Optimization of AIMD Congestion Control for Media-Streaming Applications

Media-streaming applications require smooth patterns of data transmission and also benefit from efficient resource utilization. System-wise, the underlying congestion control mechanism should achieve fairness and maintain TCP-friendliness. In this context, we optimize Additive Increase Multiplicative Decrease (AIMD) congestion control for multimedia applications, within the framework of bandwidth efficiency, smoothness, and inter-protocol fairness. We assume Scalable Streaming Video Protocol (SSVP) as the underlying congestion control mechanism. Departing from SSVP’s throughput model and based on the concepts of the knee and the cliff as defined in [4], we provide an analysis of AIMD congestion control, taking into account the role of the bottleneck queue. We observe that although multiplicative decrease is necessary to achieve fairness, it does not necessarily sacrifice the system throughput, as long as the system operates between the knee and the cliff. The proposed AIMD mechanism introduces congestion control parameters adaptable to current network conditions, preventing the system from operating below the knee, where a fraction of the available bandwidth is not utilized and smoothness is compromised as throughput fluctuates.