3.3.3.2.1 Connection Time-Out

A higher-level protocol on the server can instruct the RPC runtime to monitor the state of the connection in an implementation-dependent<117> manner above and beyond the monitoring provided by default by the RPC transport, so that if the client crashes or loses network connectivity to the server, the server can take recovery action. In the common case, the recovery action is a context handle rundown. Depending on the protocol sequence for the method call, the establishment of the timer acts only as advice to the RPC runtime system.

When this timer expires, the expiry is not noticed at the RPC protocol level, but it is noticed at the TCP/IP protocol level and shows as a local event, as specified in section 3.3.3.7.1.