Elsevier

Computer Communications

Volume 26, Issue 12, 21 July 2003, Pages 1240-1254
Computer Communications

The issue of useless packet transmission for multimedia over the Internet

https://doi.org/10.1016/S0140-3664(02)00276-1Get rights and content

Abstract

When packet loss rate exceeds a given threshold, received audio and video become unintelligible. A congested router transmitting multimedia packets, while inflicting a packet loss rate beyond a given threshold, effectively transmits useless packets. Useless packet transmission wastes router bandwidth when it is needed most. We propose an algorithm to avoid transmission of useless multimedia packets, and allocate the recovered bandwidth to competing TCP flows. We show that the proposed algorithm can be easily implemented in well-known WFQ and CSFQ fair packet queueing and discarding algorithms. Simulation of a 15-s MPEG-2 video clip over a congested network shows that the proposed algorithm effectively eliminates useless packet transmission, and as a result of that significantly improve throughput and file download times of concurrent TCP connections. For the simulated network, file download time is reduced by 55% for typical HTML files, 36% for typical image files, and up to 30% for typical video files. A peak-signal-to-noise-ratio (PSNR) based analysis shows that the overall intelligibility of the received video is no worse than that received without the incorporation of the proposed useless packet transmission avoidance algorithm. Our fairness analysis confirms that implementation of our algorithm into the fair algorithms (WFQ and CSFQ) does not have any adverse effect on the fairness performance of the algorithms.

Introduction

Internet voice and video applications, such as IP telephony and video streaming, continue to gain popularity. A direct consequence is the increased congestion in the Internet routers. In the routers, multimedia packets, i.e. IP packets carrying voice or video traffic, compete head to head with traditional data packets for link bandwidth. Two important issues that need to be addressed when multimedia traffic is multiplexed with data traffic are: (i) fairness, and (ii) useless packet transmissions (UPT). The fairness issue arises due to the fact that the data applications use TCP, which cuts back the transmission rate when packets are discarded at the router due to congestion, but most multimedia applications, use UDP, which does not cut back transmission rates. As a result, multimedia flows unfairly gets more bandwidth than the congestion-responsive TCP flows. The fairness problem in the Internet is now well recognised. Several packet queueing and discarding algorithms, such as WFQ [1], CSFQ [2] and FRED [3], have been proposed in the last few years to effectively address the issue of fairness. Performance results confirm that these algorithms can fairly distribute link bandwidth among competing multimedia and data flows. Some router vendors have already started incorporating these algorithms in their latest products [4].

The issue of UPT, however, is less understood. UPT is based on the fact that for packetised audio and video, packet loss rate must be maintained under a given threshold for any meaningful communication [5], [6], [7]. When packet loss rate exceeds this threshold, received audio and video become useless. Thus a router transmitting multimedia packets at a fair rate (using WFQ for example), while inflicting a packet loss rate beyond the threshold actually transmits useless packets. These packets are useless, because they do not contribute to any meaningful communication. UPT effectively reduces the available bandwidth to competing TCP flows, which in turn increases file download times. Longer download times increase power consumption of battery-powered devices, and hence, have direct impact on mobile computing.

The UPT problem will be more significant in future due to the following trends: (i) rising popularity of voice and video applications, (ii) increasing deployment of fair packet queueing algorithms, and (iii) rapid proliferation of battery-powered mobile devices. Hence, there is a need to investigate mechanisms to effectively address the UPT issue in fair packet queueing algorithms. In this paper, we propose a UPT avoidance algorithm (we call it UPTA) that can be easily implemented in existing fair packet queueing and discarding algorithms. We have evaluated its performance using simulation of an MPEG-2 video stream competing with a TCP connection in a congested router. Simulation results show that with the proposed UPTA algorithm in place, we can effectively eliminate UPT, and significantly reduce file download times without deteriorating the perceived quality of the video stream.

The remainder of the article is organised as follows. We discuss related work in Section 2. Section 3 formally defines the UPT problem and derives the performance parameters. In Section 4, we propose an UPTA algorithm and show its incorporation in two well-known fair packet queueing algorithms, WFQ and CSFQ. In Section 5, we derive the packet loss rate threshold for intelligible communication for the MPEG-2 video clip used in the simulation. The simulation experiment is presented in Section 6, followed by the results in Section 7. Finally, we present our conclusions and discuss open issues in Section 8.

Section snippets

Related work

How to support multimedia over the Internet is a topic of intense research. In this section, we discuss the related work.

The UPT problem

This section details the UPT problem and formally defines the performance parameters. We make the following assumptions and observations for multimedia over best-effort network service:

  • When packet loss rate exceeds a threshold q, the media, particularly voice and video, becomes unintelligible [5], [6], [7]. The exact value of this threshold may vary from application to application. For the video clip used in our experiment, we experimentally derive this threshold (see Section 5). As long as the

UPT avoidance algorithm

There are two basic ways to avoid UPT. One way is to eliminate all U intervals; the other is simply to remove UPT from the U intervals. The former approach, which needs to reserve bandwidth for multimedia connections, is pursued by the guaranteed quality of service (QoS) efforts (see related work in Section 2). However, for best-effort network services, it is acceptable to expect occasional U intervals in multimedia connections. We propose an UPT avoidance algorithm (henceforth called UPTA),

Packet loss threshold for the experimental video

Packet loss rate threshold q for intelligible communication may vary from application to application. We have carried out a series of experiments with different loss rates in the network to determine the threshold for the MPEG-2 video clip used in our simulation study. This section explains the experiment with a brief overview of the standards used for transporting MPEG-2 video over IP.

