Invoking EJB's from Oracle ESB using WSIF

A common requirement with the Oracle ESB is to use it to invoke one or more EJBs; one way to do this is to create a standard SOAP service from an existing EJB using the wizards in JDeveloper.

Whilst this is straight forward, it has the disadvantage that you introduce the overhead of using SOAP/HTTP as well as losing the ability to invoke the EJB as part of a wider transaction.

An alternative approach to SOAP, is to use WSIF (Web Services Invocation Framework) which allows us to expose the EJB through a standard WSDL interface but bind to it a run time using native bindings (i.e. RMI) as opposed to SOAP. This provides us with significantly improved performance and enables us to include the invocation of the EJB within a distributed transaction.

This document covers how to go invoke an EJB 3.0 from the ESB via WSIF. For the purpose of this document, we will create a simple EJB (Greeting) with a single method helloWorld. However to make this example a bit more a bit more realistic, we will pass in a complex type Person, which has the properties title, firstName and lastName.

Note: For the sake of this example, we have assumed you have created an application with JDeveloper containing a single project called Greeting.

Creating the WSIF WSDL File

Before you can invoke an EJB via WSIF you need to create a WSDL file which contains the appropriate binding information required by the WSIF framework to invoke the EJB.

There are two basic approaches you can take:

• Contract First – With this approach we start with by defining an abstract WSDL document which defines the service and its corresponding operations and then write an EJB which implements this contract. This EJB will typically act as a wrapper to one or more existing EJBs.

• EJB First – Here we start with one or more EJB and create a WSDL interface based on the EJB and it’s operations, and then the ESB or BPEL to assemble these operations into a meaningful service.

For the sake of this article we are going to take the contract first approach as this is generally accepted as best practice. In order to do this we will need to carry out the following steps:

• Define the abstract WSDL for the required service
• Implement and deploy the corresponding EJB
• Add the WSIF bindings for the EJB to our abstract WSDL.
• Implement and Deploy an ESB project to invoke the EJB

We cover each of these steps in detail below.

Define Abstract WSDL File

The first step is to define a simple abstract WSDL file to describe your services, for our purpose we have defined the following:


At first glance there is nothing here to indicate that the WSDL is for a service to be implemented using an EJB. The only thing worth mentioning is that we have defined our XML schema elements for our input and outputs parameters in a separate file, which we have imported into our WSDL document.

The reason for doing this it that at a later stage we will need to generate Java classes based on the XML Schema, and separating out the schema makes this process slightly simpler, especially if you need share the Schema between multiple EJBs.

The schema for Person.xsd is as follow:


Note: We have defined all our parameters as elements and NOT complexTypes since the Oracle ESB will complain if you try and use a WSDL which uses complexTypes in its message definitions.

Implement Session EJB

The next step is to implement a corresponding stateless session EJB to implement our Greeting service. The simplest way to do this in JDeveloper is right click your Greeting project and select New. This will bring up the ‘Gallery’ as shown below. Browse to the Business Tier-> EJB section and select Session Bean (as shown below) and click OK.

Figure 1 - Creating a Session Bean

This will launch the Create Session Bean Wizard. In Step 1, select ‘Enterprise Java Beans 3.0(J2EE 5.0)’ as the version of EJB you wish to use and click next.

In step 2; specify the EJB Name, i.e. Greeting, Keep the default settings for the Session EJB 3.0 options (as shown below) and click ‘Next’.

Figure 2 – Specify EJB Name and Options

Next specify the bean class name. This will have already been defaulted based on the EJB Name you specified in the previous step (i.e. it will have appended bean to it). Leave the class name as is, but specify the package name as appropriate. In our case we set the package name to com.bpelpeople.ejb and then click next.

Figure 3 – Specify Class Name

Finally specify that you want the EJB to implement a remote, local and web service endpoint interface (as shown below). Then hit ‘Finish’ to generate your EJB.

Figure 4 – Specify EJB Interfaces

For each operation defined in our WSDL file we should create the equivalent method in our EJB. For each method we need to define its input and return parameters. For operations or method that return simple XML types (e.g. xsd:string, xsd:integer) we can use the equivalent type in Java.

However for complex types, such as our Person element we need to implement the appropriate classes required for serialization/de-serialization between java and xml.

