Asterisk - The Open Source Telephony Project
18.5.0
|
The purpose of this API is to provide a way for multiple RTP stacks to be used inside of Asterisk without any module that uses RTP knowing any different. To the module each RTP stack behaves the same.
An RTP session is called an instance and is made up of a combination of codec information, RTP engine, RTP properties, and address information. An engine name may be passed in to explicitly choose an RTP stack to be used but a default one will be used if none is provided. An address to use for RTP may also be provided but the underlying RTP engine may choose a different address depending on it's configuration.
An RTP engine is the layer between the RTP engine core and the RTP stack itself. The RTP engine core provides a set of callbacks to do various things (such as write audio out) that the RTP engine has to have implemented.
Glue is what binds an RTP instance to a channel. It is used to retrieve RTP instance information when performing remote or local bridging and is used to have the channel driver tell the remote side to change destination of the RTP stream.
Statistics from an RTP instance can be retrieved using the ast_rtp_instance_get_stats API call. This essentially asks the RTP engine in use to fill in a structure with the requested values. It is not required for an RTP engine to support all statistic values.
Properties allow behavior of the RTP engine and RTP engine core to be changed. For example, there is a property named AST_RTP_PROPERTY_NAT which is used to tell the RTP engine to enable symmetric RTP if it supports it. It is not required for an RTP engine to support all properties.
Codec information is stored using a separate data structure which has it's own set of API calls to add/remove/retrieve information. They are used by the module after an RTP instance is created so that payload information is available for the RTP engine.