YOUR FEEDBACK
shirley wrote: As an ISV and service provider, we specialise in .NET based collaboration soluti...
Cloud Computing Conference
March 22-24, 2009, New York
Register Today and SAVE !..

2008 East
DIAMOND SPONSOR:
Data Direct
Frontiers in Data Access: The Coming Wave in Data Services
PLATINUM SPONSORS:
Red Hat
The Opening of Virtualization
Intel
Virtualization – Path to Predictive Enterprise
Green Hills
IT Security in a Hostile World
JBoss / freedom oss
Practical SOA Approach
GOLD SPONSORS:
Software AG
The Art & Science of SOA: How Governance Enables Adoption
PlateSpin
Effective Planning for Virtual Infrastructure Growth
Fujitsu
Automated Business Process Discovery & Virtualization Service
Ceedo
Workspace Virtualization
Click For 2007 West
Event Webcasts

2008 East
PLATINUM SPONSORS:
Appcelerator
Think Fast: Accelerate AJAX Development with Appcelerator
GOLD SPONSORS:
DreamFace Interactive
The Ultimate Framework for Creating Personalized Web 2.0 Mashups
ICEsoft
AJAX and Social Computing for the Enterprise
Kaazing
Enterprise Comet: Real–Time, Real–Time, or Real–Time Web 2.0?
Nexaweb
Now Playing: Desktop Apps in the Browser!
Sun
jMaki as an AJAX Mashup Framework
POWER PANELS:
The Business Value
of RIAs
What Lies Beyond AJAX?
KEYNOTES:
Douglas Crockford
Can We Fix the Web?
Anthony Franco
2008: The Year of the RIA
Click For 2007 Event Webcasts

SYS-CON.TV
TOP THREE LINKS YOU MUST CLICK ON


Building Better Bridges
Building Better Bridges

Enterprise Application Integration (EAI) is a very popular topic these days. Businesses, as well as government entities at the local, state, and federal level, are all struggling to integrate their critical yet disparate information systems.

Some of the most challenging integration efforts involve integrating enterprise resource planning (ERP) or customer relationship management (CRM) systems with new and existing custom applications. These enterprise software applications often have proprietary architectures and complex APIs, and typically use proprietary programming languages unfamiliar to the mainstream developer community. However, some enterprise software vendors are attempting to make the integration effort more approachable. Using widely accepted standards such as XML and Web services, these vendors are attempting to build better bridges between their respective products and the world around them.

PeopleSoft, Inc., is one of the dominant vendors in the enterprise software market. The PeopleSoft product family spans such critical business functions as human resource management, financial management, supply chain management, and customer relationship management. These products, along with competitive offerings from vendors such as Siebel, SAP, and Oracle, often manage the most important business logic and data in the enterprise. As such, they represent important assets to be leveraged whenever possible and appropriate.

This article is the first in a two-part series describing integration methods and illustrating the use of integration technology to enable mutual reusability between PeopleSoft and J2EE applications. This article will describe how the Java Messaging Service (JMS) and PeopleSoft's Integration Broker technology can be used to integrate PeopleSoft and J2EE applications seamlessly. The second article will describe how to accomplish the same goal using Web services technology.

PeopleSoft Architecture and Integration Technology
Starting with version 8, PeopleSoft products leveraged the PeopleSoft Internet Architecture (PIA), allowing the full functionality of these applications to be accessed from within a Web browser. By eliminating the need for client software, PIA's comprehensive Internet accessibility enables customers to realize substantially lower deployment costs than were previously possible under traditional client/server architectures. In addition to these benefits, PIA transitions the PeopleSoft product suite to a standard, J2EE-based operating platform built on BEA technology. PeopleSoft 8.4, the latest version (and the focus of this article), uses WebLogic Server 6.1 and Tuxedo for its Web and application servers, respectively.

