What is SFU?

A Selective Forwarding Unit (SFU) is a type of media server used in real-time communication that optimizes the handling of audio and video streams in conference calls. It receives media streams from each participant and intelligently decides which streams to forward to others. Rather than processing or mixing the media itself, the SFU forwards the streams selectively based on factors like bandwidth, participant layout, or user preferences, allowing for lower latency and more efficient bandwidth usage compared to other architectures.

SFUs differ significantly from Peer-to-Peer (P2P) connections and Multipoint Control Units (MCUs). In P2P, each participant sends and receives media directly from every other participant, which can quickly become inefficient and unmanageable as the number of participants increases, due to the exponential growth in connections. In contrast, an SFU receives media streams from each participant and selectively forwards them to others, requiring only a single connection per participant to the server, thus making it far more scalable.

Unlike MCUs, which receive media from all participants, process or mix the streams, and then broadcast the combined output, SFUs only forward the streams without altering them. This approach reduces server load and latency, making SFUs well-suited for larger conferences with multiple participants. Additionally, SFUs can leverage cascading techniques to further enhance scalability, allowing the architecture to support a significantly larger number of participants while maintaining performance and quality.

How Does A SFU Work ?

A SFU manages the flow of media between participants by receiving, analyzing, and selectively forwarding media streams. It starts by inspecting the incoming media from each participant to understand details such as the codecs used, video resolution, and frame rate. This allows the SFU to adapt to the characteristics of the streams and the network conditions.

Once the media streams are analyzed, the SFU makes intelligent decisions about which streams to forward to other participants. For instance, if multiple participants are sending high-definition video, the SFU may choose to forward only one HD stream to optimize bandwidth usage. Additionally, the SFU can perform transcoding, converting media from one format to another to reduce bandwidth consumption or ensure compatibility with different endpoints.

After making these decisions, the SFU forwards the selected media streams to the relevant participants using transport protocols like UDP or TCP.

What is a MCU ?

WebRTC Multipoint Control Unit (MCU) is a server used to facilitate real-time communication (RTC) across devices and applications. It enables multi-party audio and video conferencing by combining various media streams into a single, unified stream, making it ideal for VoIP and video conferencing applications.

Unlike Peer-to-Peer (P2P) connections, where each participant directly sends and receives media from others, an MCU simplifies communication by centralizing the media processing. It also differs from Selective Forwarding Units (SFUs), which only route incoming streams without merging them. In an MCU setup, the server handles all the audio and video streams, mixing them into a composite signal that can be easily transmitted to participants.

MCUs are advantageous for ensuring compatibility with legacy systems because they emit standard, mixed media streams that can be decoded by various devices and software. While P2P connections can quickly become inefficient with multiple participants—requiring direct connections between each pair—an MCU requires only one connection per participant, enabling support for a larger number of users.

However, the MCU architecture can be costly and challenging to scale. The process of mixing and transcoding audio and video streams into a single output demands significant computational power, potentially straining server resources as the number of participants increases. Despite these limitations, MCUs remain a valuable solution for scenarios where high compatibility and unified media streams are needed.

What is a Peer?

A peer is a computer connected to a Peer-to-Peer (P2P) netowrk. Peers can connect and disconnect from the network at any time, and they can share resources with other peers on the network.

P2P networks are different from traditional client-server networks in which each client is connected to a central server. In a P2P network, there is no central server; instead, each peer is equal and can connect to any other peer on the network

P2P networks are often used for video conferencing, file sharing and other applications where users need to transfer data between each other.

How does A P2P(MESS) network work?

In a P2P network, each peer is equal and can connect to any other peer on the network. In order for two peers to communicate, they first need to find each other on the network. To do this, each peer broadcasts a message that contains its own IP address. Other peers on the network can then reply with their own IP addresses, allowing the two peers to establish a connection.

Disadvantages of Peer vs. MCU vs. SFU

Peer-to-Peer (P2P)

1. Scalability Issues: Connection requirements grow exponentially with participants, leading to high bandwidth usage.

2. Bandwidth Consumption: Each participant maintains multiple connections, straining limited upload/download speeds.

3. Device Performance Limitations: Encoding and decoding multiple streams can overwhelm lower-end devices.

Multipoint Control Unit (MCU)

1. High Server Load: Significant computational power is needed for mixing and transcoding, increasing operational costs.

2. Latency Issues: Mixing processes can introduce latency, affecting real-time communication.

3. Expensive Infrastructure: Requires powerful hardware and more server instances for scaling.

Selective Forwarding Unit (SFU)

1. Bandwidth Requirements: Participants send media to the SFU, which can strain bandwidth for users with limited speeds.

2. No Stream Mixing: Participants receive multiple streams that need client-side mixing, complicating application logic.

3. Compatibility Limitations: May not work well with legacy systems that lack support for WebRTC or multiple streams.

Advantages of P2P, MCU, and SFU

Peer-to-Peer (P2P)

• Simplicity: Easy to set up and use.
• Cost-Free: No server costs involved.

Selective Forwarding Unit (SFU)

When choosing a WebRTC architecture, consider the following factors:

• Cost
• Scalability
• Flexibility

Benefits of Using an SFU:

• Support for Large Parties: Can handle many participants efficiently.
• Transcoding Capabilities: Converts media streams between different codecs, ensuring compatibility.
• Bandwidth Management: Allows limits on bandwidth usage for video/audio to optimize performance.
• Security Features: Enhanced security for media streams.
• Quality of Service (QoS): Ensures consistent media quality.

Difference Between MCU and SFU

• MCU: Broadcasts mixed media to all participants.
• SFU: Selectively forwards streams to participants based on their needs.

Conclusion

Each WebRTC implementation has distinct advantages:

• P2P: Simple but less efficient in bandwidth and CPU usage; requires high-speed connections.
• MCU: Reliable and resource-intensive, but complex to set up and maintain.
• SFU: More efficient, supports lower-speed connections, and balances reliability with resource usage.