Overload leads to timeout errors

Sometimes integration scenarios can receive a bulk load from an application, especially when the source application uses batches as the trigger for sending out messages. This load can lead to the following non-retryable SOAP errors in the ESB:
Fault message:
Failed to enqueue deferred event "oracle.tip.esb.server.dispatch.QueueHandlerException:
Context lookup failed "[CONFIGURATION ERROR] Invalid Destination "Topics/ESB_JAVA_DEFERRED" :
javax.jms.InvalidDestinationException: Looking up java:comp/resource/esbRP/Topics/ESB_JAVA_DEFERRED:
javax.naming.NameNotFoundException: No resource named 'esbRP/Topics/ESB_JAVA_DEFERRED'found.
Please verify configuration of adminobject or the lookup string."
Make sure the topic is mapped to a jndi tree"

This behaviour is described in an Oracle note [ref: note ID 1173584.1] with the following solution: increase the timeout settings. Together with two other Oracle notes describing how to avoid BPEL errors due to adapters response time and what timeout settings in SOA can impact AIA [ref: note ID 1074227.1 and 885114.1] this leads to the following timeout settings:

• xa_timeout in $SOA_HOME/integration/esb/config/esb_config.ini
  Default: 60; recommended setting: 3600
• jms_receive_timeout in $SOA_HOME/integration/esb/config/esb_config.ini
  Default: 30; recommended setting: 300
• Also according to another Oracle note [ref: note ID 752385.1] you might want to set PingCount and PingInterval to 30 in $SOA_HOME/integration/esb/config/esb_config.ini
• syncMaxWaitTime in $SOA_HOME/bpel/domains/default/config/domain.xml
  Default: 45; recommended setting: 600
• transaction-timeout in $SOA_HOME/j2ee/oc4j_soa/application-deployments/orabpel/ejb_ob_engine/orion-ejb-jar.xml (change this value for all 6 occurences of transaction-timeout in this file)
  Default: up to 3000; recommended setting: 3600
• transaction-timeout in $SOA_HOME/j2ee/[container]/config/transaction-manager.xml (change this value for all containers: e.g. home, designtime and runtime [oc4j_soa])
  Default: from 300 to 3600; recommended setting: 7200
• Timeout in $SOA_HOME/Apache/Apache/conf/httpd.conf
  Default: 300; recommended setting: 300, this option specifies the amount of time Apache will wait for a GET, POST, PUT request and ACKs on transmissions. The default is 300 (seconds) however this may need to be increased.

The settings in the $ORACLE_HOME/integration/esb/config/esb_config.ini file apply to all ESB instances in the Oracle Home. One small bonus feature: if you have many BPEL processes deployed and you experience long waiting times in the ESB Console: put a lazyLoad property in the esb_config.ini file and the waiting time is gone [ref: Bug 7720420].
esb.console.services.lazyLoad.Allowed=true
It's a little confusing regarding the file orion-application.xml, different notes say different things, one note mentions the location in application-deployments which overrides esb_config.ini, another note mentions the location under applications which overrides esb_config.ini. Just to be safe and consistent check the orion-application.xml file in the following locations and if they are present and not commented out apply the values mentioned above for the xa_timeout, jms_receive_timeout, PingCount and PingInterval:
$SOA_HOME/j2ee/[ESB_RUNTIME_CONTAINER]/application-deployments/esb-rt/orion-application.xml
$SOA_HOME/j2ee/[ESB_RUNTIME_CONTAINER]/applications/esb-rt/META-INF/orion-application.xml
$SOA_HOME/j2ee/[ESB_DESIGNTIME_CONTAINER]/application-deployments/esb-dt/orion-application.xml
$SOA_HOME/j2ee/[ESB_DESIGNTIME_CONTAINER]/applications/esb-dt/META-INF/orion-application.xml

The above changes also take care of the following errors in the container logfile from $SOA_HOME/opmn/logs:
ORABPEL-05002
ORABPEL-02182
JTA transaction is not present the transaction is not in active state
Message handle error

Instead of increasing the timeout settings for the BPEL processes a more durable solution would be to throtte the inbound flow. There is an activation agent (bpel.xml) property (since 10.1.3.1), which can be used to control the speed at which the adapter posts messages to BPEL [ref: note ID 1178163.1 or Oracle® SOA Suite Best Practices Guide 10g Release 3 (10.1.3.3.0) E10971-01 December 2007]
...
    <activationAgents>
      <activationAgent partnerLink="JmsDequeuePL" ... >
          <property name="minimumDelayBetweenMessages">1000</property>
      </activationAgent>
    </activationAgents>
  </BPELProcess>
</BPELSuitcase>
This setting ensures that there at least will be 1000 milliseconds delay between two consecutive messages being posted to this BPEL process.

Actually this last best practices guide provides many valuable tips, like

• Performance tuning guidelines.
• Answers to frequently asked questions about threading.
• Answers to frequently asked questions about transactions.
• How to optimize the JVM heap. The heap size controls the amount of memory the JVM can use.
• How to relieve the dehydration store by making BPEL processes synchronous.
• Description of performance persistence parameters in bpel.xml.
• WSIF binding (localhost) and EJB binding optimization.
• The relationship among some performance settings
• The objectives and best practices for creating the BPEL cluster.
• Increasing the instanceKeyBlockSize to 100,000. Doing so decreases the frequency at which Oracle BPEL Server visits the dehydration store.

Some best practices should be considered during design and development time, like the BPEL parameters that are configured per BPEL component, whether a BPEL process should be synchronous or asynchronous and whether it should participate in a transaction or not. Other best practices are SOA Suite wide and should be part of the overall configuration, like the start-up memory settings.

Updated June 6, 2011 with LazyLoad, PingCount and PingInterval, another location of the orion-application.xml file and jms_receive_timeout to 300.

