Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....boost in freedom for the design of adaptive streaming mechanisms. Another big issue for streaming in general is the potential impact of video tra#c on existing Internet tra#c. Many research projects have studied quality adaptive streaming in relationship with TCP friendly congestion control [36, 27, 4, 28, 9, 19, 31, 32, 20]. A common idea amongst them is to let the transport protocol and its congestion control dictate the appropriate sending rate. The main di#erences are in the details of deciding what to send and what to drop, and what information are used to inform these control decisions. For example, Rejaie et ....

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In INFOCOM (2), pages 737--746, 2000.

.... streaming media across networks has focused on the adaptability of applications to changes in network conditions [1] In these applications, the sender dynamically adjusts the delivery rate by controlling the quantization scales of media encoding, the frame rate, or the priority of the media parts [2, 3, 4]. Other schemes allow receivers to select from several possible playback qualities at the connection setup time [5] On the other hand, the receivers adopt the mechanism called bu#ering. This bu#ering process places data in receiver s bu#er, ensures the data synchronization for playback, and ....

D. Saparilla and K. W. Ross, "Optimal Streaming of Layered Video," In Proc. IEEE INFOCOM 2000.

....of QoS under the rate adaptive scaling yields useful expressions that can be used to help dimension networks and identify bottlenecks. Multi class admission control achieves the QoS benefits of optimal adaptation but requires accurate knowledge of system parameters. Related work includes [10], 7] 11] 12] 13] 14] 15] 8] 10] investigates optimal policies to dynamically adapt the fraction of the available bandwidth given to a base and enhancement layer. Their work differs from ours in that it takes is a client centric view while ours is a system centric view. Both [7] and ....

....useful expressions that can be used to help dimension networks and identify bottlenecks. Multi class admission control achieves the QoS benefits of optimal adaptation but requires accurate knowledge of system parameters. Related work includes [10] 7] 11] 12] 13] 14] 15] 8] [10] investigates optimal policies to dynamically adapt the fraction of the available bandwidth given to a base and enhancement layer. Their work differs from ours in that it takes is a client centric view while ours is a system centric view. Both [7] and [8] use an almost identical model for QoS as ....

Despina Saparilla and Keith Ross, "Optimal Streaming of Layered Video," in Proceedings of Infocom, 2000.

....O.C.M.B. Duarte is with the Network and Automation Group (GTA COPPE) Federal University of Rio de Janeiro, Brazil. Email: otto gta.ufrj.br. This work has been supported by CAPES, COFECUB, CNPq, UFRJ, UPMC, RNRT, and CNRS. if the number of receiver is large [8] [9], 10] 11] If all packets sent by the receivers arrive at the source, a feedback implosion occurs and the system su ers from a source collapse. Previous works have proposed some adaptive schemes to deal with the real time aspect of video distribution [1] 12] 13] 14] 15] 16] 17] ....

D. Saparilla and K. W. Ross, \Optimal streaming of layered video," in IEEE Infocom, (Tel-Aviv, Israel), Mar. 2000.

....QoS by a single parameter is not adequate because of the heterogeneous nature of the receivers. We argue that a multi criteria approach should be used in the case of multicast distribution of multi layered applications [5] A number of works have already dealt with QoS issues in such architectures [6, 7, 8, 9, 10, 11, 12, 13, 14]. The receiver driven layered multicast (RLM) 11] is a rate adaptive protocol for the distribution of layered video using multiple IPmulticast groups where the receivers subscribe to the number of layers they want. Nevertheless, they are limited to the layers the source decides to transmit. The ....

D. Saparilla and K. W. Ross, \Optimal streaming of layered video," in IEEE Infocom, (Tel-Aviv, Israel), Mar. 2000.

....the source is coded in a base layer and in one or more enhancement layers and the quality of the presentation depends on the number of layers that are played. Multi layered coding has been applied to improve the fairness of a wide variety of multicast applications, mainly for video distribution [1 10]. This approach is interesting in multicast communications because destinations with different receiving capacities (available bandwidth, local resources, etc) can choose the layers they are able to receive. Multi layered applications over multicast networks require that new mechanisms are ....

D. Saparilla and K. W. Ross, "Optimal streaming of layered video," in IEEE Infocorn, (Tel-Aviv, Israel), Mar. 2000.

....by monitoring available bandwidth and buffer for smoothing. A single constant bandwidth allocation is assumed for pre fetching in this paper. In the forward shifting technique, the fluid model is used for video stream. That is, the transmission of different layers is synchronous. Saparilla et al. [6] proposed some heuristics for finite length layered video streaming. However, the transmission of different layers is asynchronous and it is very challenging to choose the thresholds in using these heuristics. In asynchronous transmission of layers, the buffered low layers may not be used for ....

