P. Pancha and M. El Zarki. Leaky Bucket Access Control for VBR MPEG Video. In Proc. IEEE Infocom, pages 796--803, April 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....in the end systems, though our adaptation algorithm will be more complicated than the work presented in [16] Previous research in the area of flow shaping and control in the source end system has led to numerous noteworthy results in the past few years. In the work at AT T Bell Laboratories [19], for example, has examined the performance of the well known Leaky Bucket control scheme for both intraframe and mixed intra interframe variable bit rate (VBR) MPEG video sources. The work investigates the effect of varying the leaky bucket parameters, and presents a new technique for VBR video ....

P. Pancha and M. Zarki. Leaky Bucket Access Control for VBRMPEGVideo. IEEE INFOCOM '95, 2, 1995.

....a leaky bucket controlled channel. Heeke [15] and Coelho [4] aim to make the output behave like a predefined Markov chain. Pickering and Arnold [31] propose a rate control algorithm that produces a variable bitrate (VBR) traffic lying between predefined upper and lower bounds. Pancha and El Zarki [30] also advocate using rate control with an algorithm similar to that of [33] where a few frames long bucket is used to control traffic variability. The present contribution belongs to the second category. We pretend that a network cannot efficiently provide performance guarantees if video coder ....

P. Pancha and M. El Zarki, "Leaky bucket access control for VBR MPEG video," in Proc. IEEE INFOCOM'95, Boston, MA, Apr. 1995.

.... on the integrated services Internet framework can be found in [21] The more traditional view that the token bucket rate r is the average rate of the source can be encountered in the ATM traffic management specifications, where it is called sustainable cell rate [1] In some earlier works (like [11, 13]) the token bucket rate was not separated from the rate that is reserved for a flow, i.e. the token bucket was not used as a traffic descriptor or shaper, but as the resource at the User to Network interface that is granted to a flow. Those works studied the relation between the token bucket ....

P. Pancha and M. E. Zarki, "Leaky bucket access control for VBR MPEG video," in Proceedings IEEE INFOCOM, pp. 796--803, 1995.

....in the end systems, though our adaptation algorithm will be more complicated than the work presented in [16] Previous research in the area of flow shaping and control in the source end system has led to 10 numerous noteworthy results in the past few years. In the work at AT T Bell Laboratories [19], for example, has examined the performance of the well known Leaky Bucket control scheme for both intraframe and mixed intra interframe variable bit rate (VBR) MPEG video sources. The work investigates the effect of varying the leaky bucket parameters, and presents a new technique for VBR video ....

P. Pancha and M. Zarki. Leaky Bucket Access Control for VBRMPEGVideo. IEEE INFOCOM '95, 2, 1995.

....channel. Heeke [14] and Coelho [15] proposed to make the encoder output behave like a predefined Markov chain. Pickering and Arnold [16] proposed a rate control algorithm that produces a VBR stream which rate varies between predefined upper and lower bounds. Pancha and El Zarki [17] also proposed a similar rate control algorithm as [13] but using a bucket that is a few frames long to control traffic variability. However, source rate control at the encoder is not always practical or desirable, e.g. in the cases of preencoded video archive, or video traffic originating from ....

P. Pancha and M. El Zarki, "Leaky bucket access control for VBR MPEG video," Proc. IEEE INFOCOM'95, Boston, MA, Apr. 1995.

....considered deterministic traffic characterizations for VBR video traffic using the (oe; ae) traffic model that corresponds to the leaky bucket policing mechanism. Most studies use only a single (oe; ae) pair and explore the dependencies between the burstiness parameter oe and the rate parameter ae [77, 80, 87, 88, 90]. In particular, for any fixed choice of rate ae, the burst parameter oe should be selected as small as possible, that is [77] oe = inffoe j oe aet A[ t] 8t; 0g (4.1) 4.1. Related Work 48 Equation (4.1) illustrates a tradeoff between buffer space (i.e. burst) and bandwidth (i.e. ....

....parameters should be selected. Note that it is computationally demanding to determine this candidate set of (oe; ae) pairs. Many schemes select parameters oe and ae according to either network resource availability or the relative importance of bandwidth and buffer space. Pancha and El Zarki [80] choose parameters by fixing the burstiness parameter oe according to available buffer space, while the choice of (oe; ae) in [18] depends on the relative availability of unallocated bandwidth and buffer space. An approach discussed by Guillemin et al. in [46] assigns relative importance ....

[Article contains additional citation context not shown here]

P. Pancha and M. El Zarki. Leaky Bucket Access Control for VBR MPEG Video. In Proc. IEEE Infocom, pages 796--803, April 1995.

....at an acceptable level. In adaptive coding, the source traffic characteristics is controlled to stay within the allowed range defined by the access control mechanism [5] The result of this technique is very similar to CBR coding and it can lead to variations in image quality in a sequence. In [6] a leaky bucket controller is applied to shape the VBR video signal in order to control cell loss rate. Some studies have used TES modeled video to investigate the performance of multiplexed video signals [7] In [8] a frame spreading strategy is used to reduce the variability of bit rate. In ....

