TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
multitude of streams. It can switch from one stream to another very easy. Here, the client is a<br />
simple single layer decoder.<br />
3.2.1.6 Real-Time Single Stage Encoding<br />
Real-time coding means to encode the video stream right before the transmission process. The<br />
client connects to the servers and specifies the encoding parameters. This result in a high<br />
complexity at the server and rate adaptation in the network is not possible.<br />
3.2.1.7 Transcoding<br />
One of the methods to modify the media bit stream is the transcoding. The transcoding aims<br />
to transform of one media format to another media format. For example transcoding from MPEG-<br />
2 format to H.264 format. This allows bit rate reduction, frame rate reduction, and temporal and<br />
special downsampling [64]. The transcoding is applied by decoding the video to raw format<br />
(uncompressed video) and then re-encode the video to the target format. There are mainly two<br />
drawbacks when using this approach. First, the quality of the result format is generally lower than<br />
the original format. Second, media transcoding generally requires extensive computation and large<br />
storage spaces, which makes this approach very expensive. The advantage in transcoding is that it<br />
allows media adaptation both at the server and in the network. However, it can be efficient using a<br />
dedicated media gateway [65] [66].<br />
3.2.1.8 <strong>Streaming</strong> over TCP<br />
The Transmission Control Protocol (TCP) is the most widely used protocol in the current <strong>IP</strong><br />
networks. It can be used for packet video streaming. However, it has been noted that TCP is not<br />
adapted for real time traffic for theses reasons. First, TCP guarantees delivery via retransmissions,<br />
but because of the retransmissions its delay is unbounded and not suitable for real-time traffic.<br />
Second, TCP provides flow control and congestion control that make packet video streaming<br />
unpractical. Finally, TCP requires a back channel for packets acknowledgements. All theses reasons,<br />
led us to choose UDP for media streaming applications<br />
3.2.1.9 <strong>Streaming</strong> over UDP<br />
The User Datagram Protocol (UDP) is simply a user interface to <strong>IP</strong>. It is therefore unreliable<br />
and connectionless. UDP does not guarantee delivery of packet video applications, but for those<br />
delivered packets their delay is more predictable. UDP does not provide any flow control or<br />
congestion control. If the video application needs theses techniques, it should implement its own<br />
one like TCP-Friendly congestion control.<br />
3.2.1.10 <strong>Streaming</strong> over Rate Controlled UDP: TCP-Friendly<br />
In order to avoid network congestion, in which the transmitted traffic exceeds the designed<br />
limit, it is necessary to implement functions that perform congestion control such as rate control.<br />
The congestion is a phenomenon that has symptoms on packet losses, transmission delay, and<br />
delay jitter. As we have discussed in Section 3.1.6, such symptoms represent significant challenges<br />
for media streaming systems. The congestion control can be performed at many levels, by the<br />
network (e.g. edge router policy, source quench, DEC-bit, RED) or by the end-hosts. In<br />
48