Remote Procedure Calls
Adapted from:
Paul Krzyzanowski
[email protected]
[email protected]
Except as otherwise noted, the content of this presentation is licensed under the Creative Commons
Attribution 2.5 License.
Page 1
Network Transfer Protocols
Connection Oriented
• often reliable, stream based
• analogous to making a direct phone call
• TCP (Transmission Control Protocol) is a connection-based
protocol that provides a reliable flow of data between two
computers
Connectionless
• unreliable, datagram based
• analogous to sending letters via the postal service
• UDP (User Datagram Protocol) is a protocol that sends
independent packets of data, called datagrams, from one
computer to another with no guarantees about arrival. UDP is
not connection-based like TCP.
Page 2
Internet Protocol and the TCP/IP Protocol Suite
• IP (Internet Protocol) provides an unreliable,
connectionless datagram delivery service.
• TCP/IP is a set of protocols created specifically to allow
development of network and internetwork communications
on a global scale.
• TCP/IP is the most commonly used protocols within the
internet.
• TCP/IP is normally considered to be a four-layer system
Host A Host B
Client Server
Network
TCP IP Driver Driver IP TCP
Page 3
Sockets
• Sockets ( Berkley sockets) are one of the most widely used communication APIs.
• A socket is an object from which messages and are sent and received.
• A socket is a network communication endpoint.
• In connection-based communication such as TCP, a server application binds a socket to a
specific port number. This has the effect of registering the server with the system to receive all
data destined for that port. A client can then rendezvous with the server at the server's port, as
illustrated here:
• Data transfer operations on sockets work just like read and write operations on files. A socket is
closed, just like a file, when communications is finished.
• Network communications are conducted through a pair of cooperating sockets, each known as
the peer of the other.
• Processes connected by sockets can be on different computers (known as a heterogenous
environment) that may use different data representations.
• Data is serialized into a sequence of bytes by the local sender and deserialized into a local data
format at the receiving end.
Page 4
Problems with sockets
Sockets interface is straightforward
– [connect]
– read/write
– [disconnect]
BUT … it forces read/write mechanism
– We usually use a procedure call
To make distributed computing look more like
centralized:
– I/O is not the way to go
Page 5
RPC
1984: Birrell & Nelson
– Mechanism to call procedures on other machines
Remote Procedure Call
Goal: it should appear to the programmer
that a normal call is taking place
Page 6
Regular procedure calls
Machine instructions for call & return but the
compiler really makes the procedure call
abstraction work:
– Parameter passing
– Local variables
– Return data
Page 7
Regular procedure calls
You write:
x = f(a, “test”, 5);
The compiler parses this and generates code to:
a. Push the value 5 on the stack
b. Push the address of the string “test” on the stack
c. Push the current value of a on the stack
d. Generate a call to the function f
In compiling f, the compiler generates code to:
a. Push registers that will be clobbered on the stack to save the values
b. Adjust the stack to make room for local and temporary variables
c. Before a return, unadjust the stack, put the return data in a register,
and issue a return instruction
Page 8
Implementing RPC
No architectural support for remote procedure
calls
Simulate it with tools we have
(local procedure calls)
Simulation makes RPC a
language-level construct
instead of an
operating system construct
Page 9
Implementing RPC
The trick:
Create stub functions to make it appear to the user
that the call is local
Stub function contains the function’s interface
Page 10
Remote Procedue Calls
• Enable procedure calls across host boundaries
• Call interfaces are defined using an Interface
Definition Language (IDL)
• RPC compiler generates presentation and
session layer implementation from IDL
Page 11
ISO/OSI Presentation Layer
Resolution of data heterogeneity
Common data Transmission of
representation data declaration
Marshalling and Unmarshalling
Page 12
Marshalling and Unmarshalling
• Marshalling: Disassemble
(encode) data structures into
transmittable form
• Unmarshalling: Reassemble
(decode) transmitted form into
original complex data structure.
Page 13
Method Call vs. Object Request
Caller
Caller Called
Called
Caller
Stub
Stub
Stub
Called
Called
Transport
TransportLayer
Layer (e.g.
(e.g.TCP
TCP or
or UDP)
UDP)
Page 14
Stubs
• Creating code for marshalling and
unmarshalling is tedious and error-prone.
• Code can be generated fully automatically
from interface definition.
• Code is embedded in stubs for client and
server.
• Client stub represents server for client,
Server stub represents client for server.
• Stubs achieve type safety.
• Stubs also perform synchronization.