As illustrated in Figure 1, PeopleSoft's Web tier components are deployed as J2EE Web applications under WLS 6.1, while its back-end components fall under Tuxedo control and are accessible via Tuxedo's Java-based Jolt API. These servers can be colocated or distributed, allowing additional flexibility in establishing environments that are optimized for scalability and redundancy. PeopleSoft Web applications handle a number of critical functions, from serving standard application pages and managing the distribution of batch-generated report output, to retrieving and aggregating user-requested content through the PeopleSoft Portal product. One Web application, the PeopleSoft Integration Gateway, plays a key role in integrating PeopleSoft products with each other and with third-party systems, including J2EE- and Microsoft .NET-based applications.

The Integration Gateway is a component of PeopleSoft's Integration Broker (IB) technology, introduced in PeopleSoft 8.4 (see Figure 2). The Integration Broker is an XML-based messaging hub that supports both synchronous and asynchronous messaging between PeopleSoft and external applications. IB is composed of several components and services spanning the Web and application server environments. The Integration Gateway resides on the Web server, providing inbound and outbound pathways for PeopleSoft applications. Other components, such as the internal publication and subscription agents that actually move information across the application server boundary, compose the Integration Engine, and reside on the application server under Tuxedo control. Ultimately, one should be familiar with both environments to successfully develop and administer an IB integration solution. Due to the constraints of this article, I'll focus on the Web tier and the Integration Gateway, where the bulk of the developer-relevant configuration tasks lie.

The gateway provides the interface to the broker, managing the actual receipt and transmission of messages between integrated systems. The gateway consists of several listening and target connectors that support inbound and outbound communications between external systems and the integration engine. The connectors support a range of communication protocols, including HTTP, JMS, POP3, SMTP, and FTP. Additional PeopleSoft and PeopleSoft 8.1 connectors enable gateways to communicate with other PeopleSoft 8.4 and older 8.1 systems, respectively. Some of these protocols are supported for both inbound and outbound communications, while others are supported in only one direction (see Figure 2).

The Gateway Manager processes every message that enters through the Integration Gateway, invoking gateway services such as XML parsing, message validation, and logging. The Gateway Manager uses the content of the message object to determine the intended message recipient, and then uses the appropriate target connector to transmit the message. After receiving a response from the target connector, it forwards the response to the calling listening connector. During all of this processing, the contents of incoming messages are logged to a disk file, along with any errors encountered during processing.

A number of different physical architectures are possible, depending on the nature of the desired integration effort. For example, a simple integration, behind a company's firewall, may require a single local Integration Gateway. More elaborate integrations across firewall boundaries, LANs, and WANs could very well involve multiple remote Integration Gateway instances and several different listening and target connectors. As such, real-world PeopleSoft integration can get very complex, very quickly. This article presents some of the basics but will not delve into all of the possible integration scenarios, optimizations, and consequences.

Integration Broker Definitions and Configurations
Every IB integration requires some degree of definition. The integration may involve several different systems communicating via multiple protocols and varying semantics. Hence, a number of conceptual definitions must be created and configured to appropriately model and implement the desired integration. The major IB definitions include gateways, nodes, transactions, relationships, messages, and message channels.

  • Gateway: Gateways handle incoming and outgoing messages through the broker. Each gateway has an address, which is the URL for that gateway's PeopleSoft listening connector, http://<gatewayservername>/PSIGW/PeopleSoftListeningConnector. Gateways use their PeopleSoft connectors to communicate with Integration Engines and other (remote) gateways. Additional connectors can be used to exchange messages with external systems using a wide range of protocols. Gateways can be remote or local. However, only one local gateway may be defined for any given PeopleSoft instance.

  • Node: Nodes represent connection points in IB integrations, for example, other PeopleSoft applications or external systems. Nodes can be remote or local. However, for IB, all nodes are considered remote, except for the Default Local node representing the PeopleSoft instance being configured. A node's definition can reference a target connector, indicating the protocol (ex: JMS, HTTP, SMTP) the broker must use in order to reach that node.

  • Transaction: Transactions are the most basic unit of work in an integration. They are associated with nodes, and designate which messages the node receives (inbound) and transmits outbound), and what semantics (i.e., synchronous, asynchronous) are used during the process. Every integration requires at least one transaction at each IB node either to send or to receive a message at the node.