Using schemac to generate serialization/de-serialization classes

As part of the SOA Suite, Oracle provides the schemac utility (bundled with BPEL PM) which you can use to generate the serialization/de-serialization classes. To use this utility open the BPEL Developer Prompt and then run the following command:

schemac –noCompile –sourceOut < dir=""> < file="">

This will generate the source Java for serialization/de-serialization classes for the specified schema file (e.g. Person.xsd in our case) and places these within the specified source directory. Within JDeveloper you will need to import these classes into you project containing your EJB.

You will notice that for each element, schemac will generate three classes (e.g. Person, IPerson and PersonFactory); we will use the actual concrete class (i.e. Person) for the input parameter to our method.

So for our Greeting EJB we have defined the following method:

Setting Project Classpath

Once you have imported the generated Java classes into your project, in order to compile them you will need to add the orabpel.jar file to your project classpath, this is located at:

$SOA_HOME/bpel/lib/orabpel.jar

Create Deployment Descriptor

Finally we need to create a deployment descriptor for our EJB, within the Application Navigator right click on our Greeting project and select ‘New’. From the gallery, browse to General -> Deployment Profiles and select ‘EJB Jar File’. This will bring up the Create Deployment Profile window.


Figure 7 – Specify Deployment Profile Name

Specify a profile name, i.e. ‘Greeting’ in our example and click ‘OK’. This will launch the ‘EJB JAR Deployment Profile Properties’ window, as shown below.


Figure 8 – EJB JAR Deployment Profile Properties

Specify a name for the application, in our case we have chosen ‘GreetingApp’ (note: you will need to make a note of this value as you will use it when defining the WSIF Binding for your service) and click ‘OK’.


Note: you will need to define an Application Server connection to the OC4J instance to which you want to deploy your EJB first.

Deploying the EJB

You can now use JDeveloper to deploy your EJB, however before doing this you should define an Application Server connection with JDeveloper to the oc4j instance on which the SOA Suite is deployed (e.g. oc4j_soa).

You are now ready to deploy your EJB. You can now use the deployment file you just created to deploy your EJB. Right click on this file and select deployTo-> oc4j_soa (where oc4j_soa is the name of your application server connction).

As part of the deployment process, JDeveloper will bring up the Configure Application window. However if you click ‘OK’ and follow the standard deployment process, you will need to include the orabpel.jar within your .ear file. This can be a bit cumbersome, particularly if you have several EAR files to deploy.

The other option is to deploy your application as a child of the orabpel application. By designating orabpel as the parent application, your EJB will inherit the set of shared libraries imported by the parent including orabpel.jar. To do this select ‘GreetinApp’ within the ‘Configure Application’ window and then select orabpel as the parentApp as show below in figure 9.


Figure 9 – Setting the Parent Application

Then click ‘ok’ and JDeveloper will complete the deployment of our EJB.

Adding WSIF Bindings

Now that we have written and deployed our EJB we are ready to add the WSIF bindings to our abstract WSDL file to enable it to be called from the ESB.

Modify Definitions Element

Within the <definitions> element of your WSDL file you need to add the following namespaces:

• xmlns:ejb=”http://schemas.xmlsoap.org/wsdl/ejb/”
• xmlns:format="http://schemas.xmlsoap.org/wsdl/formatbinding/"

The ejb namespace allows the binding of WSDL operations to methods on an EJB class. The format namespace adds support for mapping Java types to XML Schema definitions.

Add Bindings Element

This is where we bind our service to an EJB rather than a standard SOAP service. For our example, the WSDL <binding> element is defined as follows:

Ejb Binding

<ejb:binding> should be the first element within our <bindings> tag and identifies that this is service is bound to an EJB rather than a SOAP service.

Type Definitions

Next, you need to map the XML Schema elements used within the WSLD message definitions to the Java types used in the method invocations for your EJB.

The <format:typemapping> element will contain one <format:typemap> for each xml schema element that we need to map, it has two attributes encoding and style both of which should be set to ‘Java’ to indicate that we are mapping to Java classes.

The <format:typemap> element has two attributes, typeName which hold the name of the xml schema element that we are mapping and formatType which contains the class name of the Java class to which we are mapping it.

In our example, we have specified two type mappings one between our Person element and the corresponding class that we generated using schemac, and the other between the Return element and the java.lang.String class.

Method Mapping

The final step is now to map the EJB method calls onto the WSDL operations. This is done using the <ejb:operation> tag to identify which EJB method should be used to support a given operation

This element has the following attributes:

• methodName – which should be set to the equivalent method within the EJB.
• interface – which should be set to ‘remote’.
• parameterOrder – which should be set to name of the <part> contained with the input message for the operation.
• returnPart – which should be set to the name of the <part> contained with the output message for the operation.

