DB2 magazine Chinese version to gain flexibility through design

On March 16, IBM announced rational Data Architect (RDA), a new eclipse based tool that allows you to discover, model, visualize, correlate, and develop distributed data assets in different formats.

RDA is part of the WebSphere information integration platform, a concrete implementation of the information Integration Services layer in the IBM information on Demand landscape.

The WebSphere information integration Briefcase enables organizations to perform five key integration functions:

To connect data or content stored anywhere

Understand and analyze information, its meaning and relationships

Clean the information to ensure the quality and consistency of the information

Transform information to provide rich, customized information

Make the information federated so that people, processes, and applications that need it can access it

Under these features is a common metadata and parallel processing infrastructure layer. By publishing shared services, these features can be used in a service-oriented architecture (SOA).

RDA belongs to the understand (understanding) feature. It helps create a "big picture" of the entire information infrastructure. With RDA, you can design new databases, reverse-engineer existing data sources, perform mapping between data sources, and generate the code needed to implement federated concepts in a database.

RDA can be run on Windows (XP, 2003) and Linux (SUSE Linux 9, RedHat ES Linux 3.0) platforms, using the same platform as other Rational design and construction tools, and therefore scalable and ripe The user experience. For example, Workspace is the root location of all the information that RAD will permanently save, and it is common to all Rational tools. Project is another concept that is familiar to Rational users, and projects are created to describe the type of work that will be handled. For example, in a Workspace, you will find a data design project and a data Development project. This structure is shown in Project Explorer, which is a navigation tool for RDA and most of the Eclipse based tools.

In this article, I summarize some of the key features of RDA into the following categories:

Design and modeling

Understanding Relationships

Change Management

Design and modeling

Design and modeling capabilities include logical data modeling, physical data modeling, and the mapping of different data sources, and integration of different data sources into a federated schema. Standards, such as naming, can be implemented together with other rules, such as first, second and third paradigms, and excessive use of indexes.

Logical data modeling. RDA allows data architects to perform logical data modeling tasks, such as defining entities with attributes, keys, constraints, and attributes. By defining relationships, entities can be interconnected, and relationships between entities can be equivalent (identifying), non-equivalence (nonidentifying), many-to-many relationships, or represent a generalization (generalization). A designer can create an entity diagram that starts with a blank logical model, or a template, or a reusable model that contains standard or core enterprise entities. These entity diagrams can include descriptions of entities and attributes, with comments (see Figure 1).

Figure 1. This logical data model illustrates attributes, primary keys, foreign keys, equivalence relationships, non-equivalence relationships, and cardinality.

RDA provides a set of predefined data types for the logical data model. In addition to these predefined data types, you can add user-defined data types, called domains (domain). A domain can be defined in a logical model or defined in a different domain model, and other models can reuse the domain model. RDA allows the creation of a standardized data dictionary containing user data types that can be shared by different project data models. The types of fields include atomic fields (based on logical data types, with constraints), list fields (with constrained enumeration values), and Union Fields (a federation of logical data types with constraints).

By using constraints and relationships, you can define simple business rules. But RDA can do more. The Analyze model feature displays a modal analysis window that allows you to specify which predefined analysis Windows should be executed. The built-in RDA rules include:

Duplicate relationship

First Paradigm

Second Normal form

Third normal form

Excessive indexing

Model Syntax Check