Keio University, Graduate School of Media and Governance
MAUI Project
Ph.D. Dissertation

[ English | Japanese ]
Back to Index Page

TITLE Robust High-Performance Real-Time Streaming Control

High-quality and high-performance real-time video streaming over internet protocol (IP) networks requires 1) maintaining the highest data transmission rate while effectively utilizing network resources and 2) minimizing data packet loss and transmission delay to achieve the best possible streaming quality. However, because current typical congestion controllers based on TCP-friendly rate control (TFRC) prevent occurrences of network congestion reacting susceptibly to packet loss, they cause a significant degradation of streaming quality due to low-achieving throughput and data packet losses.

In this dissertation, motivated by the deployment of wide-area high-speed networks, we propose the two streaming control mechanisms: 1) the dynamic probing forward error correction (DP-FEC) and 2) GENEVA, the streaming control algorithm using generalized multiplicative-increase/additive-decrease (GMIAD). DP-FEC can be utilized for the purposes of IPTV, e-learning and international symposiums that strongly require maintaining the best possible streaming quality. DP-FEC estimates network conditions by successively observing variation in the intervals between packet loss events, and adjusts the degree of FEC redundancy to make a packet loss tolerance as high as possible while minimizing the performance impact of competing TCP flows. GENEVA assumes a usage environment where real-time streaming flows with various data transmission rates and round trip times (RTT) compete (e.g., in the Internet). Using GMIAD mechanism for the adjustment of the degree of FEC redundancy, GENEVA utilizes expected network resources that competing TCP flows fail to copy, and combats bursty packet losses to improve FEC recovery capabilities while cooperating with other competing flows. The evaluation results show that DP-FEC and GENEVA enable high-performance streaming flows to retain higher streaming quality while minimizing an adverse impact on competing TCP performance. The proposed mechanisms can contribute largely to further growth and promotion of high-performance real-time streaming which is subject to an increasing demand for higher streaming quality.

Keywords: 1. Real-Time Streaming, 2. Forward Error Correction, 3. Congestion Control, 4. Internet

CONTACT To obtain the dissertation, please contact;
MATSUZONO, Kazuhisa ( kazuhisa [at] )

MAUI Proejct
Last update:

Back to Project Home Page