Add Services Binding

Finally we have to add the element to our WSDL to specify where to locate the service. This looks pretty normal, except that instead of a <soap:address> element, we need to specify an <ejb:address> element.

This contains one attribute jndiName; which specifiec the JNDI Name of the deployed EJB. In our example the <service> element is defined as follows:

Calling your EJB from ESB

We are now ready to invoke our EJB from within our ESB. To do this, create a SOAP Service based on our WSDL file within your ESB project in the normal way (note you will need to import the Schema into your ESB project first).

However before registering your ESB project you will need to define the following endpoint properties on your service:

• java.naming.factory.initial - This is used to specify the initial context factory and should be set to com.evermind.server.rmi.RMI InitialContextFactory.
• java.naming.provider.url - Used to specify the URL for the EJB provider, the structure of this is covered below.
• java.naming.security.principal - Specifies the user id to be used to invoke the EJB and should be set to the appropriate value (e.g. oc4jadmin)
• java.naming.security.credentials - Specifies the corresponding password to be used to invoke the EJB and should be set to the appropriate value (e.g. welcome1).

This is so that the ESB is able to locate and invoke the EJB at run time.

java.naming.provider.url

Specifies the URL for the provider (or application) which contains our EJB, this takes the form:

opmn:ormi://<hostname>:<opmn request port>:<oc4j container>/<application>

Where

• <hostname> is the host on which the Oracle Application Server is deployed.
• <opmn request port> is the runtime port for OPMN requests on the Oracle Application Server (e.g. 6003).
• <oc4j container> is OC4J container to which we have deployed our EJB Application (e.g. oc4j_soa).
• <application> is the name of the application in which our EJB is deployed, we specified this as part of our deployment descriptor (i.e. GreetingApp).

Unfortunately JDeveloper won’t allow you to specify these specific properties, so you will need to enter the .esbsvc file outside of JDeveloper.

To do this first close down JDeveloper (otherwise you can get sync issues) and then open the appropriate .esbsvc file in your favourite text editor and specify the endpoint properties at the end of this file as follows:



Once done, you can open up JDeveloper and deploy your ESB project.

Deploy EJB Classes to ESB Engine

Before we can call the EJB from the ESB, we must first deploy the remote interface class of our EJB (Greeting.class) and our Java serialization/de-serialization classes to the ESB engine, the simplest way to do this is to copy the java classes into the directory:

<soa_home>\bpel\system\classes

Once done you will need to re-start the SOA Suite so that it picks up the classes.

Deploy Patch to ESB

Finally if you are using 10.l.3.3 of Oracle SOA Suite you will need to install patch 6314009 which is available at metalink.oracle.com.

SOA Down Under

Well it's been a while since I posted my last blog, and since then a lot has happened!!

On a personal note, I have moved half way round the world, re-locating from the UK to Melbourne, Australia. As you can imagine the process of tearing down your life on one side of the world and then rebuilding it on the other side is quite a time consuming but strangely cathartic activity (as well as leaving little time for blogging).

I've now been in Melbourne for 4 months and am absolutely loving it. I'm still with Oracle, but now working in Product Management for Oracle Fusion Middleware.

From a middleware perspective Oracle has announced the acquisition of BEA as well as released a technical preview of the Oracle SOA Suite 11g the next generation of the Oracle SOA Suite. No doubt I will be writing more about these in future blogs.

Well now that the dust has started to settle (in terms of my move to Australia), I plan to get back to blogging on a more regular basis, so if you have any requests then by all means let me know. In the meantime I'm currently writing one on how to call EJB's via WSIF for the ESB so look out for that one soon!!

Message oriented interoperability between WCF channels and Oracle Application Server WSIF

Jesus Rodriguez, a buddy of mine, has written a number of very interesting articles on web service interoperability between .NET and Oracle Fusion Middleware. A recent one on Message oriented interoperability between WCF channels and Oracle Application Server WSIF is well worth a read.

Querying BPEL Process Instances

The Oracle BPEL Console provides a great tool for monitoring the status of in-flight and completed process.

However it is often a requirement to present this information to a business user within a business specific view. One way of achieving this is to use the Oracle BPEL Portlets.

This is fine where you want to give the user a specific list of process instances, for example a list of all currently running Purchase Order processes. But what about when you need to give them a list more filtered to their specific business requirements, for example:

• How do I find all open purchase order processes for a specific customer?
• How do I find all help desk processes being managed by a specific customer service representative?
• How do I find all open expense processes that are waiting approval?

