Tdts04 Fo11 Ds Objektbaserade Och Corba

8/3/2019 Tdts04 Fo11 Ds Objektbaserade Och Corba

1/4

1

Tanenbaum & Van Steen, Distributed Systems:Principles and Paradigms, 2e, (c) 2007 Prentice-Hall, Inc. Allrights reserved. 0-13-239227-5

DISTRIBUTED SYSTEMSPrinciples and Paradigms

Second Edition

ANDREW S. TANENBAUMMAARTEN VAN STEEN

Chapter 10DISTRIBUTED

OBJECT-BASED SYSTEMS


Distributed Objects

Figure 10-1. Common organization of a remoteobject with client-side proxy.


Example: Enterprise Java Beans

Figure 10-2. Generalarchitecture of anEJB server.


Four Types of EJBs

Stateless session beans

Stateful session beans

Entity beans

Message-driven beans


Object Adapter

Figure 10-5.Organization of anobject serversupporting differentactivation policies.


Binding a Client to an Object

Figure 10-7. (a) An example with implicit binding usingonly global references. (b) An example with

explicit binding using global and local references.


2/4

2


Parameter Passing

Figure 10-8. The situation when passing anobject by reference or by value.


Object-Based Messaging (1)

Figure 10-9. CORBAs callback model forasynchronous method invocation.


Object-Based Messaging (2)

Figure 10-10. CORBAs polling model forasynchronous method invocation.


CORBA Object References

Figure 10-11. The organization of an IOR withspecific information for IIOP.


Synchronization

Figure 10-14. Differences in control flow for locking objects


Replication and consistency:

Totally Ordered Multicasting

Figure 6-11. Updating a replicated database andleaving it in an inconsistent state.


3/4

3


Consistency and Replication:Primary-Based, Remote-Write Protocols

Figure 7-20. The principle of a primary-backup protocol.


Consistency and replication:

Primary-Based, Local-Write Protocols

Figure 7-21. Primary-backup protocol in which the primarymigrates to the process wanting to perform an update.


Granularity of invocations:Entry Consistency

Figure 10-15. Deterministic thread schedulingfor replicated object servers.


Replication Frameworks (1)

Invocations to objects are intercepted at three

different points:

At the client side just before the invocation is

passed to the stub.

Inside the clients stub, where the interception

forms part of the replication algorithm.

At the server side, just before the object is

about to be invoked.


Interceptors, generally

Figure 2-15. Using interceptors to handleremote-object invocations.


Replication Frameworks (2):Interceptors in EJB

Figure 10-16. A general framework for separating replication

algorithms from objects in an EJB environment.


4/4

4


Replicated Invocations (1)

Figure 10-17. The problem of replicated method invocations.



Figure 10-18. (a) Forwarding aninvocation request from a

replicated object to anotherreplicated object.



Figure 10-18. (b) Returning areply from one replicatedobject to another.


Example: Fault-Tolerant CORBA

Figure 10-19. A possible organization of anIOGR for an object group having a primary and backups.


An Example Architecture

Figure 10-20. An example architecture of afault-tolerant CORBA system.


Example: Fault-Tolerant Java

Causes for nondeterministic behavior:

1. JVM can execute native code, that is, code that isexternal to the JVM and provided to the latter

through an interface.

2. Input data may be subject to nondeterminism.

3. In the presence of failures, different JVMs will

produce different output revealing that themachines have been replicated.

Documents

Tdts04 Fo11 Ds Objektbaserade Och Corba