Data Model Component

From MIKE2 Methodology

The Data Model describes the structure that holds data "at rest" within a database. One of the key factors for success in an Information Development programme is that that this data is properly structured in terms of relationships, level of granularity and completeness. The data modelling process described below is used to create these proper structures.

Contents 1 Data Modelling process within MIKE2.0 1.1 Conceptual Modelling 1.2 Logical Modelling 1.3 Physical Modelling 1.4 Physical Implementation 1.5 Summary of Approach

Data Modelling process within MIKE2.0

Data Modelling is a set of techniques used to move from a very high level expression of information requirements to the detailed physical implementation of data structures. Many new models apply either relational or dimensional techniques that are described below.

At a very high level, MIKE2.0 recommends a 4-step approach to modelling. The Conceptual Data Model should typically precede the Logical Data Model (LDM), and Physical Database Design and Database Implementation follow the development of the LDM. In some cases, a Conceptual Data Model is not developed, but it is a generally recommended approach in the MIKE2.0 Methodology that it should be undertaken during the first phase of the programme.

Conceptual Modelling

Conceptual Modelling (High Level Logical Modelling) provides a high-level business view of information, which is independent of application, organization or technical environments. Shows key entities and their interrelationships. The conceptual level is the most stable amongst the three data model levels.

Logical Modelling

Logical Modelling (LDM) is used to capture and document complete and detailed descriptions of the specific data requirements and relationships involved. This includes confirming the validity of the end users business requirements represented in the LDM (Analytical, Application, Organisational) are confirmed and validated. All entities, primary and foreign keys are specified, all attributes are fully defined, and depending on the LDM design methodology pursued (See LDM design), the model is fully normalised and cardinality is indicated.

For data warehouse implementations, logical modelling will involve additional considerations (see LDM design). The additional design considerations may extend the model to introduce time variance, derived attributes, and select aggregation and stability considerations.

A LDM provides the basic building blocks, which serve as the foundation for designing the physical database schema from.

Physical Modelling

Physical Modelling (PDM) provides specifics for the database and results in generation of DDL statements for creation of the database. May be structured different than logical model for reasons such as performance.

Physical Implementation

Physical Implementation relates to the final loading of schema into the database and any optimisation and security steps requited that are specific to the physical environment.

Summary of Approach

Through the phases of the data modelling lifecycle, the audience shifts from the end users of the proposed system to the technical staff that will design and build it. Consequently, as the model’s level of detail and complexity increases, the user presentation focus is also changing.

The basic components of a Data Model (the ones that have to come first) are Entities and Relationships. A diagramming tool is typically used to represent the model.