In addition once you’ve located a process instance, how do I provide the user with access to relevant data contained within the process? Again the BPEL Console provides a mechanism for drilling into the process and looking at the audit trail for the data. But often we want to provide this data in a summarised business view designed specifically for the needs of the business user.

The Oracle BPEL PM Server provides a series of API’s that enable you to meet these requirements.

In fact what is not often realised is that the BPEL Console itself makes use of these API’s, giving you the flexibility to completely re-write the console if that is what’s required!!!

In reality this is rarely if ever the case, typically the requirement is to provide business users with a simplified view tailored to their specific needs. A simple way of achieving this is through the use of these API’s; this is the subject of this article.

Locating a Process Instance

The Oracle BPEL PM Client API provides a Locator class for enabling a client application to search for processes, instances and activities. The Locator class provides a number of constructors, which enable you to connect to a BPEL domain hosted on either a local or remote J2EE Server.

For the purpose of locating specific process instances it provides two very useful methods:

• listInstances(WhereCondition wc)
• listInstancesByIndex(WhereCondition wc)

Each method returns an array of objects of type InstanceHandle, which can then be used to perform operations on the corresponding process instance.

The key parameter for each of these methods is the WhereCondition, which is used to build up a query to restrict which instances are returned by the method.

Note: Where condition objects may be concatenated together to form a larger query. The methods append and prepend allow the user to add a clause (in String format) or even a whole WhereCondition object to the beginning or end of the current where condition.

The following code snippet shows how to construct a WhereCondition to return all running process instances for the “LoanFlowProcess” where it’s current status is “CheckingCredit”.

    String pProcessId = "LoanFlowProcess";
    String pStatus = "CheckingCredit";

    // Constructs a where condition that searches for open instances
    WhereCondition where = WhereConditionHelper.whereInstancesOpen();

    // Extend the where condition to filter on process id
    WhereCondition whereProcessId = new WhereCondition( "process_id = ?" );
    whereProcessId.setString(1, pProcessId);
    where.append(whereProcessId);

    // Extend the condition to filter on processes with the status ‘CheckingCredit’
    WhereCondition whereStatus = new WhereCondition( "status = ‘" + pStatus + " );
    whereStatus.setString(1, pStatus);
    where.append("and").append(whereStatus)

    // Find Instances
    IInstanceHandle[] instanceHandles = locator.listInstances( where );

In the final step, the actual locator class is performing a query against the BPEL Dehydration store, similar to the one illustrated below:

    “select cikey from cube_instance where " + whereCondition.getClause();

We use the WhereCondition (which wraps a SQL prepared statement where condition), in order to restrict the result set returned by the query.

It’s worth exploring in a bit more detail the various parts of the WhereCondition.

For the first we use the WhereConditionHelper class to restrict the query to only currently running processes. This is a simple utility class which provides a variety of Static methods for creating various query fragments (e.g. return process instance whose state is open, completed, aborted, stale, etc) which can then be appended to additional where conditions to create the required query.

For our second condition we are literally adding the condition

    cube_instance.process_id = “LoanFlowProcess”

to our prepared statement. Here you can specify pretty much any of the columns in the database table cube_instance (e.g. Process_Id, Revision_Tag, Priority, Status).

In reality, rather than naming this column explicitly, we should use the appropriate constants defined in com.oracle.bpel.client.util.SQLDefs (e.g. SQLDefs.CI_process_id for our example).

The final statement is similar to the second in that we are querying on the process status. But what is the process status? Well it shouldn’t be confused with process state, which we queried on in the first WhereCondition.

Rather the process status is a variable that keeps track of where in the process a particular process instance is. When the process is first initiated, this value is set to ‘initiated’. This value is then updated every time you enter a new scope within a process to contain the name of the scope.

Note: When you have nested scopes, it will contain the name of the lower most nested scope that the process is in, i.e. it contains the last scope that was entered. Also when a process leaves a scope the status value is NOT reset, i.e. it will still contain the name of the previous scope until it enters a new scope.

Process Indexes

The listInstances method is very useful but it still doesn’t allow us to perform a query based on actual data held in the process instance, e.g. just return the loan flow process for ‘Dave’.

To solve this problem, BPEL allows for a process to have up to 6 indexes and to create a where condition across one or more of these indexes.

