Video Call will attempt to use the best network path that it can find.
The following table lists the network paths it will look for, in order of preference:
|Network path||STUN/Relay server port|
1: Direct peer-to-peer UDP, with STUN server-assisted NAT traversal
Each endpoint will discover its external Internet address using the provided STUN Server. This address is provided to the other endpoint and used to set up the connection through Network Address Translation.
Media flows over randomly selected ports over large range of UDP ports 49152 - 65535.
2: Via Video Call relay server, using UDP-routed egress
If a connection cannot be established using the above direct peer to peer, then the configured TURN server UDP port 3478 will be tried to establish a relay to the remote endpoint. This relay address is provided to the other endpoint and used to set up the connection through the relay, back through the local endpoint's connection to the TURN server.
Media flows to UDP Port 3478 on the TURN server.
3: Via Video Call relay server, using TCP-routed egress
If a connection cannot be established using UDP to the TURN Server, the connection to the TURN Server is established via TCP 443 not UDP 3478.
Media flows outwards to TCP port 443 on the TURN Server.
4: Via Video Call relay server, using TCP tunnelling through a local web proxy server
If a routed connection via NAT cannot be established to the TURN Server, a tunnelled connection to TCP port 443 will be attempted via the browsers configured web proxy server.
Media flows outwards through the web proxy, to TCP Port 443 on the TURN Server.
5a, 5b: Via Video Call relay server, using Secure TCP
As for 3 or 4 above, but using a TLS TCP connection to the TURN Server.
For more information, see Video Call relay servers.
Video Call relay server address: vcct.healthdirect.org.au
For most network paths, the negotiation will likely result in a valid media connection.
Recommended option: For many networks, allowing NAT egress to UDP port 3478 on the relay server (network path 2, above) will provide low latency with little overhead. This should require only a minor, low-risk change to your network configuration.