ABSTRACT
This chapter presents representative source-based congestion control mechanisms including transmission control protocol (TCP) and its variants which were reported to extend the applicability of the traditional TCP to network environments that may include long-distance high-speed and/or wireless connections that may intend to support quality demanding applications. The key idea of traditional TCP is for a source to gently probe the network for spare capacity by linearly increasing its rate and exponentially reducing its rate when congestion is detected. TCP Reno has performed remarkably well and is generally believed to have prevented severe congestion as the Internet scaled up by magnitudes in size and speed. The interesting property of scalable TCP is that the recovery time after a loss event is decoupled from the window size. The traditional TCP implementations rely on packet loss as an indicator of network congestion and lack the ability to distinguish congestion losses from losses invoked by noisy links.