Essentially there are two steps to this; first you need to set the index values on the actual process instance; secondly you use the index values in a query to pull back all processes for a particular index value in a similar fashion to above.

Setting the Index Value

The simplest way to achieve this is to embed a piece of Java (using the Java Embed Task) at the start of the process to call the setIndex API to set the index value based on a value in the initial message, as shown in the example below:

    // Set Index1 for Customer Name
    String customerName = ((com.collaxa.cube.xml.dom.CubeDOMText)
    getVariableData("input", "payload", "/auto:loanApplication/auto:customerName/text()")).getText();
    setIndex(1, customerName);

Note: The getVariableData method is used to retrieve the customerName from the “input” variable; the syntax of the parameters is similar to the “from” component within an assign construct.

Querying Processes

The following code snippet shows how to construct a WhereCondition to return all running process instances where index_1 is equal to ‘Dave’.

    String pCustomerName = "Dave";

    // Constructs a where condition that searches on index 1
    WhereCondition where = new WhereCondition(SQLDefs.CX_index_1 + " = ?");
    where.setString(1, pCustomerName);

    // Find Instances
    IInstanceHandle[] instanceHandles = locator.listInstancesByIndex( where );

However there is one minor issue with this; under the covers the listInstances method is performing a query on the cube_instance database table, whilst the listInstancesByIndex is performing a query on the ci_indexes database table.

The issue here is if we want to perform a query that is a join across these two tables, i.e. show me all LoanFlow process for Dave. The WhereCondition API doesn’t (naturally) allow for joins; however there are two possible workarounds:

The first is to set the index values to hold the additional data required by the query, e.g. set index_1 to hold the process name and index_2 to hold the customer id.

The second is to extend the where condition passed to the listInstance method to have an IN condition that queries against the ci_indexes database table, as show below:

    // Extend the where condition to only return open processes with the
    WhereCondition whereIndex = new WhereCondition( "cikey in (select cikey from ci_indexes where index_1 = ?");
    whereIndex.setString(1, pCustomerName);
    where.append("and").append(whereIndex)

Note: I’ve used table and column names for clarity, but in reality you should use the constants defined by SQLDefs.

Using an Index to set status

Earlier in the article we looked ay how we can use process status to keep track of where we currently are in a process (remember status contains the name of the last scope that we entered).

However, whilst this is useful it has a couple of drawbacks; one is that if we use the listInstancesByIndex method to locate a process, we can’t actually filter on the state of the process. However the other is that we rely on insuring that the scopes are correctly named, designed, etc to keep track of where we are in the process. However in reality we may only have a few key milestones that we are interested in, and these may span several scopes or we may have more than one milestone contained within the same scope.

An alternative is to use an Index to hold the status of the process, and just update the status of the process using the setIndex method at appropriate points within the process.

Accessing Process Data

Once we have our list of processes instances, the final stage is to actually access the relevant data contained within the instance to display to the business user.

This is simply the case of iterating through our array of instance handles. Once you have the instanceHandle for a process, you can then use this to access process variables contained within the process instance using the getField method.

However you need to take care that whatever variable you are trying to access is currently visible within the active scope of the process. I find the simplest way to do this is to define a global variable (i.e. define the variable at the process level) and initialise it at the beginning of the process based upon the content of the initial message received by the process. Then during the lifetime of the process update the variable as required to reflect the true state of the process.

The following code snippet shows how we can process each instance returned by the locator and access the variable “LoanApplicationSummary” defined in the BPEL process.

    // Find Instances
    IInstanceHandle[] instanceHandles = locator.listInstancesByIndex( where );

    // Process each instance
    for (int i = 0; i < instanceHandles.length; i++ )
    {
        IInstanceHandle instanceHandle = instanceHandles[ i ];

        // Get Loan Application Summary Variable
        Element loanApplicationElement = (Element) instanceHandle.getField(“LoanApplicationSummary”);

        // Create Loan Application Bean
        LoanApplication loanApplication = LoanApplicationFactory.createFacade(loanApplicationElement);

        // Process Loan Application Bean as required
        …
    }

To access the variable you use getField method on your instance Handle, as show below:

    // Get Loan Application Summary Variable
    Element loanApplicationElement = (Element) instanceHandle.getField(“LoanApplicationSummary”);

This will return a DOM (Document Object Model) representing the XML contained within the process variable. You can use the actual DOM API to parse and manipulate the XML content but for any complex structure this can be quite tricky.

