Predictive Video Caching for Low-Latency Online Playback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Viewers experience significant latency when selecting online digital videos due to poor Internet connections, inefficient content identification, and improper digital video player functioning, leading to frustration and reduced consumption.
Innovation Solution
Implement a video accelerator system that predicts user likelihood of selecting digital content, initiates mediation, unwraps video tags, caches video content, stitches videos together, and initializes players before user selection, using modules like video likelihood, mediation initiation, caching, unwrapping, and player initiation to reduce latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If video content is identified and prepared only after user selection, then system resources are conserved, but video latency increases significantly
Solution Approach 1:
The system performs preliminary actions by identifying and preparing video content before the user actually selects it. The video likelihood module predicts which videos the user will select, and the system initiates mediation, unwraps video tags, and caches video content in advance, so that when the user clicks, the video is already ready for immediate playback.
Solution Approach 2:
The system dynamically adjusts its preparation behavior based on predicted user likelihood. Instead of a static prepare-or-wait approach, the system uses the video likelihood module to determine which videos warrant advance preparation, creating a dynamic, adaptive system that optimizes the balance between reducing latency and conserving resources.
2Speed
If video content is cached on client device before selection, then playback speed improves, but device storage requirements increase
Solution Approach 1:
The system applies caching selectively rather than universally. The video likelihood module identifies which specific videos are most likely to be selected, and only those videos are cached on the client device. This local quality approach ensures that storage space is used efficiently for high-priority content while maintaining fast playback speed for predicted videos.
Solution Approach 2:
The system performs partial caching based on predicted likelihood thresholds. Rather than caching all possible videos or none at all, the system caches only the portion of video content that exceeds the likelihood threshold, achieving a balance between storage constraints and playback performance.
3Productivity
If mediation process is initiated before user selection, then video identification efficiency improves, but system processing overhead increases
Solution Approach 1:
The mediation process is initiated in advance based on predicted user behavior. The video likelihood module triggers the mediation process before the user actually selects the video, allowing the system to identify and prepare the correct video content ahead of time, improving identification efficiency while managing processing overhead through selective initiation.
Data Source
AI summary
Examples of the systems and methods described herein relate to reducing latency issues associated with playing online digital video on client devices. An example method includes: determining a likelihood that a user will trigger presentment of video content within a portion of digital content, and when the likelihood is determined to exceed a threshold, obtaining a video tag that identifies the video content or caching a portion of the video content prior to presentment of the video content.