Instance wide XML manipulations

If the input file is a plain xml file without namespaces declared the following error occurs (at least when xml validation is set to 'strict'):
<invalidVariables xmlns="http://schemas.oracle.com/bpel/extension"><part name="code">
<code>9710</code></part><part name="summary">
<summary>Invalid xml document.
According to the xml schemas, the xml document is invalid. The reason is: Error::cvc-elt.1: Cannot find the declaration of element 'TestXSD'.
Please make sure that the xml document is valid against your schemas.</summary></part></invalidVariables>

Solution: create in the consumer routing service (the component after the file-reader, postfixed with _RS) the following transformation, first check if the wanted namespace is included in the namespace declarations, otherwise map one field to get all the namespaces declared. Then add the following code in the xsl:
<xsl:stylesheet version="1.0" xmlns:test="http://www.mydomain.com/TestXSD">
<xsl:template match="*">
  <xsl:element name="test:{local-name()}" namespace="http://www.mydomain.com/TestXSD">
    <xsl:apply-templates select="@*|node()"/>
  </xsl:element>
</xl:template>

This code is based on the "identity" template, which was introduced with an example in the XSLT Recommendation itself [ref: www.w3.org]:
<xsl:template match="@*|node()">
  <xsl:copy>
    <xsl:apply-templates select="@*|node()"/>
  </xsl:copy>
</xsl:template>

Some explanation here: "node()" is shorthand for child::node(). That is node() matches all nodes (element, text, comment, processing-instruction) being children of other nodes but not attributes since they are not considered children of their parent element. To find all nodes in order to copy them, we need to match both nodes being children of other nodes, node(), and attributes, @*, that is: match="@*|node()".

This template copies everything recursive from source to target, all the attributes and nodes. Copying everything in itself is no fun but becomes very useful when we add exceptions or conditions to the copying. Here two transformations that copy all the nodes and attributes if:

• They have content: RemoveEmptyNodes.xsl
  <xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:template match="@*|node()">
    <xsl:if test=". != ''">
      <xsl:copy>
        <xsl:apply-templates select="@*|node()"/>
      </xsl:copy>
    </xsl:if>
  </xsl:template>
</xsl:stylesheet>
• They have content or the node contains an attribute with content: RemoveEmptyNodesKeepAttributes.xsl
  <xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:template match="@*|node()">
    <xsl:if test=". != '' or ./@* != ''">
      <xsl:copy>
        <xsl:apply-templates select="@*|node()"/>
      </xsl:copy>
    </xsl:if>
  </xsl:template>
</xsl:stylesheet>

These templates can replace one of Oracle's best practices [ref: Oracle® Application Integration Architecture - Foundation Pack 2.5: Integration Developer's Guide Release 2.5 Part No. E16465-01 December 2009], adding the RemoveEmptyNodes transformation after the main mapping fulfills:

Transformation logic and XPath should be designed with the possibility of missing or empty elements
<xsl:if test="normalize-space(Line3/text())">
  <Line3>
    <xsl:value-of select="Line3/text()"/>
  </Line3>
</xsl:if>

The above templates can save you the time-consuming conditional mapping per element. Just don't forget if the mapping contains empty placeholders (like EBMHeader/EBMTracking/ExecutionUnitID) then fill these first with an assign before calling the RemoveEmptyNodes. Also the element EBMHeader/DataArea/[verb] is a required field. This can be solved with both templates:

• RemoveEmptyNodes.xsl: include in the template an exclusion of this element (the [verb] is Create here):
  <xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:corecom="http://xmlns.oracle.com/EnterpriseObjects/Core/Common/V2">
  <xsl:template match="@*|node()">
    <xsl:if test=". != ''">
      <xsl:copy>
        <xsl:apply-templates select="@*|node()"/>
      </xsl:copy>
    </xsl:if>
  </xsl:template>
  <xsl:template match="corecom:Create">
    <xsl:copy-of select="."/>
  </xsl:template>
</xsl:stylesheet>
• RemoveEmptyNodesKeepAttributes.xsl: fill an optional attribute of this field, although not the most elegant solution...

Cloud Forum 2011

The Cloud Forum 2011 Spring Edition in Media Plaza, Utrecht was not really what I hoped for. My expectations were to learn about the cloud and SOA, but what I learned at the forum was mostly about the technical infrastructure. Then again, when you're consuming services from the cloud you have already a SOA (Service Oriented Architecture) in place, so the focus should change from SOA integration to SOA governance and SOA security. Unfortunately my curiosity about SOA governance and SOA security was not satisfied, maybe next year?

One session from Mike Chung was interesting, he placed the cloud hype back in perspective. To identify a cloud Mike gave us the following mnemonic 0 - 1 - ∞:
0 - zero investments
1 - one version
∞ - unlimited scalability

Mike also wrote, together with John Hermans, the KPMG Cloud advisory [ref: From Hype to Future (pdf)]. It's a nice read, this cloud computing survey of 2010.

Cloud computing offers services on various IT layers. When it is at the software layer it is also known as Software-as-a-Service (SaaS). Gmail is an example of SaaS. Platform-as-a-Service (PaaS) delivers IT services on the platform layer such as an operating system or an application framework. Additional software must be developed or installed by the customer. Infrastructure-as-a-Service (IaaS) delivers technical infrastructure components such as storage, memory, CPU and network. Additional platforms and software must be installed by the customer.

The cloud computing market is dominated by very few vendors as the 'Big Four': Amazon, Google, Microsoft and Salesforce.com
SaaS (Software + Platform + Infrastructur): Salesforce.com, Microsoft BPOS, Gmail
PaaS (Platform + Infrastructur): GoogleApps, Force.com, 3tera AppLogic, Azure
IaaS (Infrastructur): Amazon EC2, AppNexus, Citrix Cloud

Basically cloud is the new hype-in-progress and fast-evolving. Yet there is a lack of vendors/standards/certifications, a good initiative can be found in the U.S. with FedRAMP [ref: FedRAMP]:

The Federal Risk and Authorization Management Program has been established to provide a standard approach to Assessing and Authorizing (A&A) cloud computing services and products. FedRAMP allows joint authorizations and continuous security monitoring services for Government and Commercial cloud computing systems intended for multi-agency use.

When will Europe follow?

Still today I learned a nice quote from a head of architecture of a firm in the industrial markets sector:

"We are still using our locally installed ESB and BPM software. Cloud computing is great for relatively simple services such as e-mail and CRM. When it comes to complex architectures such as SOA, a cloud is pretty much useless."

What do terms like AIA, EBO, EBS, etc. mean?