Simulation

This section details the network model and the performance metrics used in the simulation study.

Results

In this section, we present the results obtained from the simulation experiments.

Conclusion

We have investigated the issue of UPT in IP routers when multimedia applications use best-effort network services. With best-effort service, available bandwidth is equally allocated to all competing flows using fair packet queueing and scheduling algorithms (e.g. WFQ and CSFQ). TCP-based data applications can tolerate any bandwidth, but multimedia become unintelligible when allocated bandwidth is too low and packet loss rate exceeds a given threshold. Multimedia packets transmitted by the

Jim Wu is a full-time PhD student at School of Computer Science & Engineering, The University of New South Wales. He submitted his PhD thesis for examination on 15 November 2002. Mr. Wu received his B.S. degree (in Electronics & Information Systems) from Zhong Shan University (China) in 1988, and his Graduate Diploma and Master degree (both in Digital Communications) from Monash University, in 1995 and 1997 respectively. He worked as a system engineer at No. 34 Research Institute of Electronic

References (26)

  • M. Atiquzzaman et al.

    Guest editorial: adaptive real-time multimedia transmission over packet switching networks

    Real-Time Imaging

    (2001)
  • D. Chiu et al.

    Analysis of the increase/decrease algorithms for congestion avoidance in computer networks

    Computer Networks and ISDN Systems

    (1989)
  • A. Demers et al.

    Analysis and simulation of a fair queueing algorithm

    Internetworking: Research and Experience

    (1990)
  • I. Stoica et al.

    Core-stateless fair queueing: achieving approximately fair bandwidth allocations in high speed networks

    Proceedings of ACM SIGCOMM, Vancouver, Canada

    (1998)
  • D. Lin et al.

    Dynamics of random early detection

    Proceedings of ACM SIGCOMM

    (1997)
  • C.S. Inc, Advanced QoS Services for the Intelligent Internet,...
  • N. Chapman et al.

    Digital Multimedia

    (2000)
  • F. Fluckiger

    Understanding Networked Multimedia: Applications and Technology

    (1995)
  • J.F. Kurose et al.

    Computer Networking: A Top-Down Approach Featuring the Internet

    (2001)
  • O. Verscheure et al.

    User-oriented QoS in packet video delivery

    IEEE Network

    (1998)
  • D. Hoffman et al.

    RTP payload format for MPEG1/MPEG2 video

    RFC 2250

    (1998)
  • Tektronix,...
  • K. Rao et al.

    Techniques and Standards for Image, Video, and Audio Coding

    (1996)
  • Cited by (5)

    • Avoiding useless packet transmission for multimedia over IP networks: The case of multiple multimedia flows

      2004, Computer Communications
      Citation Excerpt :

      Supporting multimedia over the Internet introduces a few problems. Notably, fairness and UPT are two major issues, as we pointed out in Ref. [1]. How to support multimedia over the Internet in an efficient way has become a hot topic in recent years.

    • An improved packet loss concealment technique for speech transmission in VOIP

      2018, 2nd International Conference on Natural Language and Speech Processing, ICNLSP 2018
    • A novel no-latency simple-to-implement sender-based packet-loss recovery technique for multimedia streams

      2006, Proceedings of ISCN'06: 7th International Symosium on Computer Networks
    • On reciprocal altruism and its application to QoS

      2003, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

    Jim Wu is a full-time PhD student at School of Computer Science & Engineering, The University of New South Wales. He submitted his PhD thesis for examination on 15 November 2002. Mr. Wu received his B.S. degree (in Electronics & Information Systems) from Zhong Shan University (China) in 1988, and his Graduate Diploma and Master degree (both in Digital Communications) from Monash University, in 1995 and 1997 respectively. He worked as a system engineer at No. 34 Research Institute of Electronic Industry Ministry (Guilin, China) before he came to Australia in 1994. He worked as an associate lecturer from 1998 to 2002, at Monash University and The University of New South Wales. Mr. Wu is currently employed as a research assistant at Network Research Laboratory, School of Computer Science & Engineering, The University of New South Wales. His research interests include congestion management, multimedia over IP, QoS for IP networks, etc.

    Dr. Mahbub Hassan is an Associate Professor in the School of Computer Science and Engineering, University of New South Wales, Sydney, Australia, where he directs advanced research and development activities in the Network Research Laboratory. He received PhD from Monash University, Melbourne, Australia and MSc from University of Victoria, Canada. He serves in the Editorial Advisory Board of Computer Communications journal (Elsevier Science) and previously served as an Associate Technical Editor for IEEE Communications Magazine. He was Guest Editor for Real Time Imaging Journal and Journal of Supercomputing. He has near 100 refereed publications, including four books. Dr. Hassan was the recipient of the Teaching Excellence Award from the School of Computer Science, Monash University, 1999. He has chaired SPIE conference on QoS, 2001 (Denver) and 2002 (Boston). He is a Senior Member of IEEE. Dr. Hassan also serves as consultant to industry and government. Consulting services range from advanced short course development and delivery to network analysis and design. Past and current clients include NEC, Telstra, Lucent and Canon

    View full text