Protecting Scalable Video Flows from Congestion Loss

Libæk, Bjørnar; Kure, Øivind

doi:10.1109/icns.2009.14

Cited by 4 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, as shown in [8], if a majority of the flows adds FEC to their packet stream 2 Application level packet loss after FEC recovery at the same time without lowering the video quality, there is a significant probability that the added overhead causes a growth in the channel packet loss ratio large enough to actually increase the effective packet loss at high traffic loads. It was also demonstrated that this probability diminishes as the relative amount of FEC overhead is reduced.…”

Section: Loss Protection: Shading the Probe Layermentioning

confidence: 97%

Generic Application Level Rate Control for Scalable Video Using Shadow Probing

Libæk

Kure

2009

2009 Fourth International Conference on Systems and Networks Communications

Self Cite

View full text Add to dashboard Cite

A generic application level rate control (ARC) for scalable video is presented. The ARC is designed to operate between a streaming video application (server-side) and a transport level rate control (TRC). TFRC is used as a reference TRC when evaluating the proposed mechanisms. When using scalable video such as H.264/SVC, incrementation of quality layers may cause significant increase in the sending rate. The ARC utilizes a technique we name shadow probing, which protects the established video layers from packet loss in these situations. Also, highly variable estimates of available bandwidth reported by the TRC are hidden from the application, in order to reduce the frequency of adding and dropping layers. Simulation results are presented showing that the mechanisms are able to improve perceived quality for both interactive and streaming video in networks with and without QoS support.

show abstract

Section: Loss Protection: Shading the Probe Layermentioning

confidence: 97%

Generic Application Level Rate Control for Scalable Video Using Shadow Probing

Libæk

Kure

2009

2009 Fourth International Conference on Systems and Networks Communications

Self Cite

View full text Add to dashboard Cite

show abstract

“…Much research has been conducted on new transport layer protocols for video streaming [6]- [12]. For instance, transport layer protocols proposed in [6]- [8] do not, relative to TCP, rapidly increase/decrease the congestion window size, yet remain TCP-friendly.…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, in [10]- [12], the authors propose modifications to TCP to maintain the data transfer rate requested by upper-layer applications. For instance, in [11], [12], the authors propose mechanisms to stabilize TCP throughput by concealing packet loss from TCP by installing a Forward Error Correction (FEC) layer in the lower part of a transport layer at both the sender and receiver. However, the redundancy of FEC causes these mechanisms to transmit excessive traffic to the network.…”

Section: Introductionmentioning

confidence: 99%

Network Friendly Transmission Control for Progressive Download over TCP

Hisamatsu¹,

Hasegawa²,

Murata³

2012

JCM

View full text Add to dashboard Cite

Video streaming services using Transmission Control Protocol (TCP) as a transport layerprotocol---represented by YouTube---are becoming increasingly popular and,accordingly, have come to account for asignificant portion of Internet traffic. TCP is greedy; that is, it tries to exhaust theentire bandwidth.Thus, video streaming over TCP tends to unnecessarilytake bandwidth from competing traffic. In this paper, we first investigate the data transfer mechanisms of the currentvideo streaming services using TCP and showthat they perform data transfer at much higher rates than the video playback rate.We then propose a new transfer mechanism for video streaming over TCP,one that controls the data transfer ratebased on the network congestion level and the amount of buffered video data at the receiver.Simulation results show that the proposed mechanism has two characteristics lacked bycurrent video streaming over TCP, specifically (1) a low frequency of buffer underflow at the receiverand (2) a lack of excessive bandwidth ``stealing'' from competing traffic.

show abstract

“…For instance, in [9,13], the authors propose mechanisms to stabilize TCP throughput by concealing packet loss from TCP. They do so by installing a Forward Error Correction (FEC) layer in the lower part of the transport layer at both the sender and receiver.…”

Section: Introductionmentioning

confidence: 99%