  • Relationship: Relationships reconcile transactional differences between the source and destination nodes of a given integration. For example, a given integration may define an inbound, asynchronous (inAsync) transaction of Message X at Node A, and an outbound, synchronous (outSync) transaction of Message Y, at Node B. In addition, Message Y may be a transformed, translated, or filtered version of Message X. Since, in this example, the semantics change from inAsync to outSync and the message formats differ, a relationship must be defined in order to resolve the differences and support the integration.

  • Message: Messages represent the application data to be sent and received through the Integration Broker. Messages can be rowset-based or nonrowset-based. The former conveys hierarchical database information, based on native PeopleSoft structures called records. The latter can be used to convey practically any other message content or structure, including XML and Simple Object Access Protocol (SOAP) messages.

  • Message Channel: Message channels are logical groupings of messages. Each message must belong to one and only one message channel. Message channel definitions are essential for sequenced processing, improved throughput, and message archival.

    All of the IB definitions must be defined and configured using PeopleSoft's powerful PeopleTools Development Environment. Portions of PeopleTools are accessible via the PeopleSoft Internet screens (dubbed "PIA screens"), while other portions can only be accessed via the separate Application Designer client program. Gateway, node, transaction, and relationship definitions must be defined and configured via the PIA screens, while messages, message channels, and any custom PeopleSoft programs (written in a language called PeopleCode) must be created and maintained via the Application Designer.

    An Example: Integration via JMS
    The following example will reinforce the concepts presented above. In this example, an integration scenario between WebLogic Server 7.0 hosted components and a single local PeopleSoft 8.4 instance using JMS will be explored.

    One of the first steps in implementing an IB integration is to define the messages that will be exchanged between the participating systems. Two messages will be defined: one for messages emanating from WLS 7.0 to PeopleSoft, and another for messages traveling back. For the sake of simplicity, this example will use nonrowset messages with XML content.

    Since any given message must belong to an existing message channel, the channel must be created first. After the message channel has been created, its configuration properties can be browsed. Figure 3 is an example of a message channel configuration dialog displayed in the Application Designer.

    The JKK_TX_MESSAGE_CHANNEL channel dialog displays the list of messages associated with the channel. In this example, the JKK_TX_REQUEST_MESSAGE and JKK_TX_RESPONSE_MESSAGE definitions represent the application data messages exchanged between WLS 7.0 and PeopleSoft. Typically, structured rowset messages will be defined in the Application Designer, along with a record-based content schema. However, for this simplified example, the message definitions are simply monikers acting as proxies for the unstructured content they represent.

    In this example, synchronous messaging will be used for both inbound and outbound traffic. Once messages arrive in PeopleSoft, they will be written to a disk file before an empty response message is created and then returned to the WLS-hosted caller. After this, the original message will be routed back to WLS (over JMS) via a feedback loop.

    In order to accomplish this, you must first write a PeopleCode request handler and attach it to the JKK_TX_REQUEST_MESSAGE definition in the Application Designer. As in other development environments, such as Visual Basic, PeopleCode message objects can respond to various events. A message's OnRequest event can specify a custom PeopleCode routine to handle the synchronous arrival of the message. The PeopleCode routine can perform the appropriate actions before creating and returning a status message.

    The PeopleCode request handler for the inbound JKK_TX_REQUEST_MESSAGE is shown in Listing 1(listings for this article can be found at www.sys.con.com/weblogic/sourcec. cfm). As illustrated, the first programmatic instruction uses the special IB function, getMessageXmlDoc(), to retrieve the incoming XML message. After this, the message contents are converted into an XML string before being appended to a file named IncomingRequests.log. Finally, CreateMessage and ReturnToServer are special functions used to generate and then return an empty status message through IB back to the caller.

    The next major step is to define and configure the necessary integration nodes, transactions, and relationships using the PeopleTools PIA screens.

    Two nodes, JKK_SOURCENODE and JKK_TARGETNODE, will be created to represent the WLS instance participating in the integration. This instance will be the ultimate receiver of the messages it originally transmitted. The JKK_SOURCENODE definition is fairly straightforward. However, the JKK_TARGETNODE node definition is more involved because it must ultimately transmit messages to an external destination (the WLS instance) via JMS. Figure 4 provides a look at the target connector information attached to JKK_TARGETNODE. Although there are several properties associated with the connector, we will focus on only the most important ones: JMSUserName, JMSPassword, JMSUrl, JMSTopic, and JMSFactory.