D. Saparilla and K. W. Ross, "Optimal Streaming of Layered Video", IEEE INFOCOM 2000.

....parameter is not adequate because of the heterogeneous nature of the receivers. In this paper, we argue that a multi criteria approach should be used in the case of multicast distribution of multi layered applications [5] A number of works have already dealt with QoS issues in such architectures [6], 7] 8] 9] 10] 11] 12] 13] 14] The receiver driven layered multicast (RLM) 11] is a rate adaptive protocol for the distribution of layered video using multiple IP multicast groups where the receivers subscribe to the number of layers they want. Nevertheless, they are limited to ....

D. Saparilla and K. W. Ross, "Optimal streaming of layered video," in IEEE Infocom, (Tel-Aviv, Israel), Mar. 2000.

....reduced cost. But such as the Internet was conceived, several limitations can be identified like the heterogeneity of the receivers and the scalability. New schemes must then be introduced to manage the transmission of the video streams in such a shared and not reliable environment [LI 99, AMO 00, SAP 00, RUB 99, SAR 00] In the video system proposed in [AMO 00] each multicast tree is composed of one source that communicates periodically with a number of receivers. The exchanged information allows the source to calculate the rates of the video layers. The corresponding procedures form an ....

SAPARILLA D., ROSS K. W., (( Optimal Streaming of Layered Video )), IEEE Infocom, Tel-Aviv, Israel, 2000.

....significant instability. There were some proposals [1] 4] on addressing this problem inside the network by assigning higher priority to lower layers and providing priority based dropping at routers in case of congestion. These approaches introduce some additional complexity at core routers. In [8] the available bandwidth is modeled as a stochastic process and optimal allocation of bandwidth between base and enhancement layers is studied. It was assumed that client buffer is unlimited. The work most relevant to ours was reported in [7] They assume that TCP friendly congestion control [6] ....

D. Saparilla, and K. Ross, "Optimal Streaming of Layered Video", Proc. IEEE INFOCOM'00, Tel-Aviv, Israel, Mar 2000.

....Source and receivers exchange control packets containing information about network states. Based on these packets, the source then adapts the rates of the video layers to current network conditions. Nevertheless, this approach is followed by scalability issues if the number of receiver is large [8, 9, 10, 11]. When all packets sent by the receivers arrive at the source, a feedback implosion occurs and the system su ers from a source collapse. Previous works have proposed some adaptive schemes to deal with the real time aspect of video distribution [1, 12, 13, 14, 15, 16, 17, 18] In [1, 14] a number ....

D. Saparilla and K. W. Ross, \Optimal streaming of layered video," in IEEE Infocom, (Tel-Aviv, Israel), Mar. 2000.

....the minimum delay. We argue that a multi criteria approach should also be used in the case of multicast distribution of multi layered applications [AMO 00] A number of works have already dealt with QoS issues in such architectures, but they were interested in solving other related problems [SAP 00, VIC 00, LI 99, SAR 00, RUB 99, SIL 00, MCC 96, LI 98, VIC 98, AMO 99] The receiver driven layered multicast (RLM) MCC 96] is a rate adaptive protocol for the distribution of layered video using multiple IP multicast groups where the receivers subscribe to the number of layers they want. ....

SAPARILLA D., ROSS K. W., "Optimal Streaming of Layered Video", IEEE Infocom, Tel-Aviv, Israel, 2000.

....client does not adapt at all, but imposes a scalable workload that is adapted at the server side to the current state. The amount of adaptation specific information that must be shared between video coder and protocols can be very small. Examples of these systems appear as layered video encoders [2] with selective packet dropping at the protocol level. Even without priority information we can use multiple description video coding [3] with random packet dropping. More specific information is required when trans coding is used. In forward scalable systems the generated workload must be ....

D. Saparilla & K.W. Ross, "Optimal streaming of layered video", Joint Conference of the IEEE Computer and Communications Societies (INFOCOM 2000) pp. 737-746 vol.2, 2000

No context found.

D. Saparilla and K. W. Ross, "Optimal Streaming of Layered Video," in Proc. of IEEE INFOCOM, Tel Aviv, Israel, March 2000, pp. 737--746.

No context found.

D. Saparilla and K. Ross. Optimal Streaming of Layered Video. Proceedings of IEEE Infocom 2000.

No context found.

D. Saparilla and K. W. Ross, "Optimal Streaming of Layered Video," in Proc. of IEEE INFOCOM, Tel Aviv, Israel, March 2000, pp. 737--746.

No context found.

D. Saparilla and K. W. Ross, "Optimal streaming of layered video," in Proceedings of IEEE INFOCOM 2000.

No context found.

D. Saparilla and K. Ross. Optimal Streaming of Layered Video. Proceedings of IEEE Infocom 2000.

.... also provides the server the opportunity to retransmit lost video packets before their decoding deadline [8, 13] Longer time scale bandwidth fluctuations, on the order of a few seconds to tens of seconds, can be addressed by using multiple versions of the same video [4] or a layeredencoded video [14]. When using a layered encoded video, the application needs adaptive control policies to decide which layers should be streamed at all times [11, 9, 15] in order to maximize the overall quality of the video rendered to the user. In this paper, we study the streaming of 2 layer FineGrained ....

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In Proc. of IEEE Infocom, pages 737--746, Tel Aviv, Israel, March 2000.

....research efforts (see [39, 50] for comprehensive surveys) Because of the best effort nature of the Internet, streaming video should adapt to the changing network conditions. One of the most popular technique for network adaptive streaming of stored video is using scalable video (see for instance [31, 34]) Video streaming applications should also adapt to the properties of the particular encoded video [32] Recently, rate distortion optimized streaming algorithms have been proposed (e.g. 5, 24] to minimize the end to end distortion of media, for transmission over the Internet. By analyzing ....

....tolerate an initial build up delay, during which some initial part of the video is prefetched into the client before the start of the playback. Maintaining a sufficient playback delay throughout the rendering allows the application to accommodate future bandwidth variations (see for instance [31, 34]) To account for bandwidth variability, we model bandwidth constraints and client buffering resources as follows. As shown on Figure 17, the video is partitioned into L allocation segments, with each allocation segment l containing the same number of frames [N LJ. While the server is streaming ....

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In Proc. of IEEE INFOCOM, pages 737-746, Tel Aviv, Israel, March 2000.

....any policy 1 3 S (22) A. Preliminary Results Having defined the loss probabilities for the streaming of finite length video, we now present some necessary preliminary results, which will aid in the derivation of the optimal streaming policies. For detailed proofs of these results see [24]. For these results, we again assume that C for some constant denote the average encoded rate of the base layer, i.e. Q . Similarly define . We consider the class of static policies and establish the following lemma, which parallels the results ....

.... be shown that E Q Applying Lemma 2 to the right hand side of the above two relationships yields E 9 Q 1 E Q 0 which implies that We now turn to the maximization problem in (23) 24) Using Theorem 2 we derive the following (see [24]) Corollary 1: The policy is optimal for , i.e. for any . The above corollary states that when the enhancement layer has an equal effect on the quality of the decoded video as the base layer, the optimal policy is a specific static policy, namely, the optimal policy ....

[Article contains additional citation context not shown here]

D. Saparilla and Ross K. W., "Optimal Streaming of Layered Video," http://www.eurecom.fr/ saparill, July 1999.

....stored video. In this paper we focus on the remaining two techniques, i.e. dding dropping layers and switching versions . Adding dropping layers has been proposed by many researchers as an effective scheme that provides both rate adaptation and error resilience in video communications [5, 3, 1, 9]. In the adding dropping layers scheme, the video stream is partitioned into several layers. It is composed of a base layer, which contains the most essential information for the reconstruction of the video, and results in low but generally acceptable quality, and one or several enhancement layers ....

....Control Policies In this section we develop control policies for adding dropping layers and for switching among versions. A control policy determines when to add and drop the enhancement layer, as well as how to allocate bandwidth among the layers (when both layers are being streamed) Results in [9] have shown that control policies based on the content of the prefetch buffers at the client attain good performance in a TCP friendly context. Some simple threshold control policies were introduced in [10] for layered encoded videos. In this paper we design equivalent control policies for ....

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In Proceedings o] IEEE In]ocom, Tel Aviv, Israel, March 2000.

....fundamental understanding of the impact of the two key resources (cache space and link bandwidth) The streaming of layered encoded video (without caching at proxies) has been studied in a variety of contexts. Optimal streaming strategies for the encoding layers are proposed in [30] 31] 32] [33]. Several studies have investigated the streaming of layered encoded video in the context of multicast distribution [34] 35] 36] 37] 38] Several studies have investigated the caching of single layer encoded video. Wang et al. 39] propose a video staging scheme where the part of the ....

D. Saparilla and K. W. Ross, "Optimal streaming of layered video," in Proceedings of IEEE INFOCOM 2000.

....P e maxfP b ; Q e g: 22) A. Preliminary Results Having defined the loss probabilities for the streaming of finite length video, we now present some necessary preliminary results, which will aid in the derivation of the optimal streaming policies. For detailed proofs of these results see [24]. For these results, we again assume that r b (t) Kr e (t) for some constant K. Let r b denote the average encoded rate of the base layer, i.e. r b = 1 T R T 0 r b (t) dt. Similarly define r e . We consider the class of static policies and establish the following lemma, which parallels ....