P. Pancha and M. El Zarki,"Leaky bucket access control for VBR MPEG video," IEEE INFOCOM'95, pp. 796-803

....for VBR video characterizations. Note, however, that most methods proposed in the literature do not strive for high network utilization as a design goal, and, instead, select the traffic parameters according to network resource availability. Also, most studies use only a single leaky bucket [23, 25, 28, 29, 31]. Pancha and El Zarki [25] choose parameters by fixing the burstiness parameter of a leaky bucket according to available buffer space. In [3] the choice of the traffic parameters depends on the relative availability of unallocated bandwidth and buffer space. An approach discussed by Guillemin et. ....

....however, that most methods proposed in the literature do not strive for high network utilization as a design goal, and, instead, select the traffic parameters according to network resource availability. Also, most studies use only a single leaky bucket [23, 25, 28, 29, 31] Pancha and El Zarki [25] choose parameters by fixing the burstiness parameter of a leaky bucket according to available buffer space. In [3] the choice of the traffic parameters depends on the relative availability of unallocated bandwidth and buffer space. An approach discussed by Guillemin et al. in [15] assigns ....

[Article contains additional citation context not shown here]

P. Pancha and M. El Zarki. Leaky Bucket Access Control for VBR MPEG Video. In Proc. IEEE Infocom, pages 796--803, April 1995.

....trace. The prediction errors resemble white noise, or at most short memory, but the marginals have a heavy tail. Using fixed service rate, for highly correlated input traffic with a heavy tail, if not served at a rate close to the peak, causes large queues, large delays, and excessive cell loss [1, 13, 17]. In contrast, by reserving bandwidth at least equal to the predicted value, only the errors in the prediction needs to be buffered. Because the errors resemble white noise or at most short memory, small buffers, high utilization and small delays can be achieved. Results show that the buffer size ....

....To compare queuing and delay performance of fixed and dynamic bandwidth allocation. Potential applications include: 1) Predicting and changing leaky bucket parameters. For MPEG streams, the predictions are found to better for individuals frames (I,P, and B) so multiple leaky buckets suggested in [17] in conjunction with forecasting may be used. 2) Dynamic reservation in a multi access channels, e.g. in Metropolitan area networks. The rest of the paper is organized as follows. Section 2 describes the non adaptive mean square error linear predictors. Section 3 describes the least mean square ....

[Article contains additional citation context not shown here]

Pancha P., and M. El Zarki, " Leaky bucket access control for VBR MPEG video," Proc. IEEE INFOCOM 1995, pp 796-803, Boston, April 1995.

....considered deterministic traffic characterizations for VBR video traffic using the (oe; ae) traffic model that corresponds to the leaky bucket policing mechanism. Most studies use only a single (oe; ae) pair and explore the dependencies between the burstiness parameter oe and the rate parameter ae [22, 24, 27, 28, 30]. In particular, for any fixed choice of rate ae, the burst parameter oe should be selected as small as possible, that is [22] oe = inffoe j oe aet A[ t] 8t; 0g (5) Equation (5) illustrates a tradeoff between buffer space (i.e. burst) and bandwidth (i.e. rate) when selecting ....

....parameters should be selected. Note that it is computationally demanding to determine this candidate set of (oe; ae) pairs. Many schemes select parameters oe and ae according to either network resource availability or the relative importance of bandwidth and buffer space. Pancha and El Zarki [24] choose parameters by fixing the burstiness parameter oe according to available buffer space, while the choice of (oe; ae) in [3] depends on the relative availability of unallocated bandwidth and buffer space. An approach discussed by Guillemin et al. in [14] assigns relative importance ....

[Article contains additional citation context not shown here]

P. Pancha and M. El Zarki. Leaky Bucket Access Control for VBR MPEG Video. In Proc. IEEE Infocom, pages 796--803, April 1995.

....not be limited to only Video On Demand applications and the delivery of Variable Bit Rate (VBR) video streams. Previous research in the area of flow shaping and control in the source end system has led to numerous noteworthy results in the past few years. The research at AT T Bell Laboratories [13], for example, has examined the performance of the well known Leaky Bucket control scheme for both intraframe and mixed intra interframe VBR MPEG video sources. The work investigates the effect of varying the leaky bucket parameters, and presents a new technique for VBR video transmission using ....

P. Pancha and M. Zarki. Leaky Bucket Access Control for VBR MPEG Video. IEEE INFOCOM '95, 2, 1995.

....IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 15, NO. 6, AUGUST 1997 The first approach introduces buffering delay. For real time interactive video services, the smoothing period needs to be small as the maximum tolerable end to end delay can be at most 10 15 frames [8] A previous study [9] has shown that for uncontrolled VBR video traffic, the first approach does not yield any substantial gain when the smoothing period is limited to be less than 15 frames. In this paper, we investigate the second approach in more detail. Generally, source rate control will cause picture quality ....

P. Pancha and M. El Zarki, "Leaky bucket access control for VBR MPEG video," in Proc. IEEE INFOCOM'95, Boston, MA, Apr. 1995, pp. 796--803.

No context found.

P. Pancha and M. El Zarki. Leaky Bucket Access Control for VBR MPEG Video. In Proc. IEEE Infocom, pages 796--803, April 1995.

No context found.

Pancha P., and M. El Zarki, "Leaky bucket access control for VBR MPEG video," Proc. IEEE INFOCOM, pp 796-803, Boston, April 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC