An Additive Increase Smooth Decrease (AISD) Strategy for Data and Streaming Applications

Abstract

Voice over IP and video applications continue to increase the amount of real-time traffic over the Internet. These applications utilize the UDP protocol because TCP is not suitable for streaming applications since its congestion control mechanism can change the connection's transmission rate drastically, affecting the user-perceived quality of the transmission. Nonetheless, the use of end-to-end flow and congestion control mechanisms for streaming applications has been acknowledged as an important measure to ease or eliminate the congestion collapse problem in the Internet, and the unfairness problem that exist when TCP and UDP share the same congested bottleneck link. In this paper, we propose the smooth fair TCP SACK-based (SF-SACK) protocol to address these problems. SF-SACK implements the well-known end-to-end window-based congestion control algorithm of TCP but uses an additive increase smooth decrease (AISD) strategy that considers history in the evolution of the congestion window. Through simulations and experimentation, we show that if SF-SACK is used by both streaming and data-oriented applications, streaming applications receive a smooth service, and fairness is achieved. If SF-SACK is used for streaming applications only, it is unfair to TCP (not TCP-friendly) but substantially fairer and beneficial than UDP.