Oracle has it's own guide to explain all the AIA terms, the AIA Concepts and Technologies Guide [ref: Oracle® Application Integration Architecture -Foundation Pack 2.5: Concepts and Technologies Guide Release 2.5 Part No. E15762-01 October 2009]. Here a overview of the terms AIA, EBO, EBM, ABO, ABM, EBS, EBF, ABCS, BSR, CAVS, PIP, SOA, ESB and BPEL with their explanation:

AIA - Application Integration Architecture
Oracle Application Integration Architecture Foundation Pack provides Oracle's best-practice methodology and reference architecture, which are based on a profound service-oriented foundation running on Oracle's best-in-class middleware suites. Application Integration Architecture defines a common vocabulary across applications and industries. EBOs are the key elements in this context. They canonically describe standard business entities such as an order or an invoice. Based on these generic business entities, Application Integration Architecture delivers other artifacts such as EBSs, EBMs, EBFs, and ABCSs.
Basically AIA delivers you the integration framework including a canonical model.

EBO - Enterprise Business Object
You can think of EBOs as canonical, application-agnostic representations of frequently used business entities. The architecture starts with the concept of Enterprise Business Objects (EBOs). EBOs can be considered as application-independent representations of key business entities.
Basically EBOs are the objects moving between applications.

EBM - Enterprise Business Message
An EBM is the message format that is specific to the input or output of an EBS operation. Enterprise Business Service operations require specific message formats called Enterprise Business Messages (EBMs) for service requests and responses.
Basically EBMs are the envelopes containing the EBOs.

ABO/ABM - Application Business Objects/Messages
The Oracle AIA canonical model exists of the EBOs and EBMs. Outside the canonical domain are the application-specific formats, these can be objects and messages, respectively ABOs and ABMs. Every application has it's own set of ABOs and ABMs.
Basically ABOs and ABMs are application specific objects and envelopes.

EBS - Enterprise Business Service
An EBS provides the generic operations that an EBO should have. Enterprise Business Services (EBSs) are the centerpiece of the AIA Reference Architecture. They implement the required operations on EBOs in the right coarse-grained granularity that SOAs demand. For every EBO, Application Integration Architecture also ships generic service definitions to cover standard operations such as create, update, query, delete, and sync.
Basically EBSs are the routing services.

EBF - Enterprise Business Flow
An EBF orchestrates a number of EBSs to implement a complex integration flow. Sometimes, the integration scenario is not just a simple requester-provider relationship. In those cases, an Enterprise Business Flow (EBF) orchestrates any number of EBSs required to implement a specific business flow. The EBF completely controls the business logic and calls the appropriate EBS methods.
Basically EBFs are the orchestrations spanning multiple EBO's and/or applications.

ABCS - Application Business Connector Service
An ABCS implements a particular service operation in the context of a specific application. To enable applications to integrate into these generic, application-independent structures, you can implement Application Business Connector Services (ABCSs). Such an ABCS calls or is called by the appropriate EBS, depending on whether the application is in the requestor or provider role in a particular scenario. The main responsibilities of ABCSs include: conversion between the generic EBO and the application-specific format (ABO), cross-referencing of key attributes and message validation, and conversation with the specific application.
Basically ABCSs are the application-specific services per EBM.

BSR - Business Service Registry
The BSR stores and provides information about the objects, messages, and services that compose the integration scenarios in your Oracle AIA ecosystem. An integration scenario refers to the end-to-end (requester participating application-to-provider participating application) flow of a message. The BSR presents the components of your integration scenarios in a centrally managed and searchable repository, enabling it to become the system of record for your business services.
Basically BSR stores services, test-definitions and application-information.

CAVS - Composite Application Validation System
The CAVS enables you to test integration web services without the deployed participating applications in place. Using a framework that includes initiators to simulate calls to participating applications and simulators to simulate responses from participating applications, CAVS provides a system that can test integrations while also eliminating the need to set up deployments of all participating applications that are involved in the integration. The CAVS provides a test repository, an execution engine, and a user interface.
Basically CAVS is a testing framework.

PIP - Process Integration Pack
Prebuilt process integration packs (PIPs) from Oracle use the Oracle AIA to deliver composite industry processes for specific industries, using software assets from Oracle's portfolio of applications. Most of these solutions encompass orchestrated process flows, as well as prebuilt data integration scenarios that are meant to seamlessly connect the systems.
Basically PIPs are off-the-shelf integrations.

SOA - Service Oriented Architecture
Oracle AIA relies upon a SOA for integrations. SOA is an approach for defining an architecture based on the concept of a service. SOA applications are integrated at the service specification point, not at the application interface. This allows application complexities to be separated from the services interface, and diminishes the impacts of back-end interface and application changes. Service-oriented integration leverages messages to communicate between consumers and providers and uses XML schemas and transports such as SOAP to transport messages.
Basically SOA is integrating using (web)services.

ESB - Enterprise Service Bus
In integrations where it makes sense, Oracle AIA also relies on ESB. An ESB is the underlying infrastructure for delivering SOA. Oracle AIA uses Oracle Enterprise Service Bus as the foundation for services using SOA. It enables interaction through services that encapsulate business functions and supports service routing and the substitution and translation of transport protocols.
Basically ESB and BPEL are the engines driving Oracle SOA Suite. ESB is used for application adapters and routing rules (EBSs).

BPEL - Business Process Execution Language
Business Process Execution Language (BPEL), short for Web Services Business Process Execution Language (WS-BPEL) is an OASIS standard executable language for specifying actions within business processes with web services. Processes in Business Process Execution Language export and import information by using web service interfaces exclusively.
Basically BPEL and ESB are the engines of the Oracle SOA Suite. BPEL is used for transformations (ABCSs) and orchestrations (EBFs).

