Predictive Manifest Generation for Recorded Content Playback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Managing and reconstituting a large volume of stored video content for future access in cloud-based DVR services is challenging due to the extensive computational resources required for maintaining manifests for recorded content items and the potential delay in providing content when manifests are generated on demand.
Innovation Solution
A system utilizing an intelligent cache and machine-learning module to predict when recorded content will be consumed, allowing for the generation and storage of manifests in advance, thereby reducing computational overhead and minimizing delivery delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If manifests are generated on demand for recorded content items, then storage space is saved, but delivery delay increases
Solution Approach 1:
The system performs preliminary actions by generating manifests in advance for recorded content items before they are requested. The machine-learning module predicts when content will be consumed and proactively creates manifests during this predicted time window, eliminating the need to wait for on-demand requests and reducing delivery delays.
Solution Approach 2:
The system uses the predicted consumption pattern of the content itself to trigger manifest generation. The machine-learning module analyzes historical data and predicts when a user will watch recorded content, allowing the system to automatically prepare manifests without requiring explicit user requests, making the system self-anticipating and self-serving.
2Speed
If manifests are pre-generated for recorded content items, then delivery speed is improved, but computational resource consumption increases
Solution Approach 1:
The system performs preliminary actions by generating manifests in advance for recorded content items before they are requested. The machine-learning module predicts when content will be consumed and proactively creates manifests during this predicted time window, eliminating the need to wait for on-demand requests and reducing delivery delays.
Solution Approach 2:
The system changes the parameter of manifest generation timing from reactive (on-demand) to proactive (predictive). By using machine-learning predictions to determine optimal times for manifest creation, the system aligns computational resource consumption with actual content consumption patterns, improving delivery speed while optimizing resource usage efficiency.
3Ease of operation
If the system waits for user requests before generating manifests, then computational resources are conserved, but user experience deteriorates
Solution Approach 1:
The system uses the predicted consumption pattern of the content itself to trigger manifest generation. The machine-learning module analyzes historical data and predicts when a user will watch recorded content, allowing the system to automatically prepare manifests without requiring explicit user requests, making the system self-anticipating and self-serving.
Solution Approach 2:
The system implements feedback loops where the machine-learning module continuously analyzes user viewing patterns and adjusts predictions accordingly. Historical data about when users actually watch recorded content is fed back into the system to refine future predictions, creating a self-improving system that increasingly accurately anticipates user needs.
Data Source
AI summary
Methods, systems, and apparatuses are provided for predicting consumption of recorded content. A content item may be recorded for subsequent consumption by a user. Predictive data associated with when the recorded content item will be subsequently consumed by the user may be determined. The predictive data may be determined based on the content type for the content item and user behavior history. The predictive data may include a predicted time window for receiving a request from the user to consume the recorded content item and a confidence level. Data indicating the predictive data may be associated with or included with the recorded content item, and at least a portion of the recorded content item may be prepared for delivery or delivered to a user device based on the predicted time the recorded content item will be consumed.