To make this simpler, Oracle BPEL Process Manager provides a lightweight JAXB-like Java object model on top of XML; a so called XML façade. The façade provides a Java bean-like front end for an XML document/element. Façade classes have a corresponding Factory class which parse the XML document/element to create the façade, as show in the code snippet below:

    // Create Loan Application Bean
    LoanApplication loanApplication = LoanApplicationFactory.createFacade(loanApplicationElement);

Once the façade has been created, you can use its getter methods to access the required data.

Note: Façades are generated using the schemac tool shipped with Oracle BPEL Process Manager. You can use schemac to generate the façades from WSDL or XSD files (see the Oracle BPEL PM Developer guide for further details).

Summary

As we have seen Oracle BPEL PM provides a powerful Client API that makes it relatively simple to build a business specific “console” tailored to the needs of the user.

For further information on the API you should see the Oracle BPEL Process Manager Client API Reference

"Private" BPEL Processes

When developing any BPEL based solution, good practice dictates that you take a modular approach to process design, which allows the sharing of sub-processes among higher level processes. For example a payment process may be used by both the Expenses Process and Order Process.

As a result you will often end up with BPEL processes that you only intend to be called by other BPEL processes, and typically BPEL processes with at least some knowledge about the underlying process. So how do you prevent other ‘clients’ from directly invoking these processes?

Now initially this may sound like a security issue, and Oracle BPEL Process Manager provides a number of ways of securing BPEL Processes; in addition Oracle Web Services Manager provides a comprehensive solution for adding policy-driven security to all Web services (not just BPEL Processes).

However security is typically intended for enabling controlled secured access to a BPEL Process (or Web Service) by authorized clients. However in this case we don’t actually want any client directly accessing the process. Now admittedly we could take a standard based security approach to this, but is there a simpler way?

Now many programming languages such as Java provide the concept of private or protected methods that control what has access to them. For example, in Java a class may declare some of its methods as being protected; indicating that only other classes in this package can access these methods. The great thing about this approach is that the developer is actually signalling a level of intent, i.e. this method should not be called directly except by related classes (or sub-classes) that can be trusted to use the method correctly.

Ideally it would be great if BPEL provided similar functionality, however unfortunately it doesn’t. So is there a way of achieving something similar?

Well it turns out there is a fairly straight forward way of getting close to the desired effect. The approach makes use of the Oracle HTTP Server embedded within the Oracle Application Server (as such this won’t work for the Developer install) to prevent access to a specific URL pattern, plus the use of domains with Oracle BPEL PM to enable us to create a simple URL pattern for all “private” processes.

Configuring Oracle HTTP Server

Oracle HTTP Server is the Web server component of Oracle Application Server and is based on the Apache infrastructure.

Any one familiar with Apache administration is aware that it provides Allow and Deny directives which let you either allow or deny access to a particular URL (or pattern) based on the host name, IP address (or partial IP address) or a fully qualified domain name (or partial domain name).

By specifying a <LocationMatch> parameter in the httpd.conf file we can create a directive to only allow access to a URL which matches the specified pattern from the host on which the Oracle Application Server itself is installed.

For example the following directive will only allow access from the host to any URL which contains the string “/orabpel/private/”:

   <Location /orabpel/private/>
      Order deny, allow
      Deny from all
      Allow from localhost
   </Location>

Note: The httpd.conf file can be found in the directory:

   <ORACLE_HOME>/Apache/Apache/conf

Configuring Oracle BPEL PM

Now the trick is here to be able to provide a simple URL pattern on which to place the restriction. This is actually rather straight forward.

Oracle BPEL PM has the concepts of domains into which a BPEL process is deployed. A BPEL domain allows a single instance of Oracle BPEL Process Manager to be partitioned into multiple virtual BPEL sections (each identified by an ID and protected by a password). When Oracle BPEL Process Manager is installed, an initial domain named 'default' is created.

If you inspect the location of any deployed process, you will see that the domain makes up part of the URL, for example:

   http://[hostname]:[port]/orabpel/[domain]/[process name]/[process version]

So using this approach we can create a domain called “private”. By doing this we can define a directive to Apache to prohibit access to any URL that contains the pattern “/orabpel/private/”.

You can create a new domain from the BPEL console. From the initial login screen, instead of login into a particular domain, select the link Goto BPEL Admin (the default password is oracle). From here select the ‘BPEL Domains’ tab and then ‘Create New BPEL Domain’.

