Lecture 8-1 Lecture 8-1
Computer Science 425
Distributed Systems
CS 425 / ECE 428
Fall 2013
Indranil Gupta (Indy)
September 19, 2013
Lecture 8
RPCs and Distributed Objects
Reading: Section 4.3, parts of Chapter 5
2013, I. Gupta, K. Nahrtstedt, S. Mitra, N. Vaidya, M. T. Harandi, J. Hou
Lecture 8-2
HW1s due now
Lecture 8-3
RMI/RPC - Motivation
You write a program where objects call each other
Works well if the program runs on one process
What if you split your objects across multiple
processes?
Can Object1s still call Object2.MethodA()?
Why (not)?
Solution
RMIs: Remote Method Invocations (Object-based)
RPCs: Remote Procedure Calls (non-Object-based)
Access libraries of reusable code across hosts
Pros
Supports code reuse
Standard interface, independent of applications and OS’s
Lecture 8-4
Middleware Layers
Applications
Middleware
layers=
Provide
support to the
application
Run at all servers
@user level
Request reply protocol
External data representation
Operating System
RPCs and RMIs, e.g., CORBA
RPC = Remote Procedure Call
RMI=Remote Method Invocation
CORBA=Common Object Request Brokerage Architecture
Lecture 8-5
Local Objects
Within one process address space
Object
consists of a set of data and a set of methods.
E.g., C++ object, Java object.
Object reference
an identifier via which objects can be accessed.
i.e., a pointer (e.g., virtual memory address within process)
Interface
provides a definition of the signatures of a set of methods (i.e.,
the types of their arguments, return values, and exceptions)
without specifying their implementation.
Lecture 8-6
Remote Objects
May cross multiple process address spaces
Remote method invocation
method invocations between objects in different processes
(processes may be on the same or different host).
Remote Procedure Call (RPC): procedure call between
functions on different processes in non-object-based system
Remote objects
objects that can receive remote invocations.
Remote object reference
an identifier that can be used globally throughout a distributed
system to refer to a particular unique remote object.
Remote interface
Every remote object has a remote interface that specifies
which of its methods can be invoked remotely. E.g., CORBA
interface definition language (IDL).
Lecture 8-7
A Remote Object and Its Remote Interface
interface
remote
m1
m2
m3
m4
m5
m6
Data
implementation
remote object
{
of methods
Example Remote Object reference=(IP,port,objectnumber,signature,time)
Lecture 8-8
Remote and Local Method Invocations
invocation invocation
remote
invocation
remote
local
local
local
invocation
invocation
A
B
C
D
E
F
Local invocation=between objects on same process.
Has exactly once semantics
Remote invocation=between objects on different processes.
Ideally also want exactly once semantics for remote invocations
But difficult (why?)
Process
Object
Process
Process
Host A
Host B
Lecture 8-9
Failure Modes of RMI/RPC
Execute
Reply
correct
function
Execute,
Crash
Request
Crash
Request
Request
Execute
Reply
Execute
Reply
crash
before
reply
crash
before
execution
lost
request
Channel
fails
during
reply
Client
machine
fails
before
receiving
reply
(and if request is received more than once?)
Lecture 8-10
Invocation Semantics
Fault tolerance measures
Invocation
semantics
Retransmit request
message
Duplicate
filtering
Re-execute procedure
or retransmit reply
No
Yes
Yes
Not applicable
No
Yes
Not applicable
Re-execute procedure
Retransmit old reply At-most-once
At-least-once
Maybe
Whether or not to
retransmit the request
message until either
a reply is received or
the server is assumed
to be failed
when retransmissions
are used, whether to
filter out duplicate
requests at the server.
whether to keep a
history of result
messages to enable
lost results to be
retransmitted without
re-executing the
operations
Java RMI,
CORBA
CORBA
Sun RPC
Transparency=remote invocation has same behavior as
local invocation
[Birrell and Nelson, inventors of RPC, 1984]
Very difficult to implement in asynchronous network…
(ok for idempotent operations)
Idempotent=same result if applied repeatedly, w/o side effects
Lecture 8-11
Proxy and Skeleton in Remote Method
Invocation
object A
object B
skeleton
Request
proxy for B
Reply
Communication
Remote
Remote reference
Communication
module module
reference module
module
for B’s class
& dispatcher
remote
client
server
Process P1
Process P2
MIDDLEWARE
Lecture 8-12
Proxy and Skeleton in Remote Method
Invocation
object A
object B
skeleton
Request
proxy for B
Reply
Communication
Remote
Remote reference
Communication
module module
reference module
module
for B’s class
& dispatcher
remote
client
server
Process P1 (client)
Process P2 (server)
Lecture 8-13
Proxy
Is responsible for making RMI transparent to
clients by behaving like a local object to the
invoker.
