Distributed Conference Stream Retrieval for Massive Participant Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio and video conference systems based on selective forwarding units (SFU) face limitations in data distribution pressure when a large number of participants join, restricting the number of simultaneous microphone or video enablers and requiring redirection, leading to a non-smooth user experience.
Innovation Solution
Implementing a method that distributes the load across multiple device groups, each with interface machines, using retrieval devices to manage media data streams and reduce data distribution pressure by sharing interactions, allowing millions of participants to simultaneously use microphones or videos without redirection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single server is used to forward audio and video streams in an SFU-based conference architecture, then the system structure is simple, but the data distribution pressure increases greatly when the number of participants is large, limiting the conference capacity
Solution Approach 1:
The patent divides the single server into multiple interface machines distributed across different data centers. Each interface machine handles a portion of the stream forwarding tasks, segmenting the data distribution pressure. This allows the system to support larger conference capacities while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent introduces a new dimension of spatial distribution by deploying interface machines across multiple data centers rather than concentrating all functionality in a single server. This dimensional expansion from centralized to distributed architecture enables the system to handle large-scale conferences by leveraging geographic and infrastructure diversity.
2Quantity of substance
If multiple interface machines are deployed across different data centers, then the data distribution pressure on each machine is reduced, but the system complexity and interaction management increase
Solution Approach 1:
The patent merges the functionality of multiple interface machines into unified device groups, where each group collaboratively provides stream forwarding services. This consolidation approach reduces the effective complexity by presenting a unified interface while maintaining the distributed architecture benefits for handling large conference capacities.
Solution Approach 2:
The patent introduces retrieval devices as intermediaries between interface machines and the control system. These retrieval devices manage the complexity of interactions by handling stream retrieval and distribution coordination, thereby reducing the communication overhead and management complexity among numerous distributed interface machines.
3Ease of operation
If a single server forwards all media streams, then the interaction management is simple, but the user experience deteriorates when many participants enable microphones or videos simultaneously due to redirection requirements
Solution Approach 1:
The patent segments the data distribution pressure across multiple interface machines, allowing simultaneous handling of multiple high-bandwidth streams without requiring redirection. This segmentation enables all participants to enable microphones and videos concurrently, significantly improving user experience while distributing the computational load.
4Quantity of substance
If retrieval devices are introduced to manage stream distribution, then the data distribution pressure on interface machines is reduced, but the system complexity increases
Solution Approach 1:
The patent designs retrieval devices with multi-functionality, enabling them to perform stream retrieval, caching, and distribution coordination. This universal approach consolidates multiple functions into single components, reducing the overall system complexity while effectively managing data distribution pressure for large-scale conferences.
Data Source
AI summary
This application discloses a method for implementing audio and video conference, an audio and video conference system, and a related apparatus. The method includes receiving, by a retrieval device in a first device group, an application to pull data stream from a first interface machine in the first device group, the application to pull data stream comprising first identification information; and forwarding, by the retrieval device in the first device group, the media data stream corresponding to the first identification information to the first interface machine, the first interface machine forwarding the media data stream to the data accessing terminal; the media data stream being obtained by the retrieval device in the first device group from a second interface machine in a second device group, a data generating terminal that generates the media data stream accessing the conference room by using the second interface machine.