Learning Method of Performance-oriented Congestion Control (PCC) for Video Streaming Analysis

Abstract

Transmission Control Protocol (TCP) congestion control architecture suffers from performance issues that are not optimal. Thus making the TCP and the variations have little hope of achieving high performance. This is due to the use of hardwired TCP mapping where each event has been assumed to be a specific disruption and TCP must cope with the incident without understanding the real condition of the network. This assumption results in performance degradation. Performance-oriented Congestion Control (PCC) is a new congestion control that makes every sender observing the action and network performance empirically to be able to take action that yields high performance. PCC has been tested in several cases. One case is video streaming. The experiment is designed to obtain the performance for video streaming in terms of throughput, delay, and packet loss for the PCC and the TCP to determine better performance results. Results from the testing of each metric in which TCP and PCC throughputs are 1064.841 and 150.825 kbps respectively. Delay of TCP and PCC are 5.326 ms and 3.843 ms respectively. Packet loss of TCP and PCC are 0.905% and 0.016%, respectively. So the PCC achieves good performances on the parameters of delay and packet loss. Whereas the TCP is shown to perform better in terms of throughput.