Sending an email from BPEL

Sending an email message from a BPEL Component is described in the Oracle BPEL guide [ref. Oracle® BPEL Process Manager Developer’s Guide 10g (10.1.3.1.0) B28981-03 January 2007]. But here are two different methods described: sending an email with a body and sending an email with two attachments. How to send an email with both a body and an attachment? The solution here is to use the second method (two attachments) and empty the BodyPartName of the first attachment, this attachment will then become the body of the email:
xmlns:notifsrvc="http://xmlns.oracle.com/ias/pcbpel/NotificationService"

<partnerLinks>
  <partnerLink name="NotificationService" partnerRole="NotificationServiceProvider" partnerLinkType="notifsrvc:NotificationServiceLink"/>
</partnerLinks>

... (population of varNotificationReq/EmailPayload)

<assign name="Remove_BodyPartName">
  <copy>
    <from expression="string('')"/>
    <to variable="varNotificationReq" part="EmailPayload" query="/EmailPayload/notifsrvc:Content/notifsrvc:ContentBody/notifsrvc:MultiPart/notifsrvc:BodyPart[1]/notifsrvc:BodyPartName"/>
  </copy>
</assign>

<invoke name="InvokeNotificationService" partnerLink="NotificationService" portType="notifsrvc:NotificationService" operation="sendEmailNotification" inputVariable="varNotificationReq" outputVariable="varNotificationResponse"/>

EBO customizations

Sometimes the EBO does not contain the elements for the information you have as part of this business object. For this case Oracle offers customization xsd's for each EBO and generic customization xsd's for common components to extend the EBO with needed elements and dataypes.

When you are going to extend an EBO, first try to find an existing common component to use for your attributes. Example when extending the EBO with street and housenumber, use the address datatype from common components. Sometimes the datatype of the used element is not known in the custom xsd, for this you have to add an import as well, for example if a common component is needed in CustomCommonComponents.xsd:
<xsd:import namespace="http://xmlns.oracle.com/EnterpriseObjects/Core/Common/V2" schemaLocation="../../../Common/V2/CommonComponents.xsd"/>

Also determine whether the information is new/updated or a reference to information already known by the systems. Example when sending a new or updated address information use AddressDatatype, when sending a reference to a known address then use AddressReference as referenceType.

If the addition is a just one element then choose the datatype from one of the Infrastructure datatypes (\EnterpriseObjects\Infrastructure\V1\DataTypes.xsd). Example when an element is needed to indicate whether something is fragile, add a fragileIndicator of datatype IndicatorType from this DataTypes list.

In order to maintain backwards compatibility always make customizations optional (minOccurs='0') to avoid validation errors in current messages. Check similar EBOs for consistency.

Besides the good practice to document the customizations in the custom xsd itself (as a comment in the header), it is very useful to have a spreadsheet with all the customizations documented. For example with the following columns: Filename of EBO, Type extended, Name of element added, User added and Date added. To keep track of removals the following two columns can be added as well: User removed and Date removed, just be careful when removing customizations as they can be used by other projects already.

AIA configuration properties

AIA brings us a very nice solution for maintaining variables in one global AIAConfigurationProperties.xml file (kind of ini-file) which are used at runtime. This is the perfect place for e.g. the systemID (application system id) which is used in the code to identify the application for DVM and XRef translations. Also is this systemID used to populate the EBMHeader Sender segment with information from the BSR (Business Service Repository) when the EBM is created in the BPEL RequesterABCSImpl component.

All the aia functions are described and explained in the AIA Development Guide [ref: Oracle® Application Integration Architecture - Foundation Pack 2.5: Integration Developer's Guide Release 2.5 Part No. E16465-01 December 2009].

Creation of the EBM Header in the transformation xsl file:
<xsl:variable name="serviceName" select="'{http://xmlns.oracle.com/ABCSImpl/[Application]/Core/[ABCSName]/V1}[ABCSName]'"/>
<xsl:variable name="systemID" select="aia:getServiceProperty($serviceName,'Default.SystemID',true())"/>
<xsl:variable name="senderNodeVariable" select="aia:getEBMHeaderSenderSystemNode($systemID,'')"/>
<xsl:variable name="messageIdVariable" select="orcl:generate-guid()"/>

<corecom:EBMHeader>
<corecom:EBMID>
 <xsl:value-of select="$messageIdVariable"/>
</corecom:EBMID>
<corecom:EBMName>
 <xsl:text disable-output-escaping="no">{http://xmlns.oracle.com/EnterpriseObjects/Core/EBO/Object/V1}SyncObjectListEBM</xsl:text>
</corecom:EBMName>
<corecom:EBOName>
 <xsl:text disable-output-escaping="no">{http://xmlns.oracle.com/EnterpriseObjects/Core/EBO/Object/V1}ObjectEBO</xsl:text>
</corecom:EBOName>
<corecom:CreationDateTime>
 <xsl:value-of select="xp20:current-dateTime()"/>
</corecom:CreationDateTime>
<corecom:VerbCode>
 <xsl:text disable-output-escaping="no">Sync</xsl:text>
</corecom:VerbCode>
<corecom:Sender>
 <corecom:ID>
  <xsl:value-of select="$senderNodeVariable/ID"/>
 </corecom:ID>
 <corecom:Description>
  <xsl:value-of select="$senderNodeVariable/Description"/>
 </corecom:Description>
 <corecom:IPAddress>
  <xsl:value-of select="$senderNodeVariable/IPAddress"/>
 </corecom:IPAddress>
 <corecom:Application>
  <corecom:ID>
   <xsl:value-of select="$senderNodeVariable/Application/ID"/>
  </corecom:ID>
  <corecom:Version>
   <xsl:value-of select="$senderNodeVariable/Application/Version"/>
  </corecom:Version>
 </corecom:Application>
 <corecom:ContactName>
  <xsl:value-of select="$senderNodeVariable/ContactName"/>
 </corecom:ContactName>
 <corecom:ContactEmail>
  <xsl:value-of select="$senderNodeVariable/ContactEmail"/>
 </corecom:ContactEmail>
 <corecom:ContactPhoneNumber>
  <xsl:value-of select="$senderNodeVariable/ContactPhone"/>
 </corecom:ContactPhoneNumber>
</corecom:Sender>

The configuration properties can also be used to retrieve the service-endpoints, pretty convenient when you'll have to deploy the code to other environments with different endpoints. When the CAVS framework is not used in your organization you can remove the whole CAVS endpoint javacode in the BPEL sourcecode which was generated by the ABCS Service Constructor. Set a webservice endpoint as follows:
<variable name="EndpointReference" element="wsa:EndpointReference"/>

<assign name="SetServiceEndpoint">
  <copy>
    <from>
      <wsa:EndpointReference xmlns:wsa="http://schemas.xmlsoap.org/ws/2003/03/addressing">
        <wsa:Address/>
      </wsa:EndpointReference>
    </from>
    <to variable="EndpointReference"/>
  </copy>
  <copy>
    <from expression="aia:getServiceProperty('{http://xmlns.oracle.com/ABCSImpl/[Application]/Core/[ABCSName]/V1}[ABCSName]','Routing.[PartnerLinkName].EndpointURI',true())"/>
    <to variable="EndpointReference" query="/wsa:EndpointReference/wsa:Address"/>
  </copy>
  <copy>
    <from variable="EndpointReference"/>
    <to partnerLink="[PartnerLinkName]"/>
  </copy>
</assign>

