hosts connected in the intra net will take necessary steps to avoid further worst case congestion in the gateway of the network. Our proposed scheme can be viewed as a two tier model, where admittance of various connections is done according to the traffic parameter declaration by the hosts in the first step and once accepted, periodic signals can be sent to the hosts to regulate their flow of traffic in accordance to the declared parameters. The main idea in our approach is to observe network states at different point of time at the gateway before sending any information to the hosts inside the network. In order to observe these states at the gateway, two important parameters such as process estimation and process parameters need to be calculated and to get them correctly, we propose to use Kalman filter in our approach. Kalman filter is derived from discrete time space analysis and is different from other filters in use, which estimates a process by using a form of feed back control. We believe, such an approach can be widely used in today’s Internet technology making use of Resource Reservation Protocol (RSVP) and other signalling protocols to achieve a desired Quality of Service (QoS) amongst a variety of users.