FEC recovery performance for video streaming services over wired-wireless networks

https://doi.org/10.1016/j.peva.2008.12.002Get rights and content

Abstract

This paper considers the packet recovery performance of forward error correction (FEC) for video streaming services over wired-wireless networks. Focusing on a wireless base station, we model it as a single-server queueing system with a Markovian service process in which the state of the server alternates between Good and Bad states. The system has two independent input processes: one is a general renewal input process and the other is a Poisson arrival process. We analyze the packet- and block-level loss probabilities to investigate the recovery performance of FEC at block level. The analysis is validated with simulation experiments driven by real traffic traces. Numerical examples show that the block-loss probability is greatly affected by the system capacity and the mean Bad-state period, and that the recovery performance of FEC deteriorates according to the fluctuation in the packet transmission rate at a wireless base station.

Introduction

Recent advancement of video coding techniques and widespread use of broadband networks have accelerated the development of real-time applications such as video streaming, sports live broadcasting and web conference. With the rapid growth of wireless networks, these applications are expected to be widely deployed over wireless networks. Because video applications have stringent delay constraints, many techniques have been studied to guarantee video image quality over wireless networks where burst packet loss occurs due to mobility and interference.

There are two basic techniques for packet-loss recovery: Automatic Repeat reQuest (ARQ) and Forward Error Correction (FEC). ARQ is a typical acknowledgement-based error recovery technique. In ARQ, lost data packets are retransmitted by the sender host. However, this retransmission mechanism is activated by receiving duplicate acknowledgement (ACK) packets or timer time-out, causing a large end-to-end delay. This large delay is not suitable for real-time applications such as video streaming and web conference.

On the other hand, FEC is a well-known coding-based error recovery scheme [1], [2]. In FEC, redundant data is generated from original data, and a sender host transmits both the original and redundant data to a receiver host. When some part of the original data is lost, it can be recovered from the redundant data at the receiver host if the loss is below a prespecified level. In this paper, we focus on packet-level FEC scheme [3]. When N redundant data packets are generated from D original data packets, the lost data can be recovered completely if the number of lost packets is less than or equal to N. Because FEC requires no retransmission mechanism, it is a suitable packet-loss recovery scheme for video applications with stringent delay constraint.

With the recent development of optical networking technology such as wavelength division multiplexing (WDM), the bottleneck of data transmission shifts from backbone networks to access ones (the last mile bandwidth bottleneck [4]). On the other hand, wireless mesh networks (WMNs) have attracted considerable attention as a solution of the last mile issue for access networks [5]. WMNs consist of wireless mesh routers and mesh clients, and are expected to support a variety of applications to end users.

Now, consider video streaming services over the wired-wireless network consisting of optical backbone networks and wireless access networks. Here, video streaming servers are placed on the backbone networks, while a client node is connected to a wireless base station with one-hop wireless link. In this situation, the wireless base station is likely to be the bottleneck of data transmission for video streaming service due to interference and user mobility. The quality of service (QoS) of video streaming over the wired-wireless network is significantly affected by the packet loss process at the wireless base station. Note that data packets for video streaming are sent to the client node at a constant bit rate. Because the optical backbone networks provide a high-speed transmission, it is important to consider the case where inter-arrival times of packets to the edge wireless base station are almost the same.

In this paper, focusing on the wireless base station, we consider the packet recovery performance of FEC at the wireless base station. We consider a single-server queueing system with finite buffer, in which the service time of a customer is governed by a two-state Markovian service process. The system has two inputs: main traffic and background traffic. Main traffic consists of original packets and FEC redundant ones. We assume that the inter-arrival times of packets in main traffic are independent and identically distributed (i.i.d.) according to a general distribution. On the other hand, arriving packets in background traffic form a Poisson process. Note that the assumption of main traffic enables us to describe various arrival processes including constant inter-arrival times. We analyze the packet- and block-level loss probabilities of main traffic, evaluating the recovery performance of FEC.

The rest of this paper is organized as follows: Section 2 gives an overview of the previous studies on the performance analysis of FEC recovery. Section 3 describes our analytical model, and Section 4 derives performance measures. Numerical examples are presented in Section 5, and finally Section 6 provides some conclusions.

Section snippets

Related work

It is well known that the recovery performance of FEC is significantly affected by the packet loss process, and much effort has been devoted to the development of adaptive QoS control schemes to improve the video quality [6], [7], [8], [9]. The relation between the recovery performance and the redundancy of FEC has also been extensively studied in the literature. A pioneering work is [10] in which the distribution of the number of lost packets within a block of packets is analyzed for an

GI+M/MSP/1/K queueing model

We consider the transmission of a data block consisting of D packets. Suppose N redundant packets are generated from the original D packets, and a set of M packets are transmitted as main traffic, where M=D+N. We assume that a packet loss occurs only at the wireless base station. We also assume that if the number of lost packets among the M packets is less than or equal to N, the original data block can be retrieved at the destination by FEC decoding and otherwise cannot be retrieved, resulting

Derivation of performance measures

In this section, we first consider the packet-loss probability of main traffic. We then derive the block-loss probability of a block, which consists of D original data packets and N redundant packets.