Besides populating the BSR with the application information to populate the EBM Sender elements, by logging into the /AIA site, go to Setup and then tab System, the key field here is System Code = systemID. The properties mentioned above have to be configured in the AIA file before lookup:
$AIA_HOME/config/AIAConfigurationProperties.xml, just above the </AIAConfiguration> tag:
<!-- [ABCSName] -->
<ServiceConfiguration serviceName="{http://xmlns.oracle.com/ABCSImpl/[Application]/Core/[ABCSName]/V1}[ABCSName]">
 <Property name="Default.SystemID">APPLICATION_01</Property>
 <Property name="Routing.[PartnerLinkName].EndpointURI">http://[HTTP_HOST]:[HTTP_PORT]/webserviceaddress</Property>
</ServiceConfiguration>
In the transformation source code (xsl) only the serviceName has to be specified, not the systemID (APPLICATION_01). This systemID gets retrieved from the AIAConfigurationProperties xml file. After editing this file don't forget to reload the AIA Configuration properties: log in the /AIA site, go to setup, go to configuration and after scrolling down you'll find the Reload button.

It's worth to mention that this AIA Configuration Properties file contains environment-specific properties, like partnerlink endpoints. Use the same procedure here as you have for managing connection pools and other physical endpoints/jdbc urls/id's per application per environment.

JNDI locations jms database

In the source code it's good practice to use Enterprise Information Systems (EIS) connection factories. In the Enterprise Manager this eis JNDI (Java Naming and Directory Interface) location points to the datasource JNDI location. Therefor the physical endpoint information can be kept outside the sourcecode, either in the connection factory (quick and dirty) or in a connection pool (multi-layered) to which the connection factory points to, via a datasource.

Here two examples using connection pools, one for a database connection and one JMS connection with the database as datastore.

Database connection, the source JNDI location is eis/db/XREF with the following connection factory $SOA_HOME/j2ee/oc4j_soa/application-deployments/default/DbAdapter/oc4j-ra.xml:
<connector-factory location="eis/db/XREF" connector-name="DbAdapter">
  <config-property name="xADataSourceName" value="jdbc/XREFDataSource"/>
  <config-property name="dataSourceName" value=""/>
  <config-property name="platformClassName" value="oracle.toplink.platform.database.Oracle9Platform"/>
  <config-property name="usesNativeSequencing" value="true"/>
  <config-property name="sequencePreallocationSize" value="50"/>
  <config-property name="defaultNChar" value="false"/>
  <config-property name="usesBatchWriting" value="true"/>
  <config-property name="usesSkipLocking" value="false"/>
  <connection-pooling use="none">
  </connection-pooling>
  <security-config use="none">
  </security-config>
  <connectionfactory-interface>javax.resource.cci.ConnectionFactory</connectionfactory-interface>
</connector-factory>
There is also a GUI available in the Enterprise Manager, goto instance [oc4j_soa], tab Applications and click on Default application. Then click module DbAdapter and tab Connection Factories.

The connection factory points to the JNDI location jdbc/XREFDataSource in $SOA_HOME/j2ee/oc4j_soa/config/data-sources.xml:
<managed-data-source connection-pool-name="XREFConnectionPool" jndi-name="jdbc/XREFDataSource" name="XREFDataSource"/>
<connection-pool name="XREFConnectionPool">
  <connection-factory factory-class="oracle.jdbc.xa.client.OracleXADataSource" user="aia" password="[password]" url="jdbc:oracle:thin:@[DB-HOST]:[DB-PORT]:[DB-SID]" commit-record-table-name=""/>
</connection-pool>
There is also a GUI available in the Enterprise Manager, goto instance [oc4j_soa], tab Administration and under Services go to task JDBC Resources.

That's it. Now the jms connection, the source JNDI location is eis/jms/JMSProducer with the following connection factory $SOA_HOME/j2ee/oc4j_soa/application-deployments/default/JmsAdapter/oc4j-ra.xml:
<connector-factory location="eis/jms/JMSProducer" connector-name="JmsAdapter">
  <config-property name="connectionFactoryLocation" value="java:comp/resource/JMSProducer/QueueConnectionFactories/QCF"/>
  <config-property name="factoryProperties" value=""/>
  <config-property name="acknowledgeMode" value="AUTO_ACKNOWLEDGE"/>
  <config-property name="isTopic" value="false"/>
  <config-property name="isTransacted" value="false"/>
  <config-property name="username" value="jmsuser"/>
  <config-property name="password" value="[password]"/>
  <connection-pooling use="none">
  </connection-pooling>
  <security-config use="none">
  </security-config>
  <connectionfactory-interface>oracle.tip.adapter.jms.IJmsConnectionFactory</connectionfactory-interface>
</connector-factory>
There is also a GUI available in the Enterprise Manager, goto instance [oc4j_soa], tab Applications and click on Default application. Then click module JmsAdapter and tab Connection Factories.

As you can see there is no JNDI location the connection factory refers to. This is done in the resource provider (you can see the name inside the config property connectionFactoryLocation, between the slashes after resource), the resource provider offers no Enterprise Manager GUI, only the xml file $SOA_HOME/j2ee/oc4j_soa/config/application.xml:
<resource-provider name="JMSProducer" class="oracle.jms.OjmsContext">
  <description>oc4j-jms AIASystem Queue Producer resource provider</description>
  <property name="datasource" value="jdbc/JMSDataSource" />
</resource-provider>

This resource provider points to the JNDI location jdbc/JMSDataSource in $SOA_HOME/j2ee/oc4j_soa/config/data-sources.xml:
<managed-data-source connection-pool-name="JMSConnectionPool" jndi-name="jdbc/JMSDataSource" name="JMSDataSource"/>
<connection-pool name="JMSConnectionPool">
  <connection-factory factory-class="oracle.jdbc.xa.client.OracleXADataSource" user="jmsuser" password="[password]" url="jdbc:oracle:thin:@[DB-HOST]:[DB-PORT]:[DB-SID]" commit-record-table-name=""/>
</connection-pool>
There is also a GUI available in the Enterprise Manager, goto instance [oc4j_soa], tab Administration and under Services go to task JDBC Resources.