Then any process we want to make private we simply deploy to this domain. This obviously makes it very simple to make a process “private” with the added benefit is that it also indicates that the process itself is intended to be private.

Additional Considerations

Now this approach only works when you call the BPEL process via the SOAP stack. Oracle BPEL PM also provides a Java RMI interface; requests made via this interface don’t go via the Oracle HTTP Server. However access via RMI tends to be more tightly controlled so this should not present a serious problem.

It doesn’t prevent other clients on the same box from calling the BPEL Process, however again if the client is hosted on the same box then I would hope you have a reasonable handle on what its doing.

Finally it doesn’t prevent you from submitting the BPEL process from the BPEL Console as the invocation is made from the BPEL Server itself; however each domain is password protected so this shouldn’t really be an issue.

We also mentioned that this approach doesn’t work for the developer install of Oracle BPEL PM as it doesn’t embed the Oracle HTTP Server. However as a default this is probably a good thing as if the process was secured then it would prevent the BPEL Designer in JDeveloper from actually being able to read the WSDL which is required in order to develop the BPEL processes which calls the “private” processes.

Of course you can come up with variations of this approach, for example you could conduct developments against a mid tier install and configure the http directive to only allow certain developers (or at least there boxes) to be able to access processes deployed to a particular domain.

Conclusion

On its own this isn’t a perfect approach to securing a BPEL process, but then it’s not intended to be. But as a relatively simple approach to creating “private” BPEL processes where it is clear that this is the intention, then I believe this is certainly one way of achieving that.

However I would be interested in hearing any comments or suggestions on this approach or any alternatives.

Writing a Recursive BPEL Process

Recently I was working with a client who wanted to implement a recursive process (i.e. one that calls itself). Now Recursion is a classic programming pattern, and in theory it should be pretty straight forward for a process to call itself. However at first sight it’s not so obvious.

The issue here is that the way a BPEL Process calls out to another process is to drag a Partner Link on to your BPEL Process and then use the Service Browser to select the deployed process that you wish to call.

Of course with our scenario, the first issue we hit is that we haven’t yet deployed the process as we’re still writing in it! The obvious thing to do here is create a stub process, with just the basic receive and reply operations defined and then deploy the stub. This works great, we can now select the process within the ‘Service Browser’ and finish implementing the process.

However the next problem occurs when we try and re-deploy the completed process. Here the deployment fails with an error that it is unable to validate the WSDL for the Partner Link. The issue here is that as part of the deployment process we are “over-writing” the old stub version of the process, with a new version (as we want to keep the version number the same).

The way BPEL PM achieves this, is to un-deploy the old version of the process, before deploying the new version. As part of the deployment, the BPEL engine validates the WSDL for all Partner Links including our recently un-deployed version of the process and as a result fails!

Fortunately there is a simple work around. First create and deploy the sub process as before, then once deployed go to the BPEL Console and select the process from the dashboard. Then select the WSDL tab and click on the URL for the WSDL Location.

This will open a browser containing the WSDL for the BPEL Process, save the file to your local file system (File-> Save As in IE) as .wsdl file. Now when you create your Partner Link in the process, instead of using the ‘Service Browser’ use ‘Browse WSDL files from your Local File System’ and select the WSDL file you just saved.

Note: When prompted to make a local copy specify ‘yes’.

From here you will be able to implement and deploy your BPEL process without any problems.

Example

I’ve created a simple example process, based of course on the classic Factorial example. You can download this here.

Deployment Considerations

The one drawback with this approach is that the WSDL file will now contain the Endpoint location for the service within in it. Thus is you were to deploy the process on a different server it would fail at run-time.

So you need to modify the WSDL file at deployment time so that the endpoint reflects the hostname and port number of where the process is actually being deployed. The simplest way to do this is update your build script so that ant will automatically do this for you.

Final Thought

Now in theory you could have incredibly nested processes using this approach, however I would advise is bad practice and is likely to have performance implications.

For example; if we ran the Factorial process to work out 50 Factorial, that would result in 50 process instances. Now if we expected to handle 1000 process running in parallel, this would result in 50,000 process instances – so the actual number of process instance could increase very dramatically.

So I would recommend using this approach with caution to ensure that it doesn’t result in a dramatic increase in the overall number of process instances.

Using Analytics to Modify In-Flight Processes

I recently had the pleasure of presenting the Oracle Key Note at the Butler Business Process Management & Integration symposium. The subject of the keynote was to look at how we can use business analytics to enable us modify processes already in-flight.

