What are the differences between Mirotalk and the SFU version?

[EXT-OWL]

If the title didn't give it away, I am very new to the whole self hosted scene. I have been looking for an affordable way to create calls between my clients, and Mirotalk ticks all the boxes. However, I have no idea what the difference between the normal one and SFU are. I'm using the normal (works perfectly for me), but I'd like to know if there are any additional features with SFU that I am missing out on.

[jdaviescoates]

@EXT-OWL said in What are the differences between Mirotalk and the SFU version?:

SFU

= Selective forwarding unit

In short the SFU version should work much better than the P2P version for meetings with more than 2 people.

see e.g. https://getstream.io/blog/what-is-a-selective-forwarding-unit-in-webrtc/

[MiroTalk]

@EXT-OWL said in What are the differences between Mirotalk and the SFU version?:

However, I have no idea what the difference between the normal one and SFU are

The key difference between MiroTalk P2P and MiroTalk SFU lies in the way they handle the transmission of media streams between participants in a video conferencing environment. Here's a breakdown:

---

1. MiroTalk P2P (Peer-to-Peer)

• Architecture: Directly connects each participant to every other participant in the session.
• How it works:
  • Each participant sends their audio/video stream directly to every other participant.
  • For example, in a 3-person room, each person sends their stream to 2 others and receives 2 streams.
• Pros:
  • Low latency due to direct connections.
  • No centralized server is required for media relay, reducing infrastructure costs.
  • Good for small group meetings (2-4 participants).
• Cons:
  • Not scalable: Bandwidth usage grows significantly with more participants since each participant must upload multiple streams (N-1 streams, where N is the number of participants).
  • Participants with limited upload bandwidth may experience performance issues in larger rooms.

---

2. MiroTalk SFU (Selective Forwarding Unit)

• Architecture: Uses a centralized server to manage media streams.
• How it works:
  • Each participant sends their audio/video stream to the MiroTalk SFU built in server, which then selectively forwards streams to other participants.
  • For example, in a 3-person room, each person sends 1 stream to the SFU, and the SFU distributes the streams to the others.
• Pros:
  • Scalable: Each participant uploads only one stream regardless of the number of participants, as the SFU handles distribution.
  • Better suited for larger meetings or when participants have limited upload bandwidth.
  • Allows for additional features like transcoding and quality control (e.g., adaptive bitrate).
• Cons:
  • Introduces some latency since streams are relayed through a server.
  • Requires server resources and maintenance, increasing infrastructure costs.

---

When to Use Each?

• MiroTalk P2P:

  • Ideal for small, private calls (e.g., 1-on-1 consultations or small meetings).
  • Lower infrastructure costs and simpler setup.

• MiroTalk SFU:

  • Suitable for larger rooms or scenarios where bandwidth efficiency and scalability are crucial.
  • Offers more advanced features compared to P2P.

---

We have developed a range of MiroTalk solutions, including SFU, P2P, C2C, BRO, and WEB, allowing users to select the most suitable option based on their specific use cases and requirements. For an in-depth look at our WebRTC architecture, please visit our documentation.

[EXT-OWL]

@mirotalk-57bab571- thank you so much !!