BPELConsole session timeout

Want to disable the continuous and annoying relogins into the BPELConsole?
The solution is simple, just change the session timeout so you'll not logout automatically after some time anymore:

Login to the Enterprise Manager (em), open the oc4j container oc4j_soa and click on application orabpel, then click on module 'console' and go to tab Administration. Next click on the configuration properties - Go to Task to View/edit configuration properties for this web module. Here you can set the value of the Session Timeout (seconds): a value of -1 will disable the timeout, the default is 1200 (20 min.), 12 hours will be 43200.

As usual: the change will become effective after a restart.

This setting can be changed for all sites (em, esb, AIA), just find the right module by clicking the oc4j container oc4j_soa and click on the tab Administration. Next click on the J2EE Websites - Go to Task to Manage the J2EE websites in this OC4J instance. Click on the default-web-site and you see the list of all Web Modules, one per Application and their web-address (root-context). For the root-context BPELConsole you see application orabpel and module console, you can change the timeout for every application (em - ascontrol - ascontrol, esb - esb-dt - esbconsole, AIA - AIAApplication - AIAUserInterface) the same way as described above.

Don't forget to restart, once more:

1. Login to the Enterprise Manager (em).
2. Click the oc4j container [oc4j_soa].
3. Click on the tab Administration.
4. Click on the J2EE Websites - Go to Task to Manage the J2EE websites in this OC4J instance.
5. Click on the default-web-site.
6. You see the list of all Web Modules, click on th web module of the application you want to change the timeout of. Example click on the web module [console] of Application orabpel.
7. Click on the tab Administration.
8. Click on the configuration properties - Go to Task to View/edit configuration properties for this web module.
9. Set a new value of the Session Timeout (seconds): a value of -1 will disable the timeout, the default is 1200 (20 min.), 12 hours will be 43200.
10. Restart.

WSDL artifact errors while deploying

You'll already know that the Oracle JDeveloper contains a nice tool to create the base-mapping (XSL Transformation), unfortunately advanced logic within the mapping get screwed up by this JDeveloper pretty often. Therefor use an advanced text-editor with xml highlighting instead after you finish the base-mapping in JDeveloper. Or stick to the Source-view, no more Design-view.

This leads me to write about a more serious issue with JDeveloper: during development of the integrations all the developers working with ESB components need to have identical workspaces in their JDeveloper. If not, you might run into unwanted errors when deploying. This is known by Oracle and described in a note [ref: note ID 875558.1]:

Applies to: Oracle ESB - Version: 10.1.3.4 to 10.1.3.4.9. This problem can occur on any platform.
Symptoms: When an ESB project is registered using JDeveloper and then moved to a new JDEV Workspace with a different Workspace name, re-registration of the ESB project and its related artifacts (XSL Stylesheets, XML Schemas, WSDL files) might fail.
Cause: ESB artifacts designed with JDeveloper were moved to a different JDeveloper workspace with a new name.
Solution: DO NOT move ESB projects to a different JDEV workspace name. This is not supported by Oracle ESB in the current version.
Explanation: The way ESB identifies a file update in a project at runtime is by using the convention <workspacename>_<projectname>/<filename>. If the user is changing the workspace name, but keeps the project name and file name the same, it will not be possible for JDev to update the related artifacts in the slide repository because the identifier has changed - external references in form of esb:// were created. This is a known limitation.

You might also run into an error while deploying BPEL components that contain a referral to an ESB component, for example in the transformation the target is the ESB component WSDL. This url in the xsl contains the <workspace>, if in development this <workspace> differs from the actual deployed ESB component you might end up seeing the BPEL component in the BPELConsole, but not in the ESB console in BPELSystem. The advice here is: use always the same <workspace> in the JDeveloper, for example let the whole development team develop all ESB components (Projects) under Application AIASystem.

Talking about deployments, there is also a known bug in a jar-file, which is often used by automated deployments, see the note [ref: note ID 1308408.1] which describes the following error:
java.lang.Exception: Programming Error : Adapter Service <WriteData> does not have an associated WSDL artifact
at oracle.tip.esb.jdev.artifacts.model.ESBAdapterServiceArtifactContent.getDeploymentProperties(ESBAdapterServiceArtifactContent.java:159)
For example if you use osbs - Oracle Soa Build Server by Marc Kelderman [ref: http://orasoa.blogspot.com/2009/04/new-oracle-soa-build-server-osbs.html], you'll might receive the error as described in the note above. The solution here is to get from Oracle patch 11871410, extract the file ESBMetadataMigration.jar from the deployments.zip file in this patch and put this jar file in the location mentioned in the obbuild.sh script.

BPEL instance title in BPELConsole

For monitoring purposes it's a good practice to set the BPEL Instance title with an unique and descriptive meaning. For example: for order and invoice messages you might want to see the orderNumber in the instance title, like OrderNumber - [orderNumber].

To achieve this use the following code in the BPEL process:
<variable name="Title" type="xsd:string"/>
and
<assign name="GetTitle">
 <copy>
  <from expression="concat('OrderNumber - ',bpws:getVariableData('InputMessage','InputPart','/msg:rootElement/msg:orderNumber'))"/>
  <to variable="Title"/>
 </copy>
</assign>
<bpelx:exec name="SetTitle" language="java" version="1.5">
 <![CDATA[setTitle((String)getVariableData("Title"));]]>
</bpelx:exec>

