RPC

When a process invokes or calls a process on a remote system, that call is a remote procedure call (RPC).

The calling process is a client (that is, a process requesting a service); the remote process is a server (a process offering a service). As shown in Figure 1, the client sends a request for a procedure to be run, and supplies parameters for that particular run. Once the server has run the procedure, it returns the reply.
Figure 1. Basic RPC operation
Diagram showing the flow of an RPC request as described in the preceding paragraph.

In the RPC model, there is no provision for coordinating changes to recoverable resources in different servers, nor for coordinating changes to recoverable resources in successive calls to the same server. Committing changes to recoverable resources is under the control of the remote procedure, not the client application.

Several RPC implementations have been developed and are now available on a variety of systems. RPC allows a programmer to network an application by distributing the procedures that make up the application across different processors. This is done without the programmer becoming involved with the details of the communication interface required to transmit the parameters to and from the remote procedures.



dfhtmc00188.html | Timestamp icon Last updated: Sunday, 4 June 2017