The evolution of software development has taken a paradigm shift from the monolithic architectures of the past that used conventional procedure-based programming. After procedural programming came the era of client-server computing, where there was a notable distinction between data and the client and business logic. This is the era where the programmers began looking at data as a separate entity, but more at an application level. Next, the e-business era paved the way for manipulating data in a more secure fashion.
Today, in the current SOA era, data is regarded as a service at the enterprise level, and the importance of data has increased dramatically. Every business alignment service must have a data model. State-of-the-art products such as DB2® 9 help to unleash the power of data as direct XML content, as required by Web services in the SOA environment.
Rational Data Architect is one of the flagship products of the IBM Software Development Platform. It is part of the IBM Information Management software portfolio and enables SOA development and implementation. RDA works with many industry-standard database servers, incluing DB2, Informix® Dynamic Server, Sybase, Oracle, SQL Server, and Cloudscape™. It supports DB2 on System z™ and System i™ platforms.
RDA is Eclipse-based and has a standardized user interface, like other products within the SDP. RDA also integrates with Rational RequisitePro® for requirements management, and ClearCase® for versioning and team work.
In many development projects, data is not given much importance compared with the application architecture. In many cases, industry-standard tools are not used for the data architecture portion of the project. Rational Data Architect helps data architects close the gap between the application architecture and information architecture, and ensures seamless alignment of data in the overall solution architecture.
RDA provides transformation from Unified Modeling Language (UML) to the Logical Data Model and vice versa. Proper usage of data modeling tools helps to avoid re-factoring of the application at later stages. Rational Data Architect is an end-to-end tool that can be used for data modeling, transformation, DDL script generation, and to build, debug, and manage database objects such as SQL stored procedures and functions.
In the SOA environment, data is critical factor for successful implementation, as the entire process choreography is all about invocation of services through messages that have the data encapsulated within them. Messaging between services and data models is closely related. The data models must have sufficient detail as required by the service. If the service has to execute complex queries or stored procedures to get the required data, then the performance of the service will be at stake, reducing the quality of service.
Like the Rational Software Development Platform suite of products, IBM Information Management software has a suite of products which handles the information layer in an SOA environment. This can be best understood with respect to the IBM SOA foundation illustrated in the figure below.
Figure 1. IBM SOA foundation
The SOA lifecycle consists of four phases
We'll look at each of the four phases below.
This phase typically deals with requirements gathering with customers. The requirements include business process models, functional requirements, and translating the requirements into architectures of various domains, including application, information, and infrastructure. Rational Data Architect helps in this domain to study the existing information model or build a new one based on the business process model and functional requirements.
The other SDP tools which are used in this phase are Rational RequisitePro, WebSphere® Business Modeler, and Rational Software Architect. Rational Data Architect has a tight, seamless integration with RSx (Rational Systems Developer, Rational Software Architect and Modeler) in the sense that it can perform a logical data model to UML transformation and vice versa. This feature helps to ensure that every application component or service identified has an underlying data component as well. ClearCase, one of the products in Rational SDP, is used for configuration management. Rational Data Architect can be integrated with it for data model version control and team development.
This phase involves the development of services, components, the underlying physical data model, and database objects. RDA supports several database servers such as DB2 (including System i and System z), Informix, and Cloudscape. RDA has wizards, an editor, and a debugger for DB2 that helps in the development of SQL statements, stored procedures and user-defined functions.
This phase involves the deployment of the components which are assembled in the previous phase. Typically the WebSphere stack is used in this layer.
The services and business processes which are deployed are managed and monitored primarily by products like WebSphere Business Monitor. Service level characteristics are calculated to ensure smooth running of the application in the runtime environment.
Let's look at the capabilities of Rational Data Architect much more closely as an accelerator for information architecture development. RDA helps in the following areas:
Most of the time, the development environment is not a "green field" environment, because there will be existing assets and databases which need to be considered for new development. Often, due to development timeline pressure, documentation might not exist, or, if it does exist, it might not be in synchronization with the physical data model in production. RDA can be helpful to handle such scenarios. RDA can be connected to the existing database using standard JDBC connectivity. The Database Explorer discovers the existing physical data model in real time, without requiring a complete reverse engineering of the existing database
Figure 2. Existing database discovery
The discovery allows users to analyze the objects that are part of the physical data model. This feature helps in identification of Information which might be required for the target application that is being developed. Using this feature, data optimization can also be achieved at an enterprise level.
