Jul 05

In some views SOA is represented as a series of 4 layers: Presentation Layer (SOA 1), Business Process Layer (SOA 2), Business Service Layer (SOA 3) and Technical Layer (SOA 4). Typically each layer higher up in the hierarchy consumes services exposed by the layer under it. So the Presentation Layer would consume services provided by the Business Process or Business Service Layers. Service interfaces are described using Web Services Description Language (WSDL), sheltering service consumers from details of service implementation. Web Services are seen as the technical means to implement the decoupled functional layers in a SOA development. Decoupling allows implementations of business functionality at different layers to be swapped in and out without disturbing other layers in the stack.

In this document I will implement a multi-operation Web Service that will allow patient information to be upserted into a database table and will return all patient details for a patient whose Facility+Local ID are specified in the request. This service will be used to populate the patient table and to implement patient lookup portlets, discussed in other writeups in this series. This is a basic Patient Service that hides the specifics of interaction with the patient data store form applications that need to interact with it, by providing a defined interface and web service-based implementation. Thus the data store may change but the service consumers need not. We use the Database BC with select, insert and update operations and Database BC with SQL File-based parameterized SQL prepared statement. We handle null value insertion on missing data. We also use the SOAP/HTTP BC and the BPEL SE.
The business idea is that patients are looked after in various healthcare facilities. Information about patients is stored in a relational database. This information must be inserted, for new patients, and updated, for existing patients, as required. Frequently applications need to search for a patient and display details to human operators. To shelter application developers from the details of the data store the upsert functionality and patient details lookup functionality will be made available as a multi-operation web service.

Walkthrough Document: 02_PatientSvc_GFESBv21.pdf
Companion Archive: 02_PatientSvc_companion_archive.zip

Jul 02

“Progress” notwithstanding, Healthcare environments still extensively use the HL7 v2.x Delimited messages for conveyance of patient and patient-related information between applications. The GlassFish ESB provides support for HL7 v2.x messaging in the form of the HL7 Encoder, which allows conversion between HL7 v2 Delimited and HL7 v2 XML message formats, and in the form of the HL7 Binding Component, which allow connectivity between the GlassFish ESB-based healthcare solutions and healthcare applications that support HL7 over TCP connectivity.

In this document I will walk through the process of generating HL7 v2.3.1 delimited messages from pipe-delimited records containing patient information, sending and receiving HL7 v2.3.1 delimited messages using the HL7 Binding Component, parsing HL7 v2.3.1 delimited messages and writing HL7 v2 delimited messages to a file. To create and process HL7 messages I show how create a custom ADT A04 XML Schema and a custom “any HL7 v2 message” XML Schema. This gives me an opportunity to use the File Binding Component (File BC), the HL7 BC, the HL7 Encoder, the Custom Encoder and the BPEL Service Engine (BPEL SE). This also gives me an opportunity to demonstrate a HL7 v2.3.1 delimited message sender solution and to demonstrate a HL7 v2.3.1 delimited message receiver solution. At the end of the process we will have a file containing HL7 v2 delimited ADT A04 messages, which we will use in related writeups.

Here is the document: 02_PatientSvc_MakeHL7v2DelimDataFromCustomDelimRecords_0.4.pdf

Here is the companion archive containing input files, the output file and the projects: 02_PatientSvc_MakeHL7v2DelimDataFromCustomDelimRecords_data.zip

The writeup document has been updated and version changed to 0.4.

Jun 30

This Quick Note discusses a simple solution to the use case provided by Leonard Barkley. The question goes like this:

“I dont have any idea how to implement BPEL process but the BPEL deployed as a subscriber of a topic. usually I implement the BPEL process and deployed it as web service.”

We produce a simple GlassFish ESB v2.1-based solution which reads a file, sends its content to a JMS Topic and another simple GlassFish ESB v2.1-based solution which subscribes to the same JMS Topic, receives the message and writes it to a file. Both solutions will use BPEL to implement the simple logic, though it is possible to implement both solutions without BPEL, so we have File BC -> BPEL SE -> JMS BC -> JMS Provider (Topic) -> JMS BC -> BPEL SE -> File BC.

Here is the note: QuickNote003_forLeonardBarkley.pdf

Hire is an archive with the project group containing all the projects developed in the Note: JMSTopicSampleProjGrp.zip

As Leonard Barkley pointed out to me, having implemented the sample, the Note is incorrect on Pages 23 and 24. The JMSSubscriber_JMSIn WSDL should use the Receive type, not Send type as the docuent states. The solution still works, it appears, but the configuration as documented is confusing. Thanks Leonard.

