The endpoint specifies the communication port clients use to make remote procedure calls to a server.
The server application specifies endpoint information at the same time it specifies the protocol sequence by calling the RPC routine that starts with the prefix "RpcServerUseProtseq" or "RpcServerUseAllProtseqs."
A finite number of endpoints are available for any protocol sequence. Some of these are usually assigned by the authority responsible for the protocol. The syntax of the endpoint string depends on the protocol sequence you use. For example, the endpoint for TCP/IP is a port number, and the endpoint syntax for named pipes is a valid pipe name.
The major design decision you must make regarding the endpoint is whether it is well known or dynamic. Your choice of option determines whether the distributed application or the run-time library specifies the endpoint the application will use.
Most applications should use dynamic endpoints so the endpoint-mapping service can dynamically map a distributed application to an endpoint available for the protocol. In this way, this limited system resource can be assigned to a distributed service at run time as needed, instead of being dedicated to a distributed service when the service is developed.