SFU Media Server Bandwidth Drop Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-party videoconferences using selective forwarding units (SFUs) face suboptimal quality of experience due to delayed detection of sudden drops in available egress bandwidth, leading to packet losses and delivery delays.
Innovation Solution
A media server with an SFU architecture detects a sudden drop in egress bandwidth and sends a request to a video sender device to reduce its transmission bitrate, determines whether to reset the video stream, stops forwarding, sets a wait time, and then restarts the stream using an intra-frame request to mitigate congestion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the central server computer detects a drop in available egress bandwidth and sends a request to reduce transmission bitrate, then the quality of experience is improved, but the detection and response are delayed causing packet losses and delivery delays
Solution Approach 1:
The patent implements preliminary action by proactively resetting the video stream forwarding and requesting I-frames before congestion fully develops. When bandwidth drop is detected, the system immediately stops forwarding and prepares recovery mechanisms in advance, rather than waiting for packet losses to accumulate. This proactive approach reduces the effective response time and mitigates the impact of detection delays.
Solution Approach 2:
The patent applies the skipping principle by rapidly transitioning through stream reset states and forcefully requesting I-frames to skip over the congested period. The system rushes through the necessary recovery actions (stop forwarding, reset stream, request I-frame, restart forwarding) in rapid succession to get past the congestion window, minimizing the time spent in a degraded state.
2Reliability
If the central server computer continuously monitors and adjusts transmission bitrate, then packet losses are reduced, but system complexity increases
Solution Approach 1:
The patent implements feedback mechanisms where the central server computer monitors network conditions (bandwidth availability, packet delivery status) and uses this information to adjust stream forwarding decisions. RTCP receiver reports provide feedback about packet delivery quality, which triggers bandwidth drop detection and subsequent stream reset actions. This closed-loop feedback system automatically adjusts to network conditions without requiring complex manual configuration.
Solution Approach 2:
The system applies self-service by automatically detecting bandwidth drops and executing recovery procedures without external intervention. The central server computer autonomously monitors its own egress bandwidth, detects congestion conditions, and initiates stream reset and I-frame request sequences independently, reducing the need for complex external control mechanisms.
Data Source
AI summary
Techniques for handling sudden changes in available bandwidth in videoconferences implemented with media servers employing SFUs. The techniques involve detecting a sudden drop in available egress network bandwidth from a media server to a video receiver, sending a request to a video sender currently sending a video stream to the media server for forwarding to the video receiver to reduce its transmission bitrate, determining whether to reset the forwarding of the video stream to the video receiver, and, having determined to reset the forwarding of the video stream to the video receiver, stopping the forwarding of the video stream to the video receiver, sending a request for an I-frame to the video sender for use in restarting the forwarding of the video stream to the video receiver, and, having received the requested I-frame, restarting the forwarding of the video stream from the video sender to the video receiver.