When you consider the traditional closed loop BPM Lifecycle, as illustrated below, the emphasis has always been very much on using analytics about how processes have performed in the past, in order that we can modify the actual process definition in order to improve / optimize future versions of the process.



Whilst this approach has many benefits, the challenge was how we could modify processes already in-flight. In order to achieve this we need to overcome a number of challenges; firstly we need to collect analytics in (near) real time in order to base our decisions on up to date information; secondly we need a controlled way of modifying the in-flight process based on the data.

Collecting Real Time Business Analytics

Business Activity Monitoring (BAM) provides us with the tool to collect the near time analytics on which we can base our decision on how we wish to modify our in-flight processes.

For those of you who are not familiar with Oracle BAM, it enables the business to gain a real-time view of what’s happening with the business. To achieve this it provides the following key components:

• Capture Real Time Data - Within a BPEL process you can place a sensor on any activity with a process, when triggered this generates an event which is picked up by BAM in real time.
  
  Note: You can actually use events generated pretty much by any event based system (e.g. database triggers, messaging systems, etc), it’s just that BPEL makes it very easy.
  
• Analyze Processes, Trends, and Context - Next Oracle BAM is able to correlate all these events and synthesize them into meaningful data objects within it’s active data cache
  
• Interface for Business Users - The business can then build interactive real-time dashboards and proactive alerts on top of this active data to enable them to monitor the business processes.

For more information on BAM see http://otn.oracle.com/bam.

Modifying In-Flight Processes

Once we have the real time analytics, the next challenge here is to use the data to modify the processes already in flight.

Now when we talk about modifying In-Flight Processes; most people assume that this involves quickly writing a new version of the process, testing it, deploy it, and then migrating the existing in-flight processes to this new improved version. The reality is that is rarely as quick as required!!!

Rather what’s really required from a business perspective is to be able to modify the flow through a process (and its sub-processes) in order to obtain the desired business outcome. There are three basic patterns here which provide a way of achieving this, these are:

• Modify Process Flow – This uses business rules, which can be modified externally to the process to change the flow through a process instance.
• Exception Management – This uses BAM to launch a process to manage an exception which is (typically) occurring across multiple processes.
• Dynamic Process Assembly – Here we use BAM as source of real time data against which to evaluate Business Rules and use the results to dynamically call sub processes and assemble the end to end process on the fly.

Modify Process Flow

With this approach we are not looking to modify the actual process, rather modify the “path” through the process. If we look at any process, then at various points the process will hit a decision point which will determine which route it should take the through the process. Rather than building these decision points directly into the process we can externalise them in a rules engine such as Oracle Business Rules or iLog JRules, as illustrated below.



This then enables the business to modify the rules based on what’s currently happening within the business (as indicated through BAM) and thus get the process to take a different route.

An excellent case study for this is Cattles (a UK company which lends money to the secondary market), which uses a combination of BPEL, BAM and Rules for this purpose.

Exception Management

BPEL already provides a comprehensive way to handle exceptions, within the context of a process. However on some occasions we want to take a more holistic view to managing exceptions. This is certainly the case where we are suddenly “hitting” a common exception across multiple processes.

For example, if we have a loan flow process that’s going out to an external credit rating agency, and for some reason that credit rating check fails. Within the individual process BPEL process, we could have built in a simple re-try mechanism that simply waits a period of time before re-trying the service.

However if we start having a high number of failures (e.g. we may 1000’s of these processes running at any time), rather than handle the same exception multiple times (at least the same root cause) we can use BAM to detect that we have an issue and kick-off a single process to handle that exception (as illustrated below).

Dynamic Process Assembly

Rather than use BAM to feed a Dashboard, we can use it as a real time data source. A BPEL process can then query this real time data and pass this to the rules engine; enabling us to evaluate rules based on real time data.
At this point we could just use the result from the rules engine to dictate the flow through a process (as we did with the first example – Modify Process Flow).

However (as the title implies) we can take it a step further and dynamically assemble the process. The trick here is to have multiple “sub” processes all with the same WSDL definition. The rules engine rather than returning a “decision” now returns the end-point of the sub process to call, which the main BPEL process can then call dynamically (as illustrated below).



For more details on how to implement dynamic routing, see my previous blog; Using BPEL to Implement Dynamic Content Based Routing.

Final Thoughts

You can download a pdf of the full presentation from here. In the mean-time I would be curious to learn of any other strategies that are being adopted for modifying in-flight processes.