Content Delivery Platform Fragment Indexing Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current on-demand video streaming technologies face delays in content playback due to the need for fragment-level metadata transmissions and the HTTP round trip for manifest files, which increase network traffic and processing load, and result in longer 'time to first frame' for content delivery.
Innovation Solution
A content delivery platform that eliminates fragment-level metadata transmissions by allowing viewing devices to request and retrieve content fragments based on provided indexes, using a content packager to package media assets into fragments and an intermediary device to manage metadata, enabling pre-fetching and post-fetching strategies to reduce latency and improve response times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fragment-level metadata transmissions are used for content delivery, then content fragments can be properly identified and delivered to viewing devices, but network traffic increases and processing load increases
Solution Approach 1:
The patent extracts and eliminates fragment-level metadata transmissions from the content delivery process. Instead of transmitting metadata for each fragment, the system uses a simplified manifest file that provides only essential information (starting index and file name), allowing viewing devices to calculate fragment indexes independently. This extraction removes the harmful metadata transmission component while preserving the ability to deliver content fragments accurately.
Solution Approach 2:
The patent implements a virtual copy mechanism where the manifest file contains a reference (starting index) to the first fragment, and viewing devices create virtual copies of the manifest for subsequent fragments by incrementing the index. This allows the system to provide fragment identification information without physically transmitting separate metadata files for each fragment, thereby reducing network traffic while maintaining delivery accuracy.
2Reliability
If fragment-level metadata transmissions are used for content delivery, then content fragments can be properly identified and delivered to viewing devices, but device complexity increases due to parsing requirements
Solution Approach 1:
The patent extracts and removes the complex fragment-level metadata parsing requirement from the viewing device. By eliminating separate metadata files for each fragment and using a simplified manifest with only starting index and file_name, the system reduces the parsing complexity to minimal operations while maintaining the ability to identify and deliver content fragments accurately.
3Reliability
If content buffering is implemented to reduce playback issues, then connection reliability is improved, but time to first frame increases due to buffer creation delay
Solution Approach 1:
The patent applies preliminary action by enabling viewing devices to pre-fetch subsequent content fragments before they are officially available. The system allows devices to calculate and request fragment indexes in advance based on the starting index provided in the manifest, so that when fragments become available, they are already in the process of being transmitted or have arrived, eliminating the traditional buffering delay while maintaining playback stability.
Solution Approach 2:
The patent segments the content delivery process into independent fragment requests based on calculated indexes. Instead of requiring a complete buffer to be filled before playback, the system allows sequential fragment retrieval with each fragment being independently requested and delivered. This segmentation enables playback to begin as soon as the first fragment is received, reducing time to first frame while maintaining reliability through continuous fragment delivery.
Data Source
AI summary
A non-transitory computer-readable medium may comprise instructions which, when executed by one or more hardware processors, may cause execution of operations. A viewing device may receive fragment-level metadata associated with a first fragment of content. A request for a second fragment of the content immediately subsequent to the first fragment may be received from the viewing device. It may be determined that the second fragment is not stored at a content storage device but is expected to be stored within a defined time period. The request may be held for the defined time period. Prior to expiration of the defined time period, it may be determined that the second fragment is stored at the content storage device. The second fragment may be transmitted to the viewing device.