    JMSUrl specifies the host and listening port of the target JMS server. In this case, the server is WLS 7.0-based and resides at http://wls01:7001. JMSUserName and JMSPassword are valid user ID/password combinations on the target WLS 7.0 instance. JMSFactory and JMSTopic reference the factory and topic defined within the target JMS server's config.xml file. A JMS listener or message-driven bean (MDB) listening to the same combination of JMSFactory, JMSTopic, and JMSServer will rendezvous with the IB and receive any messages transmitted by the broker via JKK_TARGETNODE.

    The next step is to define the necessary inbound and outbound transactions and associated relationships. IB transaction mappings must be made from the perspective of the PeopleSoft instance's default local node. Hence, an inSync transaction conveying JKK_TX_REQUEST_MESSAGE and returning JKK_TX_RESPONSE_MESSAGE must be defined at JKK_SOURCENODE, while an outSync transaction with identical message characteristics must be defined at JKK_TARGETNODE.

    In addition, a relationship, shown in Figure 5, will be created to tether the two nodes together so inbound message traffic is routed to the target node and ultimately back to the WLS instance via the JMS target connector. Relationships, like many PeopleSoft objects, can be activated or deactivated, and can even become applicable upon a given effective date. All of this provides considerable flexibility, as multiple relationships can be defined across a set of node and date combinations.

    In addition, relationships can be used to reconcile differences in the data formats exchanged between a combination of source and target nodes. For example, the data emanating from JKK_SOURCENODE may need to be transformed, translated, or even filtered en route to JKK_TARGETODE. A relationship can perform these data manipulations using PeopleSoft batch programs called application engine programs. These programs can use a combination of PeopleCode, Extensible Stylesheet Language Transforms (XSLT), or SQL actions to perform the required data manipulations between a given combination of nodes. Such programs would be specified in the Transformations section of Figure 5.

    The next step is to configure the Integration Gateway to use the appropriate inbound JMS topics and queues. This configuration information is stored in the integrationGateway.properties text file, which is located under the Integration Gateway subdirectory of the installation tree. The gateway can listen to any combination of JMS topics and queues, and can use different JMS providers. For this example, the WebLogic JMS provider will be used, along with a single inbound JMS topic and a special gateway error topic. The gateway will publish processing errors to this topic, in addition to writing the errors to a cumulative error log file. This design would accommodate a custom error topic subscriber that could handle gateway errors programmatically.

    Listing 2 shows the portions of an integrationGateway.properties file relevant to the integration example. The ig.isc.serverURL, ig.isc.userid, ig.isc.password, and ig.isc.toolsRel properties specify the Jolt address, along with a valid user ID/password combination for the target application server. These properties effectively connect the gateway, which resides on the Web server, to the appropriate application server. Messages that enter the gateway will, under normal conditions, be routed to the target application server for processing by PeopleCode message subscriptions and OnRequest handler programs.

    The ig.jms.JMSProvider.JNDIFactory.WebLogic property setting instructs the gateway to use WebLogic as the JMS provider. Next, the ig.jms.Topics, and ig.jms.Topic1 lines define the total number of JMS listening topics and specific topic connection information, respectively. The subsequent ig.jms.ErrorTopic lines specify the JMS error topic to which gateway errors should be published. The JMS connection properties shown in Listing 2 will cause the gateway to rendezvous with topic publishers and subscribers on the WLS instance server.

    The final step of this example is to create the actual JMS topic publishers and subscribers that will rendezvous with the gateway via the topics defined in the integrationGateway.properties file. Specifically, a JMS subscriber will be needed to cover both the error topic and the PostingStatusMessage topic specified in the JMS target connector. These JMS subscribers can be standard JMS message listeners or even MDBs. A standard JMS topic publisher can be used to send JMS TextMessage instances into the gateway.

