In order to achieve real-time face to face communication, video conferencing the endpoints carry out a significant amount of processing in order to reduce the latency of when the video and audio signals are compressed into a data stream and then sent across a network. There is however a limit to the endpoints processing capability and can only deal with a certain amount of delays across networks, therefore protection is needed for the traffic moving from one endpoint to another.
The following metrics are used to determine a network’s performance.
In video conferencing latency is the coding / decoding delay of the video and audio being compressed / decompressed.
An acceptable latency using today’s high definition codecs is up to 300 milliseconds. This ensures the video and audio data can transfer from endpoint to endpoint through the network and cause little disruption to a conference. Reducing the latency is vital to ensuring a true-to-life video communication experience.
Packet loss is when one or more packets of data traveling across a network are unable to reach their destination. Losing packets from compressed data has a greater impact on the quality of the stream and becomes apparent at the stage of decompression.
The network solution for reducing loss of data relies on the Transmission Control Protocol which ensures the delivery of traffic between the video conference devices. Unfortunately ensuring reliability using the Transmission Control Protocol can result in an increase in latency. When devices communicate there is handshake that occurs to ensure all information has been sent and received. This handshake can produce fluctuations in the data signal because one device is waiting for confirmation from the other.
The increase in latency is not suitable for video conferencing. To deliver real-time time traffic over a network the User Datagram Protocol is used which allows packets to be sent and received whilst the video conferencing hardware carries out the processing work to minimise the impact of lost or delayed packets.
The delivery of packets in an uncongested network is pretty consistent. However, in a congested network the delay between packets can vary. This inter packet delay variation is usually measured in milliseconds with an acceptable delay being around 20ms.
In order to achieve high definition video conferencing many systems require as little as 1mbps dedicated bandwidth for a point to point call. Systems hosting a HD multipoint call often require a minimum of 3-4mbps. An acceptable video quality can be achieved at as little as 500kbps.
Dedicating the necessary bandwidth for video conferencing will help reduce the possibility of latency and packet loss.
Often when a network bandwidth is close to capacity a quality-of-service can be implemented to effectively reserve bandwidth for video conferencing traffic.
For a network to support video conferencing the points in this article need to be seriously considered. Provisioning your network correctly will help keep latency and packet loss to a minimum which is vitally important to ensure interaction remains smooth and consistent during video calls.