IBM Branch Transformation Toolkit 5.1 Migration and Usage Guidelines

Running H/F 1 461

* @see com.ibm.dse.base.OperationXValidate#validate(java.lang.String,

com.ibm.dse.base.DataField, com.ibm.dse.base.Context)

*/

public void validate(String arg0, DataField arg1, Context arg2)

throws DSETypeException {

// TODO Auto-generated method stub

}

}

4. Modify withdrawalClientOperation defined in Java Perspective →

BTTBankApplicationClient → appClientModule → dseoper.xml to include

the reference to this validation class. See Example 8-49.

Example 8-49 withdrawalClientOperation

<operation

context="withdrawalClientCtx"

id="withdrawalClientOp"

serverOperation="withdrawalServerOp"

impClass="btt.bank.client.appl.WithdrawalClientOperation"

xVal="labs.WithdrawalValidate">

......

8.4.6 Creating the invoker

Instantiated by the Bean Invoker Factory, invokers enables Struts actions or

Java request handlers to access business processes and Single Action EJBs

through an EJB call.

When a request comes from a requester, which could be a request handler or a

Struts action, the request brings a request ID and a session ID to the Bean

Invoker Factory. The request ID indicates what kind of transaction the client is

requesting, and the session ID identifies the session of this transaction request.

The Bean Invoker Factory generates or allocates an invoker with the request ID

and session ID. The Bean Invoker Factory then returns the invoker to requester

so that the requester can send the request to the application logic layer.

462 IBM Branch Transformation Toolkit 5.1 Migration and Usage Guidelines

Bean invoker pattern

This section describes some common ways that you can work with invoker

beans.

򐂰 Invoker bean overview

– EJB methods are strongly typed and each EJB has its own corresponding

access information. In order to provide a generic way to access EJB,

invocation architecture is required.

– The logic for creating Bean Invoker and bean invocation logic are

encapsulated within the Bean Invoker. Developers may invoke EJB

through a Bean Invoker with a generic interface.

– The composition of Bean Invoker Pattern includes Bean Invoker Factory,

Basic Invoker Class (super class), end-user extended Bean Invoker class

and Bean Invoker Registry.

The Bean Invoker Pattern is shown in Figure 8-189.

Figure 8-189 Invoker pattern architecture

򐂰 Base invoker class

– This is the abstract class that encapsulates some basic functions about

EJB accessing and cache mechanism.

Bean Invoker Pattern

Invoker Pool

EJB EJB

EJB EJB

Single Action

EJB

Bean

Invoker

Factory

Invoker

Registry

Invoker

Object

Invoker

Object

EJB Proxy Cache

EJB

Proxy

EJB

Proxy

requestID

Running H/F 1 463

– It also provides all the format/unformat functions, for example,

StringFormatter, XMLFormatter, FloatFormatter, and so on.

– Each bean invoker should be extended from the base invoker class. You

should manually modify it for each EJB method accessed.

򐂰 Interface for channels

– Different channels should implement different invoker interfaces. The

defaults are:

• BeanInvokerForJavaRequest:parseRequestData(String requestData)

• BeanInvokerForStrutsAction:parseRequestData(Context requestData)

򐂰 End-user extended invoker

– For each EJB method in Single Action EJB, the end-user must extend the

super invoker class. The logic for accessing the EJB method is defined in

this extended class.

– executeEJB() should be overridden by the end-user.

– End-user extended invoker implements different interfaces according to

the channels that use the Bean Invoker.

򐂰 Bean Invoker Factory

– This is responsible for the Bean Invoker life cycle.

– Only one instance of Bean Invoker Factory should be available in a JVM

(Factory Pattern).

– Bean Invoker Factory maintains the invoker object cache and EJB proxy

object cache.

– Exception handling.

򐂰 Bean Invoker Pool

– It is used to store Bean Invoker Object.

– It improve performance with caching.

򐂰 Bean Proxy Cache

– It stores EJB Proxy Object.

– It improves performance with EJB interaction.

– It is session-relevant.

– Session timeout handling

򐂰 Bean Invoker Registry (see Figure 8-190 on page 464 )

– This is a resource bundle that stores the information that is required for

creating a Bean Invoker.

– Each type of Bean Invoker should have its corresponding registry file.

– It links a request ID to a given resource bundle name.

464 IBM Branch Transformation Toolkit 5.1 Migration and Usage Guidelines

– There are two ways of accessing resource bundle file: ResourceBundle

mode and Properties file mode.

Figure 8-190 Invoker Registry

Creating the Withdrawal invoker

To create the Withdrawal invoker, perform the following tasks:

1. Import a Web project for building the Invoker.

a. From the WebSphere Studio Application Developer menu, select File →

Import.

b. In the dialog box that opens, select WAR file and click Next.

c. From WAR file field, browse to the directory

c:\7160code\cha8\BTTBankAppClientWeb.war.

d. For the Project field, click New... to create a New Dynamic Web Project

named BTTBankAppClientWeb, whose EAR project should be set as

Invoker Registry

BeanInvokerRegistry

Mapper.properties

requestID

#startupServerOp.properties

implClass=com.ibm.dse.samples.appl.StartupServerInvoker

jndiName=ejb/com/ibm/shanjh/cstest/StartupServerActionHome

factory=com.ibm.websphere.naming.WsninitialContextFactory

location=iiop://localhost:2809

homeClassName=com.ibm.cstest.StartupServerActionHome

isLocal=false

HelloWorld=/properties/HelloWorld.properties:PROP

startupServerOp=com.ibm.btt.cs.invoker.base.startupServerOp:RB

Running H/F 1 465

BTTBank by selecting the Configure advanced options check box and

clicking Next, as shown in Figure 8-191.

Figure 8-191 Import EAR and new BTTBankAppClientWeb