    Listing 3 shows three message properties that must be set prior to publishing JMS TextMessage instances to the gateway. The first property, MessageName, specifies the message associated with the textual message being published. The DestinationNodes property specifies the default local node for the target PeopleSoft system. Finally, the RequestingNode indicates the IB node sending the published messages. Since the MessageName and RequestingNode shown in the listing match the inSync transaction previously defined at JKK_SOURCENODE, the broker will accept and route the published messages to the JKK_TARGETNODE node.

    Conclusion
    The concepts, definitions, and procedures outlined in this article enable J2EE and PeopleSoft 8.4 applications to be seamlessly integrated using JMS. The benefits of taking this approach are clear: standards-based protocols and data (e.g., JMS and XML), transactional integrity, assured message delivery (both within PeopleSoft IB and WebLogic JMS), and high scalability using a combination of mechanisms such as message channels and pools of MBDs. Another benefit is that it can be used to integrate J2EE with any PeopleSoft 8.4-based product, including Financials, HRMS, SCM, and CRM.

    It's important to remember that this article provides an outline, not an exhaustive treatment of this topic. Important issues such as security have not been discussed. Others, such as transaction management, are only implied by virtue of using JMS, but have not been covered. Both J2EE and PeopleSoft's Integration Broker are vast, complex technologies that cannot be covered in any appreciable detail within the constraints of this article. However, this outline can become the first step in the journey toward mastery of these technologies. The next article in this series will explore the integration of J2EE and PeopleSoft 8.4 applications using Web services.

    About Joseph K. Krozak
    An expert systems architect with over fifteen years of full lifecycle software engineering experience in various applications including banking, supply chain management / manufacturing, law, insurance, and telecommunications. I have spent years mentoring management and development teams on the architecture, design, implementation, and planning of large-scale, distributed, enterprise applications.

  • BEA WEBLOGIC LATEST STORIES
    Okay, here's the deal. When you observe the big software guys and see how quickly they adopt emerging technologies, which will change IT the way we know it today, here is what we see. Larry Ellison invested millions in old SaaS / cloud companies, which gave him zippo in return, and he ...
    SYS-CON Events announced today that more than 40 Cloud technology providers, as well as Virtualization and SOA companies will exhibit at the upcoming 1st International Cloud Computing Conference & Expo (www.CloudComputingExpo.com), November 19-21, in San Jose, California. The conferenc...
    SYS-CON Events announced today that the leading global SOA, Virtualization, Cloud Computing and Open Source technology provider FreedomOSS named "Gold Sponsor" of SYS-CON's SOA World Conference & Expo which will take place November 19-21, 2008, at the Fairmont Hotel in the heart of Sil...
    Cassatt, the company started by BEA founder Bill Coleman, is redirecting its data center widgetry into creating internal clouds comparable to Amazon or Google out of infrastructure customers already have in-house. Coleman observed that most IT professionals aren’t comfortable outsour...
    Just as people begin to understand the difference between web ops and IT, we are entering a period where clouds promise "Ops-Free" computing. Because it’s easy, scalable, available and disposable, the cloud is well on its way to becoming “technology’s next big thing.” However, ...
    As far as the software industry goes, these tough economic days give the biggest business advantage to those companies who contribute directly to the solution of the big global problem and they will be the first to flourish as we dig ourselves from the ditch. Call that the new Y2K prob...
    SUBSCRIBE TO THE WORLD'S MOST POWERFUL NEWSLETTERS
    SUBSCRIBE TO OUR RSS FEEDS & GET YOUR SYS-CON NEWS LIVE!
    Click to Add our RSS Feeds to the Service of Your Choice:
    Google Reader or Homepage Add to My Yahoo! Subscribe with Bloglines Subscribe in NewsGator Online
    myFeedster Add to My AOL Subscribe in Rojo Add 'Hugg' to Newsburst from CNET News.com Kinja Digest View Additional SYS-CON Feeds
    Publish Your Article! Please send it to editorial(at)sys-con.com!

    Advertise on this site! Contact advertising(at)sys-con.com! 201 802-3021

    SYS-CON FEATURED WHITEPAPERS

    ADS BY GOOGLE
    BREAKING NEWS FROM THE WIRES

    In the graph before the boilerplate, the first sentence should read: The Evans Data...