Media Gateway Selection via Quality Prediction Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for selecting media gateways for voice/video over IP calls fail to accurately predict network performance, leading to sub-par quality due to factors like jitter, latency, and bandwidth fluctuations, and lack control over end-user media experience in cloud collaboration services.
Innovation Solution
A two-step modeling process that predicts quality metrics and user ratings for media gateways, using historical and real-time network data to rank available gateways and route incoming connections to the one with the highest predicted user rating, thereby improving call quality and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If media gateway selection is based on location or load-balancing, then routing simplicity is maintained, but media quality prediction accuracy deteriorates
Solution Approach 1:
The system performs preliminary quality prediction and modeling before actual media transmission begins. Historical network metrics and real-time data are analyzed in advance to forecast quality metrics (packet loss, jitter, latency) and generate quality scores for different gateways and routes, enabling proactive selection rather than reactive routing
Solution Approach 2:
A machine learning modeling system acts as an intermediary between the routing system and the media transmission system. This intermediary analyzes historical and real-time network data, predicts quality metrics, and provides quality scores that guide gateway selection, decoupling the complexity of quality analysis from the routing decision-making process
2Ease of manufacture
If conventional routing methods are used in cloud collaboration services, then service deployment is simplified, but control over end-user media experience deteriorates
Solution Approach 1:
The system implements continuous feedback loops where real-time network metrics are collected during media transmission, compared against predicted values, and used to update quality models. This feedback mechanism enables dynamic adjustment of routing decisions and proactive mitigation of quality issues, giving service providers control over user experience while maintaining cloud-based simplified deployment
3Reliability
If network conditions are monitored in real-time, then media quality can be improved, but system complexity increases
Solution Approach 1:
The quality prediction system is segmented into modular components: data collection modules that gather network metrics, modeling modules that process historical and real-time data, prediction modules that forecast quality metrics, and selection modules that determine optimal gateways. This segmentation allows real-time monitoring to be implemented incrementally and maintains system manageability while improving media quality
Data Source
AI summary
Disclosed is a system, method and computer readable medium enabling collaboration service providers to more accurately predict packet loss, jitter and delay based on current session, historical session and user location parameters. The prediction can be used to forecast the occurrence of poor media quality at the current location and potential future locations.