Bandwidth Prediction Source Selection for Video Streaming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video streaming technologies face challenges in dynamically selecting the best bandwidth prediction source for optimal video quality, as they often rely on either local historical data or network-based predictions, which may not accurately reflect real-time conditions, leading to inconsistent user experience.
Innovation Solution
A method that dynamically selects the best bandwidth prediction source by combining local historical throughput data, network historical data, and real-time network measurements, allowing for validation against actual throughput and adaptive decision-making to choose the most accurate prediction source for each video session.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If local historical throughput data is used for bandwidth prediction, then the prediction is based on device-specific patterns, but it may not reflect real-time network conditions accurately
Solution Approach 1:
The patent combines multiple bandwidth prediction sources (local historical data, network historical data, and real-time network measurements) into a unified prediction system. The video player evaluates predictions from different sources and selects the most accurate one dynamically, merging the advantages of each source to achieve both precision and adaptability.
Solution Approach 2:
The system dynamically selects which prediction source to use based on current performance validation. The video player continuously monitors actual throughput and adjusts its preference for different prediction sources in real-time, making the system adaptive to changing network conditions rather than relying on a fixed approach.
2Adaptability or versatility
If network-based prediction is used according to historical throughput data, then network-wide patterns are captured, but device-specific conditions may be overlooked
Solution Approach 1:
The system merges network-based historical data with local device-specific historical data, allowing the video player to benefit from both network-wide patterns and device-specific characteristics. This combination enables the system to capture macro-level trends while maintaining sensitivity to individual device conditions.
Solution Approach 2:
The patent applies local quality by allowing different prediction sources to serve different purposes. Local historical data provides device-specific customization while network historical data provides broader context. The system selectively applies each source based on which better suits the current prediction needs, treating each data source with its appropriate level of specificity.
3Measurement precision
If multiple bandwidth prediction sources are utilized, then prediction accuracy can be improved through validation, but system complexity increases
Solution Approach 1:
The system implements feedback by continuously monitoring actual throughput and using this information to validate and adjust predictions from different sources. The video player learns from past performance which prediction sources are most accurate under different conditions and adjusts its selection accordingly, reducing the need for complex manual configuration.
Solution Approach 2:
The video player autonomously evaluates and selects the best prediction source without requiring external intervention. The system self-manages the complexity by automatically validating predictions against actual throughput and making informed decisions about which data source to trust, eliminating the need for manual system configuration or complex external control mechanisms.
Data Source
AI summary
Aspects of the subject disclosure may include, for example, obtaining, from a first source of information, a first bandwidth prediction, wherein the first bandwidth prediction is based upon historical bandwidth data that had been provided by a plurality of devices; obtaining, from a second source of information, a second bandwidth prediction, wherein the second bandwidth prediction is based upon network measurements, and wherein the network measurements are other than the historical bandwidth data that had been provided by the plurality of devices; selecting as a source of a future bandwidth prediction one of the first source of information and the second source of information, wherein the selecting is based upon a comparison of each of the first bandwidth prediction and the second bandwidth prediction to an actually obtained bandwidth of the device. Other embodiments are disclosed.