This paper presents a source model for variable bit rate (VBR) video traffic. A finite state Markov chain is shown to accurately model one- and two-layer video of all activity levels on a per source basis. Our model captures the source dynamics, including
IEEE/ACM Trans. Networking v7, (3), p398-413, June 1999
1
Modeling One- and Two-Layer Variable Bit Rate Video
Kavitha Chandra1andAmy R.Reibman2
CenterforAdvanced ComputationandTelecommunications1
Department of Electrical and Computer Engineering
University of Massachusetts Lowell
Lowell, MA 01854Kavitha_Chandra@uml.edu
AT&TLabs Research2
100 Schultz DriveRed Bank, NJ 07701amy@research.att.com
ABSTRACT
This paper presents a source model for variable bit rate (VBR) video traf c. A nite state Markov chain is shown toaccurately model one- and two-layer video of all activity levels on a per source basis. Our model captures the sourcedynamics, including the short-term correlations essential for studying network performance. The modeling tech-nique is shown to be applicable for both H.261 and MPEG2 encoded video of a variety of activity levels. The traf cmodel is shown in a simulation study to be able to accurately characterize both the single-source buffer occupancyoverawide range of buffer sizes and the multiplexing behavior.The VBR video model is also used to model the en-hancement layer of two-layer SNR scalable video. Weshowthat two-layer encoding has signi cantly better statisti-cal multiplexing gains than one-layer video, particularly when the network admits calls based on a leaky-bucketcharacterization.
1. INTRODUCTION
Video communication services such as entertainment based video on demand, home shopping, videoconfer-encing, telemedicine and distance learning are emerging applications for broadband integrated services digital net-works (B-ISDN). Analog video signals are digitized, compressed and encoded prior to transmission overcommuni-cation networks. Current compression techniques for digital video can achieve video bit rates of acceptable qualityin the range of 1-5 Mbps. In a network carrying voice, video and data, the video traf c component can consume thedominant portion of the available bandwidth. In order to evaluate the performance of B-ISDN networks, traf c mod-els of proposed services are necessary.This work addresses the problem of developing a model for variable bit ratevideo traf c, which we then apply to both one- and two-layer video.
Video can be transported either with a constant bit-rate (CBR) or with a variable bit-rate (VBR). VBR videohas several potential advantages overtraditional CBR video: improvedimage quality and shorter delay.Inaddition,through statistical multiplexing, improvedchannel allocation may be obtained compared to CBR transport. However,statistical multiplexing invariably results in buffering delays and losses, which can signi cantly degrade video qual-ity.Tominimize the amount of delay and loss, the net-working community has focused on the development of effec-tive and implementable congestion control schemes, including connection admission control (CAC) and usageparameter control (UPC). Tominimizethe impact of delay and loss, the video community has focused on developinggood error concealment algorithms and designing ef cient two-layer coding algorithms [1,2] for use in combinationwith the dual-priority transport provided by ATM networks. For example, while one-layer MPEG-2 [3] producesgenerally unacceptable video quality with a cell loss ratio of 10 ?3 , losses at this rate with SNR scalability (one ofthe four standardized layered coding algorithms of MPEG-2) are generally invisible, eventoexperienced viewers[4].
To determine a set of techniques appropriate for CACand UPC, traf c models that accurately represent thestatistical nature of very high-speed bursty services are necessary.Avariety of application and coding-speci c mod-els of one-layer VBR video have been proposed in the literature. Existing models can be categorized into models for
June29.1999 32页 免费 roud - Issue no 49, June... 暂无评价 7页 免费 LegCo Panel on Transport... 暂无评价 11页 免费 IEEEACM Trans. Networkin.....
TO APPEAR, IEEEACM TRANS... 暂无评价 14页 免费 TO APPEAR IN IEEEACM .... TO APPEAR IN IEEEACM TRANSACTIONS ON NETWORKING 1 Robust Header Compression ...
IEEEACM TRANS. NETWORKIN... 暂无评价 21页 ...ON NETWORKING 1 End-to-end differentiation of ...one wireless link in the path between the sender...
TO APPEAR, IEEEACM TRANS... 暂无评价 14页 ... IEEEACM TRANSACTIONS ON NETWORKING 1 Inducing ...June 24, 2006, November 9, 2006 and March 25...
TO APPEAR, IEEEACM TRANS... 暂无评价 14页 ... IEEEACM TRANSACTIONS ON NETWORKING 1 Tracetree A ...There are two main approaches to discovering ...
热门文档