.... e g 0 X(t)dt : Applying Lemma 2 to the right hand side of the above two relationships yields E 1 T Z T c 0 X(t) dt E 1 T Z maxfT b ;T e g 0 X(t)dt ; which implies that F ff F We now turn to the maximization problem in (23) 24) Using Theorem 2 we derive the following (see [24]) Corollary 1: The policy ff is optimal for J , i.e. J ff J for any when de d b re r b . The above corollary states that when the enhancement layer has an equal effect on the quality of the decoded video as the base layer, the optimal policy is a specific static policy, namely, ....

[Article contains additional citation context not shown here]

D. Saparilla and Ross K. W., "Optimal Streaming of Layered Video," http://www.eurecom.fr/ saparill, July 1999.

....encoded into two layers, a base layer and an enhancement layer. In our earlier work we developed a model and theory for the problem of dynamically allocating fair share bandwidth among the two layers in order to minimize the impact of client buffer starvation on the quality of the decoded stream [7]. In this paper we evaluate the performance of two classes of interlayer bandwidth allocation schemes in simulations based on fair share bandwidth traces. We then propose a scheme that streams the enhancement layer only when we are confident that the base layer buffer at the client will henceforth ....

....at time t can be expressed as L b (t) r b Gamma b (t)X(t) 1(Y b (t) 0) Loss of enhancement layer data occurs when there is loss of baselayer data, or when the enhancement layer buffer at the client is starved. Multiple models can be used for determining loss in the enhancement layer. [7]. In the numerical work that follows, we use a model in which all enhancement layer data that reach the receiver buffer can be decoded as long as the corresponding base layer data is also available at the receiver. The loss rate in the enhancement layer at time t can be expressed as L e (t) ....

D. Saparilla and K. W. Ross, "Optimal Streaming of Layered Video," in Proc. of IEEE Infocom, Tel Aviv, Israel, March 2000.

.... maxfP b ; Q e g: 22) A. Preliminary Results Having defined the loss probabilities for the streaming of finite length video, we now present some necessary preliminary results, which will aid in the derivation of the optimal streaming policies. For detailed proofs of these results see [24]. For these results, we again assume that r b (t) Kr e (t) for some constant K. Let r b denote the average encoded rate of the base layer, i.e. r b = 1 T R T 0 r b (t) dt. Similarly define r e . We consider the class of static policies and establish the following lemma, which ....

.... Lemma 2 to the right hand side of the above two relationships yields E 1 T Z Tc 0 X(t) dt E 1 T Z maxfT b ;T e g 0 X(t)dt ; which implies that F F We now turn to the maximization problem in (23) 24) Using Theorem 2 we derive the following (see [24]) Corollary 1: The policy is optimal for J , i.e. J J for any when de d b re r b . The above corollary states that when the enhancement layer has an equal effect on the quality of the decoded video as the base layer, the optimal policy is a specific static policy, ....

[Article contains additional citation context not shown here]

D. Saparilla and Ross K. W., "Optimal Streaming of Layered Video," http://www.eurecom.fr/ saparill, July 1999.

.... e maxfP b ; Q e g: 22) A. Preliminary Results Having defined the loss probabilities for the streaming of finite length video, we now present some necessary preliminary results, which will aid in the derivation of the optimal streaming policies. For detailed proofs of these results see [23]. For these results, we again assume that r b (t) Kr e (t) for some constant K. Let r b denote the average encoded rate of the base layer, i.e. r b = 1 T R T 0 r b (t) dt. Similarly define r e . We consider the class of static policies and establish the following lemma, which ....

.... : Applying Lemma 2 to the right hand side of the above two relationships yields E 1 T Z Tc 0 X(t) dt E 1 T Z maxfT b ;T e g 0 X(t)dt ; which implies that F ff F We now turn to the maximization problem in (23) 24) Using Theorem 2 we derive the following (see [23]) Corollary 1: The policy ff is optimal for J , i.e. J ff J for any when de d b re r b . The above corollary states that when the enhancement layer has an equal effect on the quality of the decoded video as the base layer, the optimal policy is a specific static policy, ....

[Article contains additional citation context not shown here]

D. Saparilla and Ross K. W., "Optimal Streaming of Layered Video," http://www.eurecom.fr/ saparill, July 1999.

No context found.

Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

Saparilla D. and Ross K.W., Optimal Streaming of Layered Video, in Proceedings of the IEEE INFOCOM Conference, March 2000. submitted paper.

No context found.

Saparilla D, Ross KW (2000) Optimal streaming of layered video. IEEE INFOCOM

No context found.

D. Saparilla and K. W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

D. Saparalla and K. Ross, "Optimal streaming of layered video," in Proceedings of Infocom, 2000.

No context found.

Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

No context found.

Saparilla D. and Ross K.W. Optimal Streaming of Layered Video. In Proceedings of the IEEE INFOCOM, pages 737--746, 2000.

No context found.

Despina Saparilla and Keith W. Ross. Optimal Streaming of Layered Video. In Proceedings of the Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies 2000.

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