Rate Control for Streaming over WirelessJoint work with Avideh Zakhor from UC Berkeley.
Rate control is an important issue in video streaming applications for both wired and wireless networks. A widely accepted rate control method in wired networks is equation based rate control, in which the TCP Friendly rate is determined as a function of packet loss rate, round trip time and packet size. This approach, also known as TFRC, assumes that packet loss in wired networks is primarily due to congestion, and as such is not applicable to wireless networks in which the bulk of packet loss is due to error at the physical layer. Hence, for streaming over wireless, rate control is still an open issue. A traditional point of view is to distinguish between packet loss caused by congestions and that due to wireless channel error, by modifications to the network infrastructure or protocol. Thus, by not taking into account the packet loss due to wireless channel error when determining the sending rate, the new problem, i.e. rate control for streaming over wireless, is translated into an old one, i.e. rate control for streaming over wired network, for which a known solution exists. The solutions based on this point of view are not easy to deploy because they require modifications to network infrastructure or protocol. In this project we propose new insights into this old problem: one can achieve highest throughput and lowest packet loss rate, without distinguishing between packet loss caused by congestion and that due to wireless error. Specifically, we argue that from application level point of view, it is conceivable to use multiple TFRC connections to increase the total throughput to pursue full utilization, without distinguishing between packet loss caused by congestion and that due to wireless error. Further, we show this is actually one way to get the optimal performance, i.e. highest throughput and lowest packet loss rate, given we open an appropriate number of connections with an appropriate packet size. In order to determine the optimal number of connections to open, we design a simple algorithm based on average queuing delay measurements. The fundamental idea is to increase the number of connections if the average queuing delay is less than a threshold, and to decrease otherwise. The increasing and decreasing algorithm is Inverse Increase and Additive Decrease (IIAD). NS-2 simulations and real experiments on Verizon Wireless 1xRTT CDMA data network have been carried out to evaluate the performance, and to validate the basic idea. Our approach is also applicable to TCP over wireless. This work indicates new class of approaches to do flow control over wireless network, which motivates my thesis research. The idea first appeared in the following paper:
An improved rate control scheme based on mimicing MULTFRC by modifying TFRC protocol, namely All-in-one TFRC (AIO-TFRC), is also proposed as an alternative solution. The advantages are smaller overhead and better bandwidth utilization. The paper is here:
As archives, the most thorough version of the work appeared in IEEE Trans. on Multimedia, as well as in my PhD thesis:
|