Next it might be useful to search for all instances related to this ordernumber, for this a small fix is needed to automatically search with wildcards in the BPELConsole Instances:
Modify the following file: $SOA_HOME/j2ee/oc4j_soa/applications/orabpel/console/ngInstanceList.jsp
Replace
// construct where condition step 4: instanceTitle
//
String instanceTitleQ = request.getParameter( "instanceTitle" );
if ( instanceTitleQ != null && instanceTitleQ.length( ) != 0 )
{
  buf.setLength( 0 );
  tmpWhere.setClause( buf.append( n++ > 0 ? " AND " : "" )
    .append( SQLDefs.AL_ci_title )
    .append( " = ? " )
    .toString() );
  tmpWhere.setString( 1, instanceTitleQ );
  where.append( tmpWhere );
}

with

// construct where condition step 4: instanceTitle
//
String instanceTitleQ = request.getParameter( "instanceTitle" );
if ( instanceTitleQ != null && instanceTitleQ.length( ) != 0 )
{
  buf.setLength( 0 );
  tmpWhere.setClause( buf.append( n++ > 0 ? " AND " : "" )
    .append( SQLDefs.AL_ci_title )
//Start search with wildcards
//    .append( " = ? " )
//    .toString() );
//  tmpWhere.setString( 1, instanceTitleQ );
    .append( " LIKE ? " )
    .toString() );
  String wildcard = "%";
  instanceTitleQ = wildcard.concat(instanceTitleQ.concat(wildcard));
  tmpWhere.setString( 1, instanceTitleQ );
//End search with wildcards
  where.append( tmpWhere );
}

Now you can search by the ordernumber in the 'title'-field and the BPELConsole returns all instances with this number somewhere in the instance title.

How to determine the SOA Suite MLR version?

There are many ways to determine the MLR (SOA Bundle Patchsets - Merge Label Request) version of the SOA Suite, here are three ways to find your servers MLR version:
1. Go to $SOA_HOME/bpel/bin and run obversion.sh
./obversion.sh
*****************************************************************************
Oracle BPEL Server version 10.1.3.4.0
  Build: 0
  Build time: Sun Mar 07 17:51:49 PST 2010
  Build type: release
  Source tag: PCBPEL_10.1.3.4.0MLR10_GENERIC_RELEASE
2. Similarly, you can see this same information in the $SOA_HOME/bpel/domains/<domain_name>/logs/domain.log OR $SOA_HOME/opmn/logs/default_group~<container_name>~default_group~1.log file. Upon startup, the log file will contain the build information and MLR number entries similar to the output above.
3. Third way, look for the patches in opatch lsinventory: OPatch/opatch lsinventory |egrep '8372150|8589928|9317334'
Patch 8372150: TRACKING BUG FOR CUMULATIVE MLR#8 ON TOP OF 10.1.3.4.0
Patch 8589928: TRACKING BUG FOR CUMULATIVE MLR#9 ON TOP OF 10.1.3.4.0
Patch 9317334: TRACKING BUG FOR CUMULATIVE MLR#10 ON TOP OF 10.1.3.4.0

EBMTracking stamp both in BPEL as ESB

According to AIA [ref: Oracle® Application Integration Architecture - Foundation Pack 2.5: Integration Developer's Guide Release 2.5 Part No. E16465-01 December 2009] the messages get Tracking Information added within both the BPEL and ESB components. Below you find some updated code to get this EBMTracking segment populated (the Sample EBM here is of type SyncObjectEBM):

The transformation within the BPEL Component RequesterABCSImpl (xsl) which creates the EBM:

<xsl:variable name="ServiceName" select="'{http://xmlns.oracle.com/ABCSImpl/[SenderApplication]/Core/[ABCSName]/V1}[ABCSName]'"/>

<corecom:EBMTracking>
 <corecom:SequenceNumber>
  <xsl:value-of select="position()"/>
 </corecom:SequenceNumber>
 <corecom:ExecutionUnitID/>
 <corecom:ExecutionUnitName>
  <xsl:value-of select="$ServiceName"/>
 </corecom:ExecutionUnitName>
 <corecom:ImplementationCode>
  <xsl:text disable-output-escaping="no">BPEL</xsl:text>
 </corecom:ImplementationCode>
 <corecom:ActivityDateTime>
  <xsl:value-of select="xp20:current-dateTime()"/>
 </corecom:ActivityDateTime>
</corecom:EBMTracking>

After the transformation the ExecutionUnitID gets populated in the BPEL code with the following assign:

<assign name="Assign_InstanceID">
 <copy>
  <from expression="ora:getInstanceId()"/>
  <to variable="SyncObjectEBSReqMsg" part="SyncObjectEBM" query="/objectebs:SyncObjectEBM/corecom:EBMHeader/corecom:EBMTracking/corecom:ExecutionUnitID"/>
 </copy>
</assign>

The following transformation has to be created inside the ESB Component, the population of the ExecutionUnitID is described in the Best Practices Guide [ref: Oracle® SOA Suite Best Practices Guide 10g Release 3 (10.1.3.3.0) E10971-01 December 2007].
Tip: first let JDeveloper create the transformation and map one field to get all the namespaces populated, afterwards copy/paste the following code:

<xsl:template match="/">
 <ebo:SyncObjectListEBM>
  <corecom:EBMHeader>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:EBMID"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:EBMName"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:EBOName"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:CreationDateTime"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:VerbCode"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:MessageProcessingInstruction"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Sender"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Target"/>
   <xsl:copy-of select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:BusinessScope"/>
   <xsl:for-each select="ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:EBMTracking">
    <xsl:copy-of select="."/>
   </xsl:for-each>

<corecom:EBMTracking>
 <corecom:SequenceNumber>
  <xsl:value-of select="position() + 1"/>
 </corecom:SequenceNumber>
 <corecom:ExecutionUnitID>
  <xsl:value-of select="ehdr:getInstanceID()"/>
 </corecom:ExecutionUnitID>
 <corecom:ExecutionUnitName>
  <xsl:text disable-output-escaping="no">{http://xmlns.oracle.com/EnterpriseServices/Core/Object/V1}ObjectEBS</xsl:text>
 </corecom:ExecutionUnitName>
 <corecom:ImplementationCode>
  <xsl:text disable-output-escaping="no">ESB</xsl:text>
 </corecom:ImplementationCode>
 <corecom:ActivityDateTime>
  <xsl:value-of select="xp20:current-dateTime()"/>
 </corecom:ActivityDateTime>
</corecom:EBMTracking>

  </corecom:EBMHeader>
  <xsl:for-each select="ebo:SyncObjectListEBM/ebo:DataArea">
   <xsl:copy-of select="."/>
  </xsl:for-each>
 </ebo:SyncObjectListEBM>
</xsl:template>

Last thing: add the following property to your esb component:
name:    enableAccessBusinessEvent
value:   true

Routing rules

Basically routing rules are build upon the following three parts: sender, receiver and messagetype.