Numerical examples

In this section, we focus on video streaming, which is one of the most important real-time applications, and evaluate the recovery performance of FEC using the analytical results derived in the previous section. It is assumed that the transmission rate of streaming data is 4 Mb/s, and that the video frame rate is 30 frame/s. The size of a packet is 500 bytes. A frame has D=34 original data packets, and a block has the same number of packets as that of a frame. It is also assumed that the

Conclusions

In this paper, we considered the recovery performance of FEC for video streaming services over wired-wireless networks. We modeled a wireless base station as a GI+M/MSP/1/K queue. The packet- and block-level loss probabilities were analyzed through a continuous-time Markov chain. It was shown from numerical examples that the block-loss probability is greatly affected by the system capacity and the mean Bad-state period. It was observed that the recovery performance of FEC deteriorates according

Shun Muraoka received the B. Eng. degree from Kyoto University, Kyoto, Japan, in 2006. He received the degree of Master of Informatics from Graduate School of Informatics, Kyoto University, in 2008. He is currently with Nomura Research Institute, Ltd.

References (19)

  • I.F. Akyildiz et al.

    Wireless mesh networks: A survey

    Computer Networks

    (2005)
  • V. Srinivasan et al.

    Block cell loss reduction in ATM systems

    Computer Communications

    (1996)
  • G. Dán et al.

    On the effects of the packet size distribution on FEC performance

    Computer Networks

    (2006)
  • G. Carle et al.

    Survey on error recovery techniques for IP-based audio-visual multicast applications

    IEEE Network Magazine

    (1997)
  • C. Perkins et al.

    A survey of packet loss recovery techniques for streaming audio

    IEEE Network Magazine

    (1998)
  • N. Shacham, P. Pckenney, Packet recovery in high-speed networks using coding and buffer management, in: Proc. IEEE...
  • P.E. Green

    Fiber to the home: The next big broadband thing

    IEEE Communications Magazine

    (2004)
  • C. Cramer et al.

    Video quality and traffic QoS in learning-based subsampled and receiver-interpolated video sequences

    IEEE Journal on Selected Areas in Communications

    (2000)
  • C. Cramer et al.

    Low bit rate video compression with neural networks and temporal subsampling

    Proceedings of the IEEE

    (1996)
There are more references available in the full text version of this article.

Cited by (3)

Shun Muraoka received the B. Eng. degree from Kyoto University, Kyoto, Japan, in 2006. He received the degree of Master of Informatics from Graduate School of Informatics, Kyoto University, in 2008. He is currently with Nomura Research Institute, Ltd.

Hiroyuki Masuyama received the B. Eng. Degree from Faculty of Engineering, Kyoto University, Japan, in 1999. He received the M. I. and D. I. degrees from Graduate School of Informatics, Kyoto University, Kyoto, Japan, in 2001, and 2004, respectively. Since 2004, he has been an Assistant Professor at the Department of Systems Science, Graduate School of Informatics, Kyoto University. His current research interests include queueing theory, especially, asymptotic analysis of rare events. He is a member of the Operations Research Society of Japan (ORSJ). Dr. Masuyama received the Best Paper Award for Young Researchers from ORSJ in 2007.

Shoji Kasahara received the B. Eng., M. Eng., and Dr. Eng. degrees from Kyoto University, Kyoto, Japan, in 1989, 1991, and 1996, respectively. He was with the Educational Center for Information Processing, Kyoto University from 1993 to 1997. In 1996, he was a visiting scholar of University of North Carolina at Chapel Hill, NC, USA. From 1997 to 2005, he was with the Department of Information Systems, Graduate School of Information Science, Nara Institute of Science and Technology. Since 2005, he has been an Associate Professor of Department of Systems Science, Graduate School of Informatics, Kyoto University. His research interests include queueing theory and performance analysis of computer and communication systems. Dr. Kasahara is a member of the IEEE, the Institute of Electronics, Information and Communication Engineers (IEICE), the Operations Research Society of Japan, the Information Processing Society of Japan, and the Institute of Systems, Control and Information Engineers.

Yutaka Takahashi received the B. Eng., M. Eng., and Dr. Eng. degrees from Kyoto University, Kyoto, Japan. He was with the Department of Applied Mathematics and Physics, Faculty of Engineering, Kyoto University from1980 to 1995 and with the Department of Applied Systems Science at the same faculty from 1995 to 1996. From 1996 to 1999, he was a professor at the Graduate School of Information Science, Nara Institute of Science and Technology (NAIST), Nara, Japan. Since April 1999, he has been with the Department of Systems Science, the Graduate School of Informatics, Kyoto University, as a professor. From 1983 to 1984 he was with the Institut National de Recherche en Informatique et en Automatique (INRIA), France, as an Invited Professor. He was a Cochairman of IFIP TC6 WG6.3 on Performance of Communication Systems from 1992 to 2002 and is an Elected Full Member of IFIP TC6 WG7.3 on Computer Performance Evaluation as well as WG6.3. He is also an Associate Editor of Telecommunication Systems, an Area Editor of Mobile Networks & Applications, an Editor of Wireless Network Journal (WINET) and on the Editorial Board of Journal of Networks. He served as an International Advisory Committee Member of NIS (Networking and Information Systems) Journal and the project leader for the Kobe Multi-node Integrated Connection Research Center established by the Telecommunications Advancement Organization of Japan (TAO). He was awarded the Silver Core from IFIP in 2001 and a fellow from the Operations Research Society of Japan. His research interests include queueing theory and its application to performance analysis of computer communication systems as well as database systems. Dr. Takahashi is a member of the Institute of Electronics, Information and Communication Engineers, the Information Processing Society of Japan, the Operations Research Society of Japan, the Institute of Systems, Control, and Information Engineers, and the Japan Society for Industrial and Applied Mathematics.

View full text