Elsevier

Computer Communications

Volume 31, Issue 10, 25 June 2008, Pages 2621-2630
Computer Communications

Buffer-driven adaptive video streaming with TCP-friendliness

https://doi.org/10.1016/j.comcom.2008.02.011Get rights and content

Abstract

Existing streaming mechanisms have no consideration for the characteristics of streaming applications because they only consider network stability. In this paper, in order to overcome limitations of the previous work on video streaming, we propose a new adaptive video streaming mechanism. The proposed mechanism includes more sophisticated features that consider both network and user requirements. From the network viewpoint, the proposed mechanism can estimate the TCP-friendly rate suitable for current network conditions. Therefore, it is able to adjust the sending rate of video stream in a TCP-friendly manner and improves the network stability by reducing the packet losses. From the user viewpoint, our mechanism provides the smooth playback by preventing the receiver buffer underflow or overflow. Moreover, it is designed to take into consideration the streaming video content. Through the simulation, we prove that our mechanism efficiently uses the buffer resources and provides improved network stability with TCP-friendliness and smooth playback.

Introduction

The Internet has recently been experiencing an explosive growth in the use of audio and video streaming applications. Such applications are delay-sensitive, semi-reliable, and rate-based. Loss-tolerant video streaming applications, such as video conferencing, prefer UDP (User Datagram Protocol) as a transport protocol that has no congestion control algorithm. The emergence of non-congestion-controlled streaming applications threatens unfairness to competing TCP (transmission control protocol) traffic and possible congestion collapse. UDP is considered selfish and ill-behaving because TCP throttles its transmission rate against the network congestion, whereas UDP does not have such a congestion control algorithm. As a result, the available bandwidth to TCP connections is oppressed and their performance sharply deteriorates because the current Internet does not attempt to guarantee an upper limit on end-to-end delay and a lower limit on available bandwidth [1].

Studies on the congestion-controlled streaming mechanism has been increasingly done since the 1990s [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. These works attempt to guarantee the network stability and fairness with competing TCP connections. In [1], non-TCP connections are defined as TCP-friendly when “their long-term throughput does not exceed the throughput of a conformant TCP connection under the same network conditions”. TCP-friendly rate control mechanisms regulate its data sending rate according to the network condition, typically expressed in terms of the RTT (round trip time) and the packet loss probability. Therefore, they improve the network stability and fairness with competing TCP connections. However, by considering only the network stability, these works ignore the quality of video stream which is delivered to the user. Moreover, most of existing streaming mechanisms have no consideration for the characteristics of video stream which affects significantly the quality of streaming services.

In this paper, we propose a new adaptive video streaming mechanism. The proposed mechanism adjusts the sending rate of video stream-based on the current network state and controls the quality of video stream-based on the receiver buffer occupancy. It also considers the characteristics of video stream which is requested from the user. Therefore, our mechanism maintains the network stability accomplished by previous works and achieves the smooth playback by preventing the buffer underflow or overflow.

The rest of this paper is organized as follows. In Section 2, we review and discuss some of the previous works and in Section 3, we present the concept and algorithms introduced in our mechanism. Detailed description of our simulation results are presented in Section 4. Finally, Section 5 concludes the paper and discusses some of our future work.

Section snippets

Related work

While data applications such as Web and FTP (File Transfer Protocol) are mostly based on TCP, video streaming applications will be based on UDP due to its real-time characteristics. However, UDP does not support a congestion control algorithm. For this reason, UDP-based streaming mechanisms cause the unfairness with competing TCP traffic and the starvation of congestion-controlled TCP traffic which reduces its bandwidth share during overload situations. To overcome this limitation, several

Adaptive video streaming mechanism

This section briefly introduces our adaptive video streaming mechanism to adjust the sending rate of the video stream to the desired TCP-friendly rate. First, we describe the end-to-end architecture of the proposed mechanism. After that, the detailed algorithms are described.

Simulation environment

In this section, we present our simulation results. Using the ns-2 simulator, the performance of the proposed adaptive video streaming mechanism has been measured and compared with the TFRC and the buffer-driven scheme [20]. To emulate the competing network conditions, background TCP traffics are introduced.

Fig. 6 shows the network topology for our simulations. All of our experiments use a single bottleneck topology and the bottleneck queue is managed with the drop-tail mechanism. The RTCP

Conclusion

Most of streaming mechanisms are based on UDP with no end-to-end congestion control. For this reason, wide usage of multimedia applications in Internet might lead to congested networks. To avoid such a situation, several congestion controlled streaming mechanisms were proposed recently. However, by considering only the network stability, most of previous works ignore the requirements of streaming services and the characteristics of the video stream.

In this paper, in order to overcome

Acknowledgement

This research has been conducted by the Research Grant of Kwangwoon University in 2007.

References (20)

  • S. Floyd et al.

    Promoting the use of end-to-end congestion control in the Internet

    IEEE/ACM Transactions on Networking

    (1999)
  • S. Cen et al.

    Flow and congestion control for internet streaming applications

    Multimedia Computing and Networking

    (1998)
  • R. Rejaie et al.

    RAP: an end-to-end rate based congestion control mechanism for real-time streams in the Internet

    IEEE INFOCOMM

    (1999)
  • B. Mukherjee, T. Brecht, Time-lined TCP for the TCP-friendly delivery of streaming media, International Conference on...
  • S. Na, J. Ahn, TCP-like flow control algorithm for real-time applications, IEEE International Conference on Networking,...
  • I. Rhee, V. Ozdemir, Y. Yi, TEAR: TCP emulation at receivers – flow control for multimedia streaming, Technical Report,...
  • D. Bansal et al.

    Binomial congestion control algorithms

    IEEE INFOCOMM

    (2001)
  • J. Padhye, J. Kurose, D. Towsley, R. Koodli, A model based TCP-friendly rate control protocol, International Workshop...
  • S. Floyd et al.

    Equation-based congestion control for unicast applications

    ACM SIGCOMM

    (2000)
  • Q. Zhang et al.

    Resource allocation for multimedia streaming over the Internet

    IEEE Transactions on Multimedia

    (2001)
There are more references available in the full text version of this article.

Cited by (14)

  • Adaptation strategies for MGS scalable video streaming

    2012, Signal Processing: Image Communication
    Citation Excerpt :

    In case DCCP CCID3 is used, the TFRC rate calculated by DCCP can be utilized by the sender to estimate the available network rate, without requiring an additional estimation mechanism. Otherwise, the video send rate can be estimated using receiver buffer occupancy information to prevent any buffer underflow/overflow [12,21] or by combining the receiver buffer state with the bandwidth estimate [7]. A virtual network buffer between the sender and receiver is employed in [3] for video rate adaptation together with an end-to-end delay constraint; while the same virtual network buffer algorithm is used in [13] to implement source rate control and congestion control jointly.

  • Load Modelling and Generation in IP-based Networks: A unified approach and tool support

    2017, Load Modelling and Generation in IP-based Networks: A Unified Approach and Tool Support
View all citing articles on Scopus
View full text