Jun 22

In some views SOA is represented as a series of 4 layers: Presentation Layer (SOA 1), Business Process Layer (SOA 2), Business Service Layer (SOA 3) and Technical Layer (SOA 4). Typically each layer higher up in the hierarchy consumes services exposed by the layer under it. So the Presentation Layer would consume services provided by the Business Process or Business Service Layers. Service interfaces are described using Web Services Description Language (WSDL), sheltering service consumers from details of service implementation. Web Services are seen as the technical means to implement the decoupled functional layers in a SOA development. Decoupling allows implementations of business functionality at different layers to be swapped in and out without disturbing other layers in the stack.
In this document I will walk through the process of developing a SOA Composite Application, exposed as a Web Service, which will make available simple business functionality using a multi-operation service. We will use the GalssFish ESB v2.1 infrastructure. The service will use the SOAP/HTTP Binding Component, the Database Binding Component and the BPEL 2.0 Service Engine. This simple service will introduce the components and discuss how a multi-operation web service can be constructed using the GalssFish ESB.
The business idea is that patients are looked after in various healthcare facilities. Frequently applications need to allow selection of a facility and to access facility details for display to human operators. A relational database is used to hold the details of facilities which are a part of the healthcare enterprise. To shelter application developers from the details of the data store facility list and details will be made available as a multi-operation web service. This web service will be used elsewhere to construct a portlet that can be used in an enterprise portal.

The document is available as 01_FacilityService_GFESBv21.pdf

Late breaking news: The MySQL JDBC Driver, which I used in the example, mysql-connector-java-5.1.7-bin.jar, and which is distributed with the GlassFish ESB v2.1, causes connection pool exhaustion and other issues with the example. If you experience these issues pleases download the latest MySQL JDBC Driver, mysql-connector-java-5.1.11-bin.jar as at now, from http://dev.mysql.com/downloads/connector/j/5.1.html and replace the original in domains/domain1/lib/ext.

Jan 01

This Note walks the reader through development of a Java CAPS 6/JBI-based / OpenESB-based solution that addresses a Healthcare-related business problem. The Note elaborates on the healthcare background necessary to get a notion of what is being done and why, and provides detailed steps required to implement and exercise the solution to the business problem.

Updated note, where GlassFsish ESB v2.1 is used instead of Java CAPS 6, is available at 00_HL7_Example_Development_Instructions_Final_1.1.2.pdf

We will use the HL7 Binding Component, the File Binding Component, the JMS Binding Component, the SOAP/HTTP Binding Component, the BPEL 2.0 Service Engine, the JavaEE Service Engine, the HL7 Encoder and EJB-based Web Services in a JBI-based solution.

In the process we will create XML Schema Documents (XSDs), Web Services Description Language Documents (WSDLs), a BPEL 2.0 Business Process, an EJB-based “Implementation First” web service, an EJB- and WSDL-based “Interface First” web service, a bunch of Composite Applications, BPLE 2.0 mapping, BPEL 2.0-based Web Service orchestration, on-the-fly conversion of HL7 version 2.3.1 delimited messages to their XML equivalents. We will get a pretty good exposure to what OpenESB and Java CAPS 6/JBI components look like, how they work and how they can be used to create real business solutions. Above all, we will develop and test a solution that is more sophisticated then the customary “Hello World” examples but not so complex as to take too long to build and become too hard to comprehend by a novice user.

The particular business problem and the particular solution came about because once upon a time there was intent to build a series of related OpenESB projects – HL7 Processor, MDM Processor and IEP Processor – that would:

  • receive HL7 v2.x delimited messages
  • convert HL7 v2.x messages to their equivalent XML format
  • split message stream into ADT A01s, ADT A03s and other
  • convert A01s to an abbreviated Custom Patient XML format
  • convert A03s to an abbreviated Custom Discharge format
  • send Custom Patients to a JMS Queue for processing by a MDM solution
  • send Custom Discharges to a JMS Queue for processing by an IEM solution
  • have the MDM process Custom Patients into a Master Patient Index
  • have the IEP process Custom Discharges to flag excessive length of stay

The MDM Processor and the IEP Processor made it to the Sun CEC 2008 as demonstrations, with associated Tutorials by Tom Barrett, and demonstration recordings by me. The HL7 Processor did not make it. With the appearance of Java CAPS 6 Update 1 more JBI components made it into the officially supported Sun product. While the HL7 BC and the HL7 Encoder did not make it into this Update they will, eventually. Both components are already available from the OpenESB site and can be installed into the Java CAPS 6 Update 1 installation as unsupported components. This is what we will do for this Note.

