DASH and HTTP2.0
Download as PPT, PDF1 like1,089 views
The document discusses the tradeoff between latency and server load in DASH live streaming, highlighting how large segment sizes can hinder real-time experiences, while smaller sizes increase server requests. It introduces a solution using HTTP/2.0's push feature, allowing clients to define segment requests without explicit requests, and discusses various DASH push header options. Alternatives such as playing segments while receiving them are also mentioned, which may reduce client latency but not server-induced latency.
DASH and HTTP2.0
- 2. 2 Overview DASH Live has a latency / server load tradeoff: Large segment size create latency which hinders the real-time live experience Shorter segment size requires more segment requests so it loads the server HTTP2.0 has push option in which the server pushes content to the client without the client EXPLICIT request
- 3. 3 Solution Upon each request, the client will be able to define the number of segments that will be pushed by the server without an Explicit request Implementation using specific DASH push header in HTTP2.0
- 4. 4 HTTP2.0 push header options “DASH-PUSH: type=push-next; K=5” – for segments “DASH-PUSH: type=push-time; T=10” – for seconds “DASH-PUSH: type=uri-list; seg1.mp4;seg2.mp4…” – for segment names DASH-PUSH: type=uri-template; seg{1}.mp4; 1:2:3” – for template segment names DASH-PUSH: type=uri-wildcard; text/css” – for all resources of type text/css
- 5. 5 Other solutions An alternative solution would be to play a segment while still receiving it. This solution would decrease some of the latency on the client but would not lower the latency created by the serve while it creates the segment