This article provides a theoretical framework for feedback-based congestion control techniques. Two broad classes of control algorithms are presented: source- and switch (network)-intelligence models. In a source-intelligence model, traffic sources adjust transmission rates according to congestion information provided by the network. Theoretical models based on queueing theory and delay differential equations are presented and are applicable to most congestion control techniques in existing networks. In a network with a large bandwidth-delay product, a delay in the feedback congestion signal may contribute a large amount of data loss. In order to overcome the difficulty, a network-intelligence model has switches evaluate appropriate transmission rates of sources. A control theoretic framework which is based on prediction of traffic pattern is presented.