Low Latency Streaming Media via Playlist Annotation and HTTP/2
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing streaming media content delivery systems face challenges in delivering live events with low latency, resulting in significant delays, often exceeding 15 seconds, which is disappointing for users.
Innovation Solution
The proposed solution involves updated playlist requests from client devices that include requests for future media segments, allowing servers to block responses until the segments are available, and using HTTP/2 for efficient transmission, enabling simultaneous requests for updated playlists and media segments to reduce latency. This approach maintains socket connections open for continuous data transfer and employs techniques like concurrent media segment requests and rendition switching to minimize delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional HTTP compliant requests are used to retrieve media segments from servers, then the system can deliver streaming media content, but the latency relative to live events becomes large (15 seconds or more)
Solution Approach 1:
The client device sends a playlist request that includes a request for a future media segment before that segment is available. The server blocks the response until the future segment becomes available, then pushes it to the client. This preliminary action eliminates the need for separate requests and reduces latency.
Solution Approach 2:
The system combines the playlist request and the future media segment request into a single HTTP request. This merging of requests eliminates round trip time between separate requests and allows the server to push both the updated playlist and the future media segment in one response.
2Loss of time
If the server blocks responses until future media segments are available, then latency is reduced, but the complexity of the transfer protocol increases
Solution Approach 1:
The HTTP/2 protocol is used to provide multiple functions within a single request framework. It enables both playlist retrieval and future media segment requests to be combined in one request, and supports server push functionality, reducing the need for separate protocols while maintaining low latency.
3Reliability
If separate requests are made for updated playlists and media segments, then the system can handle available content, but round trip time increases latency
Solution Approach 1:
The client device combines the playlist request and the future media segment request into a single HTTP request. This merging eliminates the round trip time between separate requests while maintaining reliable content delivery through the server's blocking and push mechanisms.
Data Source
AI summary
Content streaming systems, such as systems that use HTTP compliant requests to obtain media segments for presentation of the content on a device. These content streaming systems can be optimized to reduce latency to a low level so that live events can be streamed to receiving devices in such a manner so that the time between an action in the live event and the presentation of the action on a receiving device that receives the streamed content is less than about 10 seconds. A client device can use rendition reports to tune-in to a new rendition (at a first bit rate) after presenting a prior rendition (of a second bit rate) when switching between the different bit rates; also, for example, a client device can use playlist annotations that indicate independent frames to avoid downloading depending frame media segments when switching between different renditions.