Internet bonding has become indispensable, especially in the livestream sector. The bundling of different Internet connections into one stable connection is more than helpful for a live transmission. This is because there is often no Internet via LAN on site, and the livestreamers have to rely on mobile communications. With Mine Media bonding servers, this is not a problem. We explain the principle behind bonding servers in this post. For those with general questions about Internet bonding, we recommend the blog post: “A strong connection for livestreaming: LTE bonding“.
Parallel transmission over multiple connections
First of all, the audio and video signal is compressed during a live stream transmission. Subsequently, the splitting unit groups the data packets and sends them to the bonding server. The server then bundles the data packets over several network channels. The advantages of this: the transmission pressure of each network is lower and the transmission rate is adjusted in real time to the transmission capacity of each network, ensuring stability of the live video.
Adaptation of the distribution share of connection data packets
The mobile network 4G and also Wifi cannot guarantee fast transmission speeds. This is because performance depends on how many users are on the same network at the same time. Because of this, they are susceptible to network jitter, large time delays, high packet loss rates, and other problems. It’s not for nothing that 4G and Wifi are often referred to as unreliable networks in expert circles. If the data is not delivered to the server in time during transmission via a single mobile phone connection, the image cannot be restored, among other things. This in turn means that the camera’s image does not reach the viewers of the livestream, the screen remains black in the worst case.
Mine Media’s Bonding Servers solves the above problems thanks to various techniques. Thus, after receiving the data packet, the receiving side sends back a table with information about the reception quality to the sending side. This side then makes decisions based on numerous parameters such as packet loss rate, time delay, signal quality, etc., to ensure the stability of the transmission and the maximum bandwidth usage of each connection.
Forward error correction
Let’s move on to the next step: correcting the forward error. For this purpose, the transmitting side groups the transmitted data packets into groups of n each. As a next step, the appropriate number of redundant packets M is selected according to the quality assessment of the connection in order to generate m corresponding redundant packets. The sum of n+m results in a certain number of network data packets. These packets arrive at the recipient via many different connections. Then the receiving side can restore the original data packets according to the received data packets. In this way, the receiver then compensates for the interference that may be caused by the network. At the same time, transmission reliability is also improved. This also has a great influence on the speed. This is because the average delay in transmission can be effectively reduced by using redundant data.
Bit rate adaptive real-time assessment of the entire network bandwidth
If the network bandwidth is insufficient, the bonding server adjusts the audio and video data. This is done by reducing the resolution, frame rate, bit rate and GOP value to ensure a smooth live transmission.