The proxy implements (Java term, not literally) the methods in
the interface of the remote object that it represents. But,…
Instead of executing an invocation, the proxy
forwards it to a remote object.
On invocation, a method of the proxy marshals the following
into a request message: (i) a reference to the target object, (ii)
its own method id and (iii) the argument values. Request
message is sent to the target, then proxy awaits the reply
message, un-marshals it and returns the results to the invoker.
Invoked object unmarshals arguments from request message,
and when done marshals return values into reply message.
Lecture 8-14
Marshalling & Unmarshalling
A x86 (Windows) client sends an RMI to a PowerPC (e.g.,
Unix/Mac) server
won’t work because x86is little endian while PowerPC is big-endian
Big endian: 1234 is stored as 1234
External data representation: an agreed, platform-
independent, standard for the representation of data
structures and primitive values.
CORBA Common Data Representation (CDR)
Allows Windows client (little endian) to interact with Unix server or Mac
server (big endian).
Marshalling: the act of taking a collection of data items
(platform dependent) and assembling them into the external
data representation (platform independent).
Unmarshalling: the process of disassembling data that is in
external data representation form, into a locally interpretable
form.
Lecture 8-15
Remote Reference Module
Is responsible for translating between local and remote
object references and for creating remote object references.
Has a remote object table
An entry for each remote object held by any process. E.g., B at P2.
An entry for each local proxy. E.g., proxy-B at P1.
When a new remote object is seen by the remote reference
module, it creates a remote object reference and adds it to
the table.
When a remote object reference arrives in a request or reply
message, the remote reference module is asked for the
corresponding local object reference, which may refer to
either to a local proxy or a remote object.
In case the remote object reference is not in the table, the
RMI software creates a new proxy and asks the remote
reference module to add it to the table.
Lecture 8-16
Proxy and Skeleton in Remote Method
Invocation
object A
object B
skeleton
Request
proxy for B
Reply
Communication
Remote
Remote reference
Communication
module module
reference module
module
for B’s class
& dispatcher
remote
client
server
Process P1 (client)
Process P2 (server)
Lecture 8-17
What about Server Side?
Dispatcher and Skeleton
Each process has one dispatcher. And a skeleton
for each local object (actually, for the class).
The dispatcher receives all request messages
from the communication module.
For the request message, it uses the method id to select the
appropriate method in the appropriate skeleton, passing on the
request message.
Skeleton implements the methods in the remote
interface.
A skeleton method un-marshals the arguments in the request
message and invokes the corresponding method in the local
object (the actual object).
It waits for the invocation to complete and marshals the result,
together with any exceptions, into a reply message.
Lecture 8-18
Summary of Remote Method Invocation (RMI)
Object A
Object
B
Comm.
Module
Comm.
Module
Skeleton
for Bs
Class
Server
Process
Client
Process
Proxy
Object
B
Remote
Reference
Module
Dispatcher
Proxy object is a hollow
container of Method
names.
Remote Reference
Module translates
between local and
remote object
references.
Dispatcher sends the
request to Skeleton
Object
Skeleton unmarshals
parameters, sends it
to the object, &
marshals the results
for return
Remote
Reference
Module
MIDDLEWARE
Lecture 8-19
Generation of Proxies, Dispatchers and
Skeletons
Programmer only writes object implementations
and interfaces
Proxies, Dispatchers and Skeletons generated
automatically from the specified interfaces
In CORBA, programmer specifies interfaces of
remote objects in CORBA IDL; then, the interface
compiler automatically generates code for
proxies, dispatchers and skeletons.
In Java RMI
The programmer defines the set of methods offered by a
remote object as a Java interface implemented in the remote
object.
The Java RMI compiler generates the proxy, dispatcher and
skeleton classes from the class of the remote object.
Lecture 8-20
Binder and Activator
Binder: A separate service that maintains a table containing
mappings from textual names to remote object references.
(sort of like DNS, but for the specific middleware)
Used by servers to register their remote objects by name. Used by
clients to look them up. E.g., Java RMI Registry, CORBA Naming Svc.