The complete Note is to be found here.

The archive containing supplementary material you can use to save yourself the trouble of downloading component and, above all else, containing test files with HL7 delimited records, is here.

For these interested in processing HL7 using GlassFish ESB v2.1 there is a blog entry, “GlassFish ESB v2.1, MySQL v5.1 – Make HL7 v2.3.1 Delimited Messages from Custom Delimited Records with HL7 Encoder“, which discusses, amongst other things, how to create a custom HL7 v2 ADT A04 XSD and a “match any” XSD.

Tagged with:
Dec 19

Following the CEC 2008 Conference in Las Vegas, where the Java CAPS Stream saw a bunch of presentations and demonstrations, I am happy to offer screencasts of the three demonstration sessions I recorded for the event.

The GlassFish ESB screencast is the ScreenCast of the CEC 2008 GlassFish ESB Essentials Lab demonstration. This is a recording of the demonstration described in detail by Tom Barrett in
the GlassFish ESB Tutorial and Lab document. The screencast is an extended version of what the CEC audience got to see. In this screencast I use the OpenESB distribution to discuss, design and implement an abbreviated Supply Chain solution
. I use the File BC, the SOAP/HTTP BC, BPEL 2.0 SE, the Custom Ecoder, the XSD Editor, the WSDL Editor, BPEL correlations, BPEL Pick wit Timer, CASA Editor and a bunch of other OpenESB/GlassFish ESB featiures and facilities. Watching the screencast will give you a pretty good idea what the tooling looks like, how easy it is ti use it, how a theoretical requirement can be turned into a practical design and how that design can be implemented and exercised using the tooling and infrastructure you can get free of charge and use as much as you might desire.

Data for the following two screencasts/demonstrations is produced by the solution discussed in the next blog entry, which ought to precede these two.

The Java CAPS 6/Mural Master data Management screencast is the ScreenCast of the CEC 2008 Java CAPS Essentials Master Data Management (MDM) Lab demonstration. This is a recording of the demonstration described in detail by Tom Barrett in the Java CAPS Essentials MDM Tutorial and Lab document. In the screencast I discuss what the Master Data Management (MDM) is, how a Healthcare enterprise might leverage it to improve its business and how the OpenESB or Java CAPS 6 can be used to implement MDM. I use OpenESB to design a Master Patient Index Data Model, implement it with the tool, generate Data Model-based Master Index Data Management Web Application, build an integration solution to feed the MDM solution with transactional data form Hospital Information Systems and build a broadcast processor solution that can be used to send master patient index updates to downstream systems which have a need to be kept in synch with the enterprise view of the patient. One will get a very good idea of what the core Master Data Management is about, how easy it is to create the MDM Application and related integration components using the OpenESB/Java CAPS 6 tooling, and how the business of maintaining master patient index looks and works like.

The Java CAPS 6 / Intelligent Event Processor screencast is the ScreenCast of the CEC 2008 Java
CAPS Essentials IEP Lab demonstration. This is a recording of the demonstration described in detail by Tom Barrett in the Java CAPS Essentials IEP Tutorial and Lab document.The screencast is what the CEC audience got to see. In this screencast I demonstrate how an Intelligent Event Processing (IEP) solution is built and exercised. The solution addresses a Helathcare business problem – it calculates an Average Length of Stay for each patient in a sliding time window, based on data from an ADT A03 HL7 Discharge message, works out which patients’ Length of Stay exceeds average for the patients in the window by 1.5 times, and passes records related to these patients
on while discarding ‘normal’ records.

The AVIs were recorded with Camtasia Studio. You may need a Camtasia Player to playe them on Windows. You could also try getting a Camtasia codec for your platform/player from the Camtasia site.

I had audio quality problems when directly playing the recordings through Mozilla, which used the Quicktime plugin. The best thing to do is to download the recordings and try different players until one works for you.;

Enjoy.

Tagged with:
preload preload preload