image

Tech2Tech

Insiders Warehouse

Where To?

Industry logical data models serve as maps to an organization's information.

Whether you're traveling or designing a data warehouse, maps and models can help you navigate the environment and home in on the details that could influence or affect your journey.

Just as it is important to understand the layout of the surroundings in order to overcome potential obstacles while traveling, when creating and maintaining a data warehouse, it is crucial to visualize the relationships between different components for a complete and precise view of your organization. Identifying these details and relationships through maps and models will enable you to make the best use of your time and tools.

In IT a complex information landscape is made understandable via a data model. Using symbols and text, a data model helps developers and analysts better understand a set of data elements and the corresponding business rules. While a data model can represent a specific functional area, such as order processing, it can also represent the organization's master plan in the form of an enterprise data model (EDM).

Organizations benefit from the mapping functions of an EDM. As a subject-oriented and integrated data model, an EDM describes the data produced and consumed across an entire organization. Subject-oriented means the data model concepts fit together as the CEO sees the company—the big picture—as opposed to the perspective of the individual functional or department heads.

For instance, as captured in the database, the same person could have multiple roles in an organization, like that of both customer and employee. On a department level, the person might appear in only one application. But from the subject-oriented level, while the integrated EDM will show Individual Last Name only once, it will capture the mapping back to each pertinent application, thereby identifying that person as a customer and an employee.

Resource and skill challenges are often inevitable when creating and maintaining an EDM. So instead of reinventing the wheel, more organizations are purchasing pre-built industry data models as starter EDMs. In essence, these industry data models capture how an organization in a particular industry works or should work. Teradata offers eight industry logical data models (iLDMs), which companies can build on to create an EDM tailored to their specific organizations.

Map it out

Integration is difficult to achieve for many organizations, yet it is a prerequisite for most enterprise-wide initiatives. As helpful resources, the EDM and industry models, such as the iLDM, provide companies a single, integrated view of their business information.

The EDM integration for a hypothetical organization is illustrated in the iLDM geography structure shown in the figure. This figure, with the table, shows how the iLDM maps the places where the field is stored to the EDM structure.

As portrayed in the "Source field" column of the table, three data elements map to the Electronic Address Txt listed under the "EDM field" heading: email, e_mail_addr and company. Even with such a simple example, complexities are evident—the same address is entered in multiple locations and valid e-mail addresses are parsed out from the company data element. These challenges make it obvious that the maps are pertinent to achieving enterprise-wide integration through an EDM.

The EDM and a map of where the data is stored offer the following benefits for organization-wide initiatives:

sample image

Enterprise resource planning

As an information system, enterprise resource planning (ERP) integrates all functional areas for an entire enterprise. While ERP systems manage broad areas like manufacturing, human resources and finance, an EDM becomes an invaluable tool by enabling information to be viewed at an integrated subject level. An EDM can assist ERP efforts by:

  • Evaluating whether a particular ERP package is a good fit for an organization by matching a subject-level EDM against a subject-level ERP model and highlighting the pros and cons. Also, mapping the EDM to existing applications can provide insight into which systems will replace an ERP package versus which systems will send and receive data to and from it.
  • Performing internal and external impact analyses to the ERP package. Building a high-level model as part of an ERP system within the EDM can help identify touchpoints and overlaps within ERP rollouts. It can also communicate how the ERP system interacts with other systems and functional areas.
  • Master Data Management

    All data that is shared among applications is considered master data. Master data usually refers to reference data, such as customer and product, but it can also include event data, such as orders and credits.

    An EDM forces a single set of metadata—the text that describes what the users need to see—for each data element. This metadata includes a single name, definition, format, domain and set of business rules.

    Defining all of this data at an organization level makes master data management (MDM) possible. In addition, MDM identifies which applications use which EDM structures. The EDM becomes an integration point among applications and provides a foundation for modeling efforts in new applications.

    Business Intelligence

    The Data Warehousing Institute (TDWI) defines business intelligence (BI) as "the processes, technologies and tools needed to turn data into information, information into knowledge and knowledge into plans that drive profitable business action."

    BI encompasses data warehousing, business analytic tools and content knowledge management. The data warehouse is responsible for turning data into information, which it does by using the two main ingredients of the EDMs—the required subject-oriented and integrated views. This is the reason the EDM is essential to the data warehouse. In addition, as an industry-specific yet customizable data model, iLDMs provide the underlying structure necessary for the successful implementation of an enterprise data warehouse (EDW).

    Data Quality

    By mapping data to existing applications, an EDM will expose an organization's data quality issues. It can show, for example, customer information that is updated in more than one application, or product information that is being retrieved from an incorrect source. The example in the table shows that Electronic Address Txt has three distinct sources (HR Payroll, Supplier Contact and Consumer Contact), each of which can update the e-mail address. Furthermore, data quality issues can also be prevented as new applications and projects leverage the data model.

    Employee Development

    An EDM is a great training tool for employees new to the industry, company or department. Providing insight into how an organization works enables staff members and consultants to quickly come up to speed, so they can contribute intelligently to a project.

    sample image

    Use of the iLDM

    The iLDM represents a business solution and, therefore, is independent of a specific database or reporting tool. It is constructed based on the commonality of the subject and logical EDMs for companies within a particular industry. Once implemented by the organization, the iLDM creates a unique, customized EDM that captures the concepts and rules that are important to the enterprise and maps out the data to the company's particular software and hardware infrastructure.

    Along with the geography example shown in the figure and detailed in the table, the following sections demonstrate how organizations can leverage the iLDM:

    AS the EDM

    With the exception of some customization, the organization will use the iLDM as its EDM. While the systems might look different from the iLDM Address structure as displayed in the figure, the iLDM becomes the organization's enterprise view, and any existing models must be mapped to it.

    FOR Input to the EDM

    A number of enterprise-type models most likely already exist in the organization, and the iLDM is another model that needs to be integrated. It can either be the focal point that all models map to, or it can be just another model used in the organization. Going back to the example in the figure, the Address structure becomes one of the inputs to the EDM as well as to one or more other models. Each of these must be reconciled, and the most appropriate sections of each model are combined to create an agreed-upon EDM.

    In these cases, the organization might decide to rely more heavily on the conceptual data model (CDM) to provide a high-level overview to the iLDM. The CDM is about 10% the size of an iLDM and shows only the most important entities in an iLDM and their relationships. For instance, the Teradata Manufacturing Logical Data Model has slightly more than 700 entities, while the manufacturing CDM has about 70 major entities.

    Also, if existing EDMs are integrated into one current and complete picture, the CDM could provide a high-level roadmap.

    To Identify Gaps

    As is evident in the figure, a fairly robust Address structure is part of the organization's EDM. It is by means of a detailed iLDM of this sort that the organization can identify missing data elements and skewed relationships in its model.

    In addition, the iLDM can provide different views of existing information. Perhaps "telephone number" is not currently considered a type of address, but it becomes apparent through the data model that it is beneficial to move "telephone number" to an address field.

    Destination driven

    Nearly every corporate-wide initiative needs an EDM that maintains a well-structured and coherent map to its data. An industry model, such as the iLDMs provided by Teradata, can save an organization time and money by facilitating the task of creating a tailor-made EDM. By leveraging best practices, maps can be tracked, data relationships identified and missing gaps filled. In the end, the organization can be assured that its data is accurate, current and of overall high quality.