Page 15
Stub functions
1. Client calls stub (params on stack)
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 16
Stub functions
2. Stub marshals params to net message
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 17
Stub functions
3. Network message sent to server
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 18
Stub functions
4. Receive message: send to stub
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 19
Stub functions
5. Unmarshal parameters, call server func
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 20
Stub functions
6. Return from server function
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 21
Stub functions
7. Marshal return value and send message
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 22
Stub functions
8. Transfer message over network
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 23
Stub functions
9. Receive message: direct to stub
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 24
Stub functions
10. Unmarshal return, return to client code
client functions server functions
server stub
client stub
(skeleton)
network routines network routines
client server
Page 25
Benefits
• Procedure call interface
• Writing applications is simplified
– RPC hides all network code into stub functions
– Application programmers don’t have to worry about
details
• Sockets, port numbers, byte ordering
• RPC: presentation layer in OSI model
Page 26
RPC has issues
Page 27
Parameter passing
Pass by value
– Easy: just copy data to network message
Pass by reference
– Makes no sense without shared memory
Page 28
Pass by reference?
1. Copy items referenced to message buffer
2. Ship them over
3. Unmarshal data at server
4. Pass local pointer to server stub function
5. Send new values back
To support complex structures
– Copy structure into pointerless representation
– Transmit
– Reconstruct structure with local pointers on
server
Page 29
Representing data
No such thing as
incompatibility problems on local system
Remote machine may have:
– Different byte ordering
– Different sizes of integers and other types
– Different floating point representations
– Different character sets
– Alignment requirements
Page 30
Where to bind?
Need to locate host and correct server process
1. Maintain centralized DB that can locate a
host that provides a particular service
(Birrell & Nelson’s 1984 proposal)
2. A server on each host maintains a DB of
locally provided services
Page 31
Transport protocol
Which one?
• Some implementations may offer only one
(e.g. TCP)
• Most support several
– Allow programmer (or end user) to choose
Page 32
When things go wrong
• Local procedure calls do not fail
– If they core dump, entire process dies
• More opportunities for error with RPC:
• Transparency breaks here
– Applications should be prepared to deal with RPC
failure
Page 33
When things go wrong
• Semantics of remote procedure calls
– Local procedure call: exactly once
• A remote procedure call may be called:
– 0 times: server crashed or server process died
before executing server code
– 1 time: everything worked well
– 1 or more: excess latency or lost reply from server
and client retransmission
Page 34
RPC semantics
• Most RPC systems will offer either:
– at least once semantics
– or at most once semantics
• Understand application:
– idempotent functions: may be run any number of
times without harm
– non-idempotent functions: side-effects
Page 35
More issues
Performance
– RPC is slower … a lot slower
Security
– messages visible over network
– Authenticate client
– Authenticate server
Page 36
Programming with RPC
Language support
– Most programming languages (C, C++, Java, …) have
no concept of remote procedure calls
– Language compilers will not generate client and
server stubs
Common solution:
– Use a separate compiler to generate stubs (pre-
compiler)
Page 37
Interface Definition Language
• Allow programmer to specify remote
procedure interfaces
(names, parameters, return values)
• Pre-compiler can use this to generate client
and server stubs:
– Marshaling code
– Unmarshaling code
– Network transport routines
– Conform to defined interface
• Similar to function prototypes
Page 38
RPC compiler
client
client code
code (main)
(main)
client
client stub
stub
data
data conv.
conv. compiler
compiler client
client
RPC
RPC
IDL
IDL compiler
headers
headers
compiler
data
data conv.
conv. compiler
compiler server
server
server
server skeleton
skeleton
server
server functions
functions
Code you write
Code RPC compiler generates
Page 39
Writing the program
Client code has to be modified
– Initialize RPC-related options
• Transport type
• Locate server/service
– Handle failure of remote procedure call
Server functions
– Generally need little or no modification
Page 40
RPC API
What kind of services does an RPC system need?
• Name service operations
– Export/lookup binding information (ports, machines)
– Support dynamic ports
• Binding operations
– Establish client/server communications using
appropriate protocol (establish endpoints)
• Endpoint operations
– Listen for requests, export endpoint to name server
Page 41
RPC API
What kind of services does an RPC system need?
• Security operations
– Authenticate client/server
• Internationalization operations
• Marshaling/data conversion operations
• Stub memory management
– Dealing with “reference” data, temporary buffers
• Program ID operations
– Allow applications to access IDs of RPC interfaces
Page 42
