Client-Server Throughput Prediction for Video Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video distribution systems face challenges in accurately predicting network throughput, leading to suboptimal video quality settings due to lower prediction accuracy when limited data is used, especially when relying on either client or server predictions alone.
Innovation Solution
An information processing apparatus that combines its own throughput prediction with a server's prediction, selecting the more accurate value and requesting video transmission quality based on this selection to improve network throughput prediction accuracy and adapt video quality accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If only client apparatus throughput prediction is used, then device complexity is reduced, but throughput prediction accuracy deteriorates
Solution Approach 1:
The patent combines throughput prediction results from both the client apparatus and server apparatus to determine the final video quality. The selection unit selects the prediction result with higher accuracy based on comparison with actual throughput, effectively merging two prediction systems to overcome the limitations of using either system alone.
2Measurement precision
If only server apparatus throughput prediction is used, then data availability for prediction is increased, but device complexity increases
Solution Approach 1:
The server apparatus acts as an intermediary that provides throughput prediction based on aggregated network data from multiple sources. The client apparatus receives both the server's prediction and its own prediction, then selects the more accurate one through comparison with actual throughput measurements.
3Measurement precision
If prediction data amount is increased, then throughput prediction accuracy is improved, but information processing complexity increases
Solution Approach 1:
The patent divides the prediction data collection into two separate segments: the client apparatus collects local network performance data, while the server apparatus collects aggregated data from multiple clients. This segmentation allows each system to use appropriate data volumes for their respective prediction algorithms without overwhelming either system.
Data Source
AI summary
An information processing apparatus includes a predictor configured to predict a throughput of a network, to generate a first predicted value indicative of the predicted throughput, an acquirer configured to acquire, from a server apparatus connected to the network, a second predicted value indicative of a throughput of the network predicted by the server apparatus, a selector configured to select a predicted value with higher prediction accuracy from among the first predicted value and the second predicted value X2, and a requester configured to request a distribution apparatus to transmit a video with a quality based on the predicted value selected by the selector 125.