DISTRIBUTED SYSTEMS
Communication using RPC
Intro
Independently developed IPC protocol is tailored specifically to one application and does not provide a foundation on which build a variety of distributed applications
RPC emerged out of the need of a general IPC protocol that can be used for designing several distributed applications. RPC is popular for the following reasons
1. Simple call syntax
2. Familiar semantics
3. Has a specification of a well-defined interface which supports compile time type checking.
4. Ease of use
5. Generality(it generalizes depending on the environment)
6. Efficiency due to rapid communication
The RPC Model
It is similar to the ordinary procedure call model and works as follows
1. The column places arguments to the procedure.
2. Control is transferred to a sequence of instructions.
3. The procedure is executed in a new execution environment.
4. caller
caller
callee
callee
The control returns to the caller.
Request message contains remote procedure paramenrs.
Request message contains remote procedure paramenrs.
Send reply and wait for request
Send reply and wait for request
Receive request and start execution
Receive request and start execution
Call procedure and wait for reply
Call procedure and wait for reply
Resume execution
Resume execution
Reply msg containing result of procedure execution
Reply msg containing result of procedure execution
RPC is different from a local procedure call because
1. The procedure is executed in a different environment
2. It is more prone to failures.
3. It consumes more time in terms of execution and transmission time
IMPLEMENTATION OF RPC
The implementation is based on Stubs concepts that provide a normal procedure call by concealing from the programs interface the underlying RPC system.
It has five elements
1. Clients process
2. Clients Stub(client procedure)
3. Server process
4. Server...