XdfRTC System Capabilities and Evolution of RTC Media Forwarding Architectures

XdfRTC System Capabilities

High concurrency: supports millions of users simultaneously engaging in real‑time audio/video interaction.

High user coverage: enables global users on various networks to connect with high‑quality, low‑latency access.

High interaction quality: interaction latency under 600 ms with 30 % video packet loss tolerance for smooth communication.

Elastic scaling: minute‑level scaling services handle sudden traffic spikes without user impact.

High reliability: partial core‑datacenter failures do not affect the service.

Multi‑terminal access: supports Windows, macOS, Android, iOS, and web clients.

RTC Media Forwarding Architecture Evolution

Three Common RTC Media Forwarding Architectures

MESH

All users connect directly to each other for media transmission.

Broadcaster B1 sends its audio/video to B2, B3, B4 simultaneously, causing high outbound bandwidth consumption; similar for other broadcasters.

Advantages:

No reliance on media servers for forwarding.

Low internal network latency and no outbound bandwidth usage.

Disadvantages:

Complex user network topology leads to low connection success rates.

Connection management becomes complicated as the number of users grows.

Media data must be sent multiple times, resulting in high outbound bandwidth.

MCU

Users send their media streams to an MCU server, which decodes, mixes, and re‑encodes them into a single stream for distribution.

The MCU generates a separate mixed stream for each user (each user receives a different stream).

Advantages:

Low downstream bandwidth consumption for users.

Client decoding and display resource usage is low.

Disadvantages:

High server transcoding resource consumption.

The server must mix a separate audio/video stream for each user.

Increased latency due to server processing.

Clients have limited control over media streams.

SFU

Clients send their media to a server, which only forwards the streams to other clients without decoding.

Receiving clients can flexibly select, receive, and display individual streams.

Advantages:

Server performs only forwarding, resulting in low resource consumption.

Receivers can dynamically control which streams to display.

Disadvantages:

Server bandwidth consumption is high.

Mesh Forwarding Architecture

Each service node acts as an SFU, forwarding media streams and serving as an access point for users.

Service nodes interconnect, forming a mesh; two SFUs communicate via the shortest path.

Users select the SFU node that best matches their region and ISP.

If a user or access SFU experiences network issues, automatic failover to another SFU occurs.

Advantages:

All services only forward streams, resulting in low CPU usage.

Uniform service functionality enables elastic scaling.

Inter‑service connections allow automatic optimal‑path switching, ensuring transmission quality.

Disadvantages:

Server resources vary in performance, making optimal utilization difficult.

When many service nodes are deployed, network connection management becomes complex.

Hierarchical Forwarding Architecture

Each node functions as an SFU, forwarding media streams.

Access layer: handles user connections within the same ISP/region.

Core network layer: forwards traffic across regions and ISPs.

Internal transmission lines are monitored in real time and switched to optimal paths.

Advantages:

Leverages cloud service resources by distinguishing core and edge services.

Reduces unnecessary media stream transmission.

Decreases the number of core forwarding networks, simplifying management.

Supports edge regions with same‑ISP and same‑region coverage.

Disadvantages:

Access‑layer SFUs currently do not support direct peer‑to‑peer transmission.

RTC Recording and Streaming System

Mixes interactive channel streams and pushes them to CDN for large‑scale live broadcast scenarios.

Mixes interactive channel streams for recording and playback.

RTC Monitoring System

Real‑time monitoring of user interaction quality.

Real‑time monitoring of network quality for each service node.

Real‑time monitoring of transmission line quality.

Statistical analysis of users, services, and other dimensions.