Activation of remote objects
A remote object is active when it is available for invocation within a
running process.
A passive object consists of (i) implementation of its methods; and (ii)
its state in the marshalled form (a form that is shippable).
Activation creates a new instance of the class of a passive object and
initializes its instance variables. It is called on-demand.
An activator is responsible for
» Registering passive objects at the Binder
» Starting named server processes and activating remote objects in
them.
» Keeping track of the locations of the servers for remote objects it
has already activated
E.g., Activator=Inetd, Passive Object/service=FTP (invoked on demand)
Lecture 8-21
Etc.
Persistent Object = an object that survives
between simultaneous invocation of a process.
E.g., Persistent Java, PerDIS, Khazana.
If objects migrate, may not be a good idea to have
remote object reference=(IP,port,…)
Location service= maps a remote object reference to its likely
current location
Allows the object to migrate from host to host, without
changing remote object reference
Example: Akamai is a location service for web objects. It
migrates web objects using the DNS location service
Lecture 8-22
Remote Procedure Call (RPC)
Similar to RMIs, but for non-OO/non-object-based
scenarios
Procedure call that crosses process boundary
Client process calls for invocation of a procedure
at the server process.
Semantics are similar to RMIs at least once, at most once,
maybe
Format of the message is standard (marshaled), uses request-
reply
Lecture 8-23
Client and Server Stub Procedures in RPC
client
procedure
Request
Reply
Communication Communication
module module
dispatcher
service
client stub
server stub
procedure
procedure
client process server process
procedure
Lecture 8-24
Stubs
Stubs are generated automatically from interface
specifications.
Stubs hide details of (un)marshalling from
application programmer & library code developer.
Client Stubs perform marshalling into request
messages and unmarshalling from reply messages
Server Stubs perform unmarshalling from request
messages and marshalling into reply messages
Stubs also take care of invocation
Lecture 8-25
The Stub Generation Process
Interface
Specification
Stub
Generator
Server
Stub
Common
Header
Client
Stub
Client
Source
RPC
LIBRARY
Server
Source
Compiler / Linker
RPC
LIBRARY
Client
Program
Server
Program
Compiler / Linker
e.g., in SUN XDR
e.g., rpcgen
gcc
.o, .exe
.o, .exe
.c
.c
.c
.c
.h
gcc
Lecture 8-26
Summary
Local objects vs. Remote objects
RPCs and RMIs
RMI: invocation, proxies, skeletons, dispatchers
Binder, Activator, Persistent Object, Location
Service
Next week: Leader Election and P2P systems. See
readings on course page.
HW2 released.
MP2 released.
Lecture 8-27 Lecture 8-27
Optional Slides
Lecture 8-28
Files Interface in Sun XDR
const MAX = 1000;
typedef int FileIdentifier;
typedef int FilePointer;
typedef int Length;
struct Data {
int length;
char buffer[MAX];
};
struct writeargs {
FileIdentifier f;
FilePointer position;
Data data;
};
struct readargs {
FileIdentifier f;
FilePointer position;
Length length;
};
program FILEREADWRITE {
version VERSION {
void WRITE(writeargs)=1; 1
Data READ(readargs)=2; 2
}=2;
} = 9999;
Program number
Version number
Only one argument allowed
Can specify as struct
Available with most Sun systems, and NFS
Lecture 8-29
Finding RPCs
Client
Program
Server
procedure
Server
Stub
Client
Stub
Comm.
Module
Comm.
Module
Dispatcher
SERVER
CLIENT
Finding An RPC:
RPCs live on specific hosts
at specific ports.
Port mapper on the host
maps from RPC name to
port#
When a server process is
initialized, it registers its
RPCs (handle) with the port
mapper on the server
A client first connects to
port mapper (daemon on
standard port) to get this
handle
The call to RPC is then
made by connecting to the
corresponding port
Lecture 8-30
Dealing Room System
Dealer’s computer
Information
provider
Dealer
External
source
External
source
Information
provider
Dealer
Dealer
Dealer
Notification
Notification
Notification
Notification
Notification
Notification
Notification
Notification
Dealer’s computer
Dealer’s computer
Dealer’s computer
Notification
Notification
At each dealer:
One object per
stock type
of interest
[Publish-Subscribe System]
e.g, stock market
Lecture 8-31
Architecture for Distributed Event
Notification
subscriber observer object of interest
Event service
object of interest
object of interest observer
subscriber
subscriber
3.
1.
2.
notification
notification
notification
notification