Mumbai University > Information Technology > Sem 6 > Distributed System

Marks: 10 Marks

Year: May 2016

In distributed computing a remote procedure call (RPC) is when a computer program causes a procedure (subroutine) to execute in another address space (commonly on another computer on a shared network), which is coded as if it were a normal (local) procedure call, without the programmer explicitly coding the details for the remote interaction.That is, the programmer writes essentially the same code whether the subroutine is local to the executing program, or remote.

Based on the needs of different systems, several communication protocols have been proposed for use in RPC which are mentioned below:

i.The Request Protocol

• This protocol is also called as R (request) protocol.
• It is used in RPC when a called procedure has nothing to return as a result of execution and the requirement of client confirmation about procedure execution is not needed.
• As no acknowledgement or reply message is involved, only single message is transmitted from client to server.
• The client proceeds after the request message is sent as there is no reply message.
• This protocol provides May-be call semantics and does not need retransmission of request messages.
• RPC that uses the R protocol is known as asynchronous RPC which helps to improve the combined performance of the client and server. This is done because the client does not ait for a reply and server does not need to send a reply.
• For an asynchronous RPC, the RPCRuntime does not retry a request in case of communication failure. TCP is better alternative then UDP since no retransmission is required and it is connection oriented.
• Asynchronous RPC with unreliable transport protocol are generally used in implementing periodic update services. Distributed system window is one of its applications.

ii.The Request/Reply protocol

• This protocol is also known as RR(request/reply) protocol.
• It is useful for designing systems which involve simple RPCs.
• In a simple RPC all the arguments and result fit in a single packet buffer while the call duration and intervals between calls are short.
• This protocol is based on the idea of using implicit acknowledgement to eliminate explicit acknowledgement messages.
• In this protocol a server reply is considered as an ACK for a clients request and a subsequent call from a client is considered as ACK of the client’s previous call.
• Timeout-and-retires technique is used with RR protocol for failure handling. Retransmission of request message is done when there is no response.
• RR protocol and timeout technique provides at-least-once call semantics on if duplicate requests are not filtered out.
• Exactly once semantics are supported by servers using reply cache which stores replies.

iii.The Request/Reply/Acknowledgement-Reply Protocol

• This protocol is also known as RRA (request/reply/acknowledge-reply) protocol.
• RR protocol implements exactly once semantics which requires storage of a lot of information in the server cache and can lead to loss of replies that have not been delivered.
• To overcome the limitations or RR protocol, RSA protocol is used.
• In this clients acknowledge the receipt of reply messages and the server deletes information from its cache only after it receives an acknowledgement from client.
• Sometimes the reply acknowledgement message may get lost therefore RRA protocol needs a unique ordered message identifiers. This keeps a track of the acknowledgement series sent.