SDN Multicast Tree Packing for Video Conferencing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-party video conferencing systems face challenges in delivering high-quality video with low end-to-end delay due to bottlenecks and delays caused by the multipoint control unit (MCU)-based approach, which is inefficient and prone to single-point failures.
Innovation Solution
The implementation of a software-defined networking (SDN)-enabled IP multicast solution that uses a media controller and SDN controller to construct and pack multicast trees, optimizing video sending rates and routes to maximize system-wide utility while ensuring low delay and high bandwidth, utilizing algorithms like Modified-BSMA and Proportional-Rate-Alloc to manage network resources effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If MCU-based approach is used for multi-party video conferencing, then user state management is simple and easy, but end-to-end delay increases significantly and system reliability decreases
Solution Approach 1:
The patent extracts the video mixing function from the centralized MCU and distributes it to edge servers located closer to end users. Each edge server independently mixes videos for its local users, eliminating the need for all traffic to pass through a central MCU, thereby reducing end-to-end delay while maintaining simple user state management at each edge.
Solution Approach 2:
The patent segments the centralized MCU function into multiple distributed edge servers. Instead of a single point of control, the system divides the mixing function across multiple edge locations, allowing parallel processing of video streams and reducing the delay bottleneck that would occur at a single centralized MCU.
2Ease of manufacture
If MCU-based approach is used for multi-party video conferencing, then implementation is simple, but system reliability decreases due to single point of failure
Solution Approach 1:
The patent segments the centralized MCU into multiple distributed edge servers, eliminating the single point of failure. Each edge server operates independently to mix videos for its local users, so a failure at one edge server does not affect other parts of the system, thereby improving reliability while keeping implementation relatively simple.
Solution Approach 2:
The patent introduces edge servers as intermediary components between end users and the central controller. These edge servers act as local proxies that handle video mixing for their respective users, distributing the system architecture and eliminating the single point of failure that would exist at a centralized MCU.
3Productivity
If video sending rate is increased to improve video quality, then video quality improves, but network congestion increases and delay increases
Solution Approach 1:
The patent applies local quality optimization by adjusting video sending rates at each edge server based on local network conditions and user requirements. Instead of using a uniform high rate throughout the network, each edge server optimizes the rate for its specific local conditions, improving video quality where possible while avoiding unnecessary congestion and delay in other areas.
Solution Approach 2:
The patent implements dynamic video rate adjustment at each edge server based on real-time network conditions. The system dynamically adapts the sending rate to balance video quality with network congestion and delay constraints, allowing high quality when network conditions permit while reducing rates to avoid congestion when necessary.
Data Source
AI summary
Method, device and apparatus for delivering video telephony by using a media controller and a software defined networking (SDN) controller to enable multicasting for multi-party video conferencing. A media controller generates a video sending rate for each of the plurality of client devices based on a received video conference request and received network information. The media controller determines and independently constructs a maximum-rate delay-bounded multicast tree for each of the plurality of client devices that supports the largest possible multicast rate. The media controller determines a common congested link in the multicast trees, re-routes the multicast tree with the common congested link for the client device with the lowest weight of utility divided by its video sending rate and creating two subtrees, and iteratively re-routes the multicast trees until an aggregate video sending rate of the multicast trees sharing the common congested link meets a link capacity of the common congested link.