With AIA comes a library of EBOs (Enterprise Business Objects - canonical model), EBMs (Enterprise Business Messages - EBO + envelope per action/verb) and EBSs (Enterprise Business Service - routing service per EBO, one operation for each EBM).
As the EBS contains already one operation per EBM for an EBO we can see AIA takes already care of the messagetype. The routing rule in the EBS consists of filter, mapping and target. This leaves us with no place for the sender as part of the routing rule as defined by AIA.

Therefor we want to include the sender as follows, two scenarios are described here.

First scenario: direct integration from ApplicationX to ApplicationY, apply the filter on sender and empty target to exclude calls from the EBF (the only component allowed to stamp the target).
{/ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Sender/corecom:ID = 'ApplicationX_01' and not(/ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Target/corecom:ID/text())};

Second scenario: direct integration with an EBF (Enterprise Business Flow) involved for data enrichment in ApplicationZ, apply the filter on sender and the target, the target is stamped by the EBF (the only component allowed to stamp the target).
Data enrichment - send to ApplicationZ_01:
{/ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Sender/corecom:ID = 'ApplicationX_01' and /ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Target/corecom:ID = 'ApplicationZ_01'};
Receiver - send to ApplicationY_01:
{/ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Sender/corecom:ID = 'ApplicationX_01' and /ebo:SyncObjectListEBM/corecom:EBMHeader/corecom:Target/corecom:ID = 'ApplicationY_01'};

The advantage is a clear overview and easier to extend/maintain the rules, the advantage increases over time.

Queue character limit

The AIA comes with howto advices and best practices, following AIA results in uniform developed interfaces which are easy to monitor and maintain. For example the naming convention for queues, the following naming convention applies to the Queue name [ref: Oracle® Application Integration Architecture - Foundation Pack 2.5: Integration Developer's Guide Release 2.5 Part No. E16465-01 December 2009]:

AIA_<App><ABO>ProvJMSQueueV[Version Number]
This convention forgets the verb and list, so we extend this naming convention to (optional=<>, variable=[])
AIA_<[App]><[verb]><[ABO|EBO]><List><Prov>JMSQueue<V[version number]>
Explanation: App is not applicable when the queue is between an EBS and before EBF (within the AIA domain), verb is from the list create|update|delete|query|sync etc., an ABO might be used before an ABCSRequester, EBO after an EBS, use a List when an ListEBM is used, use Prov for queues between an EBS and before an ABCSProvider, version number is optional for first version as V1 is the default when no version is mentioned.

The above conventions applied to a queue between the Create ObjectListEBS and ApplicationX ABCS Provider results in the following name:
According tot AIA: AIA_ApplicationXObjectProvJMSQueue Length=34
Extended convention: AIA_ApplicationXCreateObjectListProvJMSQueue Length=44

As the Oracle Database for the JMS store has a limit of 24 characters on the queue name (as well on the queue tablespace name), the convention has to be adjusted in order to not to exceed the 24 character limit while still being descriptive:

If length>24 then Shorten Queue to Q
If length>24 then shorten verb to Cre, Upd, Sy, Del, Qry
If length>24 then shorten List to L
If length>24 then shorten <ABO/EBO> to max 6
If length>24 then shorten <App> to max 6
If length>24 then shorten <ABO/EBO> to max 4
If length>24 then shorten <App> to max 4

Try to remain descriptive here, so the example we shorten as follows:

AIA_ApplicationXCreateObjectListProvJMSQueue (44)
AIA_ApplicationXCreateObjectListProvJMSQ (40)
AIA_ApplicationXCreObjectListProvJMSQ (37)
AIA_ApplicationXCreObjectLProvJMSQ (34)
AIA_AppliXCreObjectLProvJMSQ (28)
AIA_AppliXCreObjtLProvJMSQ (26)
AIA_AppXCreObjtLProvJMSQ (24)

For the queue tablespace simply replace the Q(ueue) with T(ablespace).

Tips to start SOA Suite developments

In order to have all the interfaces in Oracle SOA Suite developed in an uniform way there is a need for coding standards and naming conventions. Oracle delivers the Application Integration Architecture, which includes a canonical model, a test-framework, coding standards, naming conventions and more.

Some tips when starting with Oracle SOA Suite:

• Stick to one version or patch-level for all environments. Only if a new version or MLR (patch-bundle) is needed, for example to fix a showstopper, start a project to upgrade/patch all your environments.
• Developers must have read the AIA Development Guide [ref: Oracle® Application Integration Architecture - Foundation Pack 2.5: Integration Developer's Guide Release 2.5 Part No. E16465-01 December 2009] or taken an AIA course. As for some the AIA Development Guide is not an easy and fast read, a good practice is to create a summary for those developers :) Add to this summary company-specific coding standards and naming conventions.
• Implement the components in either BPEL or ESB technology based on their purpose. All ABCSs in BPEL, all EBFs in BPEL, all EBSs in ESB, all non-soap communication with applications use ESB adapters.
• Use the AIA Service Generator for the BPEL components (ABCSs). This takes care of code for AIA ErrorHandling, CAVS test framework and sets instance title etc.
• Automate the deployments using scripts or e.g. the osbs - Oracle Soa Build Server by Marc Kelderman [ref: http://orasoa.blogspot.com/2009/04/new-oracle-soa-build-server-osbs.html]
• Set xml validation to strict on the development and test environments (except while executing stress-tests).
• Define ownership for all the EBO customizations, DVMs, XRefs, AIA ConfigurationProperties and system-wide parameters.
• When DVM or XRef technology is used for other purposes then intended, prefix the name with for example Z_.
• Define ownership of all the applications within the SOA Suite, especially naming (name per application or per module within, name the middleware communicating with or the application behind), ApplicationObjectLibrary and connection details.
• Define appropriate Service Groups in the ESB System AIASystem, for example one service group per application.
• Transformations from ABM --> EBM should be comprehensive mapping of all available attributes (All ABM Attributes supported in EBM should be mapped). Instance identifying information needs to be populated with as much information as possible in every Identification/IdentificationType element. EBM header needs to be populated fully.
• Use the EBM verbs/EBS operations with care and as intended. For example master data from the System of Record uses the sync, messages to the SoR use create/update.
• The same for EBM List messages, bundling messages lowers the number of instances, but if one message fails all the objects within this message are blocked.

Hope this helps you...