1. Communication
Chapter 2

2. IPC Inter-Process Communication is the heart of all DSs.
Processes on different machines.
Always based on low-level message passing.
In this chapter:
RPC
RMI
MOM (Message Oriented MiddleWare)
Streams (due to the advent of Multimedia DSs)

3. Layered Protocols (1)
Layers, interfaces, and protocols in the OSI model.
2-1

4. Layered Protocols Protocol Protocol Stack
Connection Oriented
Connectionless
Protocol Stack
Description of the layers, Unit of exchange.

5. Layered Protocols (2) A typical message as it appears on the network.
2-2

6. Data Link Layer
2-3
Discussion between a receiver and a sender in the data link layer.

7. Transport Protocols
Makes the underlying layers usable by the application layer.
Provide a reliable or unreliable connection for the upper layer.
UDP :: TCP
RTP for real-time systems.

8. Client-Server TCP
Normal operation of TCP.
Transactional TCP.
2-4

9. Middleware Protocols
An adapted reference model for networked communication.
2-5

10. RPC
PC?
R…………….PC?
Simple idea
Complexity in provision

11. Conventional Procedure Call
Parameter passing in a local procedure call: the stack before the call to read
Count = read (fd, buf, nbytes);
The stack while the called procedure is active

12. Issues Calling Method? Call by value Call by reference
Call by Copy/Restore
Call by name

13. Client and Server Stubs
Principle of RPC between a client and server program.
The read stub is called on behalf of the real read procedure!

14. Steps of a Remote Procedure Call
Client procedure calls client stub in normal way
Client stub builds message, calls local OS
Client's OS sends message to remote OS
Remote OS gives message to server stub
Server stub unpacks parameters, calls server
Server does work, returns result to the stub
Server stub packs it in message, calls local OS
Server's OS sends message to client's OS
Client's OS gives message to client stub
Stub unpacks result, returns to client

15. Passing Value Parameters (1)
Steps involved in doing remote computation through RPC
It works fine, while the scenario is simple and straightforward; but ….
2-8

16. Passing Value Parameters (2)
Different character set standards (ASCII vs EBCDIC)
Little-Endian vs Big-Endian Architecture.
Original message on the Pentium (L. E.)
The message after receipt on the SPARC (B. E.)
The message after being inverted. The little numbers in boxes indicate the address of each byte

17. Call by Reference Parameter Passing
???

18. Parameter Specification and Stub Generation
Both sides should agree on the content of passing data structures.
Example in the next slide.
The way a message including the parameters is interpreted is the main issue!!
Client and server should agree on the representation of simple data structures.
Agreement on the actual exchange of the messages (connection-oriented or connection-less)

19. Parameter Specification and Stub Generation
A procedure
The corresponding message.
Interface Definition Language  compiling into client stub and server stub

20. Extended RPC Models
RPC becoming as de facto standard for comm. in DSs.
Popularity due to simplicity.
Two extensions
Doors
Async RPC.

21. Doors Equivalent to RPC for processes located on the same machine.
A door is a name for a procedure in the address space of a server process, called by colocated processes within the server.
Idea was originally from the Spirit OS (1994)
Same as LightWeight RPC.
The server process must register a door before use (calling door-create)

22. Doors
The principle of using doors as IPC mechanism.

23. Asynchronous RPC (1)
2-12
The interconnection between client and server in a traditional RPC
The interaction using asynchronous RPC

24. Asynchronous RPC (2)
A client and server interacting through two asynchronous RPCs
2-13

25. Writing a Client and a Server
The steps in writing a client and a server in DCE RPC.
2-14

26. Binding a Client to a Server
Client-to-server binding in DCE.
2-15

27. Performing an RPC The whole scenario! Semantics Idempotency
At-most-once operation
Idempotency

28. Remote Object Invocation
OO technology in centralized systems.
Promoting the idea of RPC to the OO technology.
Proxy as the client delegate == Client stub.
Skeleton == server stub
The object state is normally not distributed  remote object instead of distributed object

29. Distributed Objects
Common organization of a remote object with client-side proxy.
2-16

30. Message-Oriented Communication
Sometimes both RPC and RMI is not appropriate
Synchronous nature of RPC and RMI!
 Messaging.

31. Berkeley Sockets (2)
Connection-oriented communication pattern using sockets.

32. The Message-Passing Interface (MPI)
Some of the most intuitive message-passing primitives of MPI.
Primitive
Meaning
MPI_bsend
Append outgoing message to a local send buffer
MPI_send
Send a message and wait until copied to local or remote buffer
MPI_ssend
Send a message and wait until receipt starts
MPI_sendrecv
Send a message and wait for reply
MPI_isend
Pass reference to outgoing message, and continue
MPI_issend
Pass reference to outgoing message, and wait until receipt starts
MPI_recv
Receive a message; block if there are none
MPI_irecv
Check if there is an incoming message, but do not block

33. Stream-Oriented Communication
Till now, focus was on exchanging one or more independent and complete units of info.
However, consider an audio stream, CD quality is also required  the original sound has been sampled at Hz a sample in each 1/44100 Sec is required to re-produce the original sound.
Time-dependent and continuous media is required :: Temporal relationship between data items are crucial.

34. Data Stream (1)
Setting up a stream between two processes across a network.
Data stream is a sequence of data units.

35. Transmission Modes Async Trans Mode: Sending regardless of time
Synch Trans Mode: There is a max end-to-end delay for each unit: Sensor info!
Isochronous Trans Mode: Data units should be transferred on time:: A max and min end-to-end delay (bounded jitter).

36. Data Stream (2) Setting up a stream directly between two devices.
2-35.2

37. Data Stream (3)
An example of multicasting a stream to several receivers.

38. QoS Time-Dependent requirement:: QoS
Next slide as a sample QoS specification
Formulation based on the token bucket algorithm
Basic idea is that tokens are generated at a constant rate.
Token is a fixed # of bytes, an application is allowed to pass to the network.

39. Specifying QoS (2)
The principle of a token bucket algorithm.

40. Setting Up a Stream
The basic organization of RSVP (Resource reSerVation Protocol) for resource reservation in a distributed system.

41. Synchronization Mechanisms (1)
The principle of explicit synchronization on the level data units.

42. Synchronization Mechanisms (2)
The principle of synchronization as supported by high-level interfaces.
2-41

43. End of Chapter 2