IBM Workplace Client Technology Micro Edition Version 5.7.1: Application Development and Case Study

112 WCTME: Application Development and Case Study

Queues

A queue is where messages are stored before they are sent to their destination,

and after they have arrived there. A program can define its own queues or use

the standard queues provided by MQe.

A queue can be local or remote. A

local queue resides on the same system as

the program sending messages to it. It could be used, for example, to send

messages between programs running on the same computer. A

remote queue

exists on another computer, but MQe stores messages locally before forwarding

them to their destinations.

A remote queue can be synchronous or asynchronous. A

synchronous queue

requires a connection to the actual queue. An

asynchronous queue allows

messages to be sent even if there is no active connection. Messages are stored

until they are delivered.

Queue managers

Queue managers manage the queues and take care of details of sending and

receiving messages. Your program interacts with the queue manager, rather than

the queues.

In MQe, there can only be one queue manager for each JVM.

Connections

A connection acts as a pipe between queue managers, allowing messages to

flow between computers. A connection identifies the target queue manager, the

IP address and the port number of the listener for the queue manager.

Before you can define a remote queue, you must define a connection to the

remote queue manager. The connection does not have to be active. You do not

have to have a link to the other queue manager, but you must define the

connection.

Listeners

In order for a queue manager to receive messages from another, remote queue

manager, it must have defined a listener. A

listener identifies the port number

where the system will listen for incoming messages. The port number of the

listener must match the port number of any connections to that queue manager.

5.1.3 Message flow in MQe

In order to see how these pieces fit together, we can look at a simple scenario

(Figure 5-1 on page 113) where program A, running on a computer identified as

Client, with an IP address of 192.168.0.100, wants to send a message to

Chapter 5. Messaging 113

program B, which is running on a computer identified as Server, with an IP

address of 192.168.0.1.

The queue manager on Server is QM1. The queue on QM1 is Q1.

Figure 5-1 Simple Message Flow in MQe

Program A begins by defining a connection to QM1. It knows that QM1 is on

192.168.0.1, and will be listening on port 8082. Next, program A defines Q1 as a

remote queue on QM1.

Meanwhile, program B defines Q1 on QM1. On QM1, Q1 is a local queue. On

QM2, it is a remote queue. Program B then defines a listener for QM1, listening

on port 8082. Finally, it defines a message listener for Q1. The message listener

is a Java interface with a method, messageArrived() which will be called

whenever a message arrives on Q1.

Program A builds a message, puts data into it, and calls QM2 to send the

message to Q1. If there is no connection between the systems, MQe stores the

message in the remote queue on the client. The message will be sent later when

a connection exists.

When the message arrives on QM1, MQe calls the message listener. Program B

can then receive the message and process it.

Note that if program B wants to send a reply back to program A, it would have to

know the queue manager and queue name. It would also have to have

connection to QM2, and QM2 would have to be a listener. So, program B would

also have to know the IP address and port number for the listener. In other

words, the program running on the client would have to set up an MQe server.

Server

QM1

Q1

Client 192.168.0.100

Server 192.168.0.1

Listener

Port:8082

Message

Listener

QM2

QM1/Q1

Connection

QM1

192.168.0.1

Port:8082

Program A