Figure 3. Discovered PDM
Like UML models that are abstractions of the application architecture, data models represent the abstraction of the information model. In information engineering, data models are classified in the following categories, based on the elaboration levels that span from conceptual to the operational data store. RDA supports the creation and use of the following models:
Domain model: Domain models primarily contain domain types which are used in multiple places across the application model, for example for fields such as currency and sex, which have predefined data types or values. It is primarily used with logical data model.
Glossary model: This is used to ensure the data model uses pre-defined naming conventions in naming data entities. In most organizations, these data models are at the enterprise levels.
Logical data model: Logical data model represents the enterprise entities, their attributes (data) and the relationship among them. They are database-independent entity-relationship models that are target independent. The following are the relationships an entity can have:
- Identifying foreign key relationship
- Non-identifying mandatory foreign key relationship
- Non-identifying one-to-one foreign key relationship
- Non-identifying optional foreign key relationship
- Many-to-many relationship
In a SOA scenario, these logical entities would ideally map on to a service component which would in turn realize atomic or composite services. The logical data model primarily consists of packages and entities. Entities are representation of real world objects which can have constraints and relationship with another entity, as shown in Figure 4.
Figure 4. Logical data model
Physical data model: Physical data models are database specific and represent the database schema. They are transformations of logical data model where the entities and relationships are transformed into tables, columns, primary keys, foreign keys, and constraints.
Figure 5. Physical data model
RDA helps architects in normalization of the developed model. It checks the naming conventions, syntax, and normalization levels. There are pre-defined rules which are used to validate model and databases. These rules are also extensible beyond the available standard set. This feature helps in standardization across enterprise.
Figure 6. Model standardization
This feature is primarily used to compare models, databases, objects in databases and models. It also helps in synchronization of logical data models and physical models. Either a report can be generated or the models can be synchronized online.
Figure 7. Model synchronization
This feature helps in data conversion between TO-BE and AS-IS systems. where the databases are compared to find out the level of complexity required for data transformation to migrate the data to the target systems. It helps in comparison of columns and helps identify relationship among source and target objects. It also helps in types of relationship between source and target, such as aggregations, transformations, and arithmetic operations. It also helps in generation of data migration maps, to perform dependency analysis and also helps in optimization of target system data model. There are several techniques used in relation building though the interactive GUI and intelligent discovery. The existing mappings also can be combined to build a composite mapping.
Figure 8. Relate
This feature helps in generation of mappings between remote schema and federated schema. Transformation functions and mapping groups can be used for generating the code for WebSphere Federation Server.
Rational Data Architect provides the IDE required for DB2 object development such as stored procedures and user-defined functions. The IDE also has debugging capabilities for these objects. Apart from these features, RDA also supports the building and editing of DML statements using the SQL builder.
Figure 9. DB2 development
As Rational Data Architect is based on the Eclipse framework, it provides team support using standard tools such as CVS.
This article has introduced the basics of Rational Data Architect and how it fits into the IBM SOA foundation. Refer to the resources below to deepen your knowledge and learn how to implement it in your environment.
"Use Rational Data Architect to integrate data sources" (developerWorks, Mar 2006): Explore a tool-supported proces for federation design in just five steps.
"Access and integrate enterprise metatdata with Rational Data Architect " (developerWorks, Jul 2006): In part 1 of this tutorial series, follow a simple scenario to learn how Rational Data Architect can be used as a modeling and design tool to create a unified, virtual view across hetergeneous data sources.
"Generate SQL/XML queries with Rational Data Architect " (developerWorks, Sep 2006): In part 2 of this tutorial series, learn how RDA generates SQL and XML queries based on graphically defined mappings.
- In the
Information Management and Rational application development resource page on developerWorks:
Learn more about how to use RDA and other Rational tools with IBM Information Management software.
- Browse the
for books on these and other technical topics.
Get products and technologies
Download a free trial version of Rational Data Architect.
IBM product evaluation versions
and get your hands on application development tools and middleware products from
DB2, Lotus®, Rational, Tivoli®, and