About the WEBRTC video conferencing solution

In my experience, video conferencing is divided into two modes: mesh model, star type

1. Mesh model

The so-called mesh model is a point-to-point connection for every two people who attend the conference.

For example: A meeting room with three people a,b,c based on a network model, this is required:

1, A and b establish a connection
2, B and C establish the connection
3, A and C establish a connection

Such a three-person meeting would require the establishment of three connection calculation methods: (3-1)/2

The same: 4 person meetings need to establish the number of connections to the (4-1)/2 = 6, 5 person meetings need 5* (5-1)/2 = 10

　　Scenario : This model is only suitable for 3-4 of people attending meetings, because each additional person participating in the meeting increases the N-1 connection (n indicates the number of participants in the meeting), so that the terminal load increases sharply.

2. Star-shaped model

Star model is divided into: through the server synthesis forwarding and through a terminal to forward two cases.

A, in 3 people also this use the following model:

1, a,b,c attend the meeting

2, A and B establish A connection

3, B and C establish the connection

4, B forward a audio and video to c,b forward C audio and video to a

This situation in the case of B equipment performance is high, and a and C performance is weak, with B as a bridge to achieve 3-party calls, thus reducing the burden on the server. applicable Scenario : This model is only suitable for meetings of 3 people.

B. forwarding via server synthesis

Everyone attending the meeting sent the audio and video they collected to the server, and after the synthesis of the server, they were distributed to everyone attending the meeting.

The following model:

1, a,b,c attend the meeting

2, a,b,c respectively and the server to establish a connection

3, A,b,c The collected audio to the server

4, the server sent A,b,c audio and video after the synthesis to a,b,c

So no matter how many people attend the meeting, each of the participating terminals only establish a connection, put the load on the server side, applicable scenario : For more than 4 people of the meeting

　　　　There is a problem : Do not let the server send back the data sent by the terminal, such as the server should not send a to the server's audio and video sent to a, because then a will listen (see) to their own voice (video), it is unreasonable

Unreasonable place please the vast numbers of users correct: 576487833

About the WEBRTC video conferencing solution