This report introduces some basic issues regarding the transmission of variable bit rate (VBR) video in asynchronous transfer mode (ATM) networks with concentration on VBR video models and statistical characterises. Performance modeling for VBR video traff
1 Introduction Variable bit rate (VBR) video is predicted to be a major source of tra c for the broadband integrated service digital network (B-ISDN) which is based on asynchronous transfer mode (ATM). The hehavior of ATM networks carrying VBR video tra c has been of interest for a number of years, and has therefore received much attention in the literature. VBR encoders have the potential to provide a constant quality video signal if values of the encoding parameters remain constant or if there is no loss of information due to packet losses. However, VBR video tra c also imposes very stringent real-time constraints on the networks, especially in the case of interactive services. From the network point of view, the inherently complex stream structures of VBR tra c makes the evaluation of the performance of ATM networks di cult under the apperance of VBR services. Thus, analytically tractable models for VBR tra c, which takes into account the tra c charateristics from realistic services that would be carried in an ATM network, and, in particular, which could accurately represent the statistical nature of very bursty, high-speed video services, are required. Statistical characterization and modeling of the output process of VBR video codecs are fundamental steps in understanding how VBR coding a ects network performance. Modeling of VBR coded video tra c deals with the development of a mathematical model by monitoring and studying the output of VBR coded sources. A good model provides a tool to accurately characterize video sources and enables a theoretical assessment of the network performance as well. Two classes of model are needed for the VBR tra c: single source models and multiplexed source models. Single source models could be used to study what types of tra c descriptors make sense for parameter negotiation with the network at the call setup procedure, as well as to test rate control algorithms and predicte the quality of service (QoS). Multiplexed models, in contrast, capture the characteristics of statistically multiplexed bursty
seources and are used to evaluate the aggregated e ect of the superposition of busty streams on network instead of individual source. Video modeling, however, presents some di culties due to the intrinsic complexity and 1
variable-bit-rate (VBR) MPEG coded video sources...traffic characteristic parameters are not exceeded. ...Such statistical models has been proposed in [2,...
multi-layer variable bit rate (VBR) video traf?...characteristic useful in computing precise bounds on...analysis and statistical properties of wavelet coef...
(as in a statistical multiplexer) does not ...Willinger, "Long-Range Dependence in Variable-Bit...models that capture this characteristic takes many ...
characteristic of the Internet into account and is...(variable-bit-rate) video supports better quality...On one hand, statistical models demand a lower ...
over network, especially over ATM, is emerging. ...This characteristic should be taken into ...Statistical Models in Engineering. John Wiley & ...
loading the network and accurate models are needed...[6] where the selfsimilar characteristic of this ...of Variable Bit Rate Video Multiplexer”, Proc. ...
characteristic information of the stored video ...is the network which is usually ATM these days....Since the signal is variable bit rate, the ...
(LRD), a characteristic that the classical, ...Robbins, “Performance models of statistical ...of a Variable Bit Rate Video Multiplexer”, Proc...
characterize general variable bit rate (VBR) ...The burstiness characteristic comes from these di ...\Statistical properties of MPEG video tra c and ...
characterize general variable bit rate (VBR) ...The burstiness characteristic comes from these di ...\Statistical properties of MPEG video tra c and ...
热门文档