This column is adapted from the book Universal Meta Data Models by David Marco and Michael Jennings (John Wiley & Sons).
In the first two installments of this series, you were introduced to the complete meta data model and how the various model components interrelate in the universal meta model environment. In this installment, we will drill deeper by focusing on the enterprise systems component area of the model.
An enterprise is typically defined as an organizational unit, usually large, that utilizes information technology. The term's use here goes beyond the data found in operational or decision support applications; it includes all data stored and analyzed in an organization. The enterprise systems component represents data structures, data movements, data transformations and expressions constructs.
The purpose of the enterprise systems component is to define all of the data structures in the enterprise and describe how these structures interrelate in the meta data domain (see Figure 1). This model serves as the core for all subsequent component models in this series to control and manage both data and meta data enterprise-wide. The enterprise systems model offers insight into data, relationships, and the data movement and integration processes. The model serves as a single reference source for system semantics and data structures. By capturing and sharing the information about our systems, the model allows for a greater possibility of achieving integration throughout the enterprise.
The enterprise systems component model encompasses several subject areas. The Data Package subject area covers the interrelationship of data and data structures to application systems, data stewards, business rules and data movement activities. The Data Package subject area is the central focal area for the other component meta models (Business Rules, XML and IT Portfolio models). The Data Movement and Transformation subject area documents the general mapping that a set of data sources has to a set of data targets. The Expression Transformation subject area supports data detail transformations through use of an expressions model using an abstract syntax tree method. The Transformations Operational History subject area captures auditing information and data transformation related meta data. The Data Profiling subject area supports the identification of domain violations in relation to business rules and data quality during detailed expression transformations.
Figure 1: Enterprise Systems Component Model
The Data Package subject area includes the logical relationships of data abstracted for use in capturing and integrating data within the enterprise, detached from the source or target systems. This subject area provides the basis to control, manage and integrate information and meta data throughout the enterprise. The structuring of data in the enterprise model has three major components. Seeing how these objects relate to concrete structures such as a relational database provides insight into this generalization's workings. A packaging-level DATA PACKAGE holds high-level structures and corresponds to items such as relational database catalog, database schema, flat files, business transactions and messaging. The DATA GROUP is the lowest level grouping of data and corresponds to a relational database table. The DATA GROUP holds a DATA ELEMENT, which is a unit of addressable data. The DATA ELEMENT corresponds to an attribute from a logical perspective and a column in a relational database table from a physical viewpoint. The DATA ELEMENT may a have a domain that describes the data values it may inherit. The entry point of the model is the DATA PACKAGE, which has relationships with the DATA GROUP. The relationship of DATA PACKAGE to DATA GROUP to DATA ELEMENT provides the key architecture for all data structures in the enterprise and the managed meta data environment (MME). The hierarchy of data for the enterprise is documented here, allowing data objects to be grouped, packaged and reused in multiple occurrences.
The Data Movement and Transformation subject area describes the general types of data integration and mapping activities that occur in the enterprise as well as the tools and processes used to perform the tasks required. The transformations fall into two categories: 1) black-box, where the exact structure of the expression may not be known; 2) white-box, where the expression used for the transformation is parsed and mapped out explicitly. Black-box corresponds to the SOURCE TO TARGET MAP entity, and white-box to the Expressions Transformation subject area.
The Expression Transformation subject area provides a detailed breakdown of those transformation steps that are expression based. Expressions contain two types of components, operands and operators. Operands are the items that are changed, and operators are the symbols that represent specific processing. All expressions have at least one operand. An EXPRESSION NODE is denoted as an identifier, string literal, constant, string constant, array specification, function or procedure call, or any combination of these, in order with operators. Expression elements are parsed and stored as OPERATION ARGUMENT LIST, OPERATION NODE, ELEMENT NODE or CONSTANT NODE in an abstract syntax tree structure.
The Transformation Operational History subject area captures operational meta data, or the actual executions of the transformations and the groupings that executed them. This is pertinent information collected during the transformation procedure used for measuring status, processing efficiency, capacity planning, source data quality analysis, historical comparison, error diagnosis and other purposes. Operational meta data is collected for both the general transformations (the relationship to TRANSFORMATION GROUP) and for detailed transformations (the indirect relationship to EXPRESSION NODE). Meta data is collected at each hierarchal level of the transformation through the relationships between the Data Movement and Transformation and the Transformation Operational History subject areas. The relationship of a package audit to one or many process audits and process audits to one or many steps is captured through the hierarchy relationship between the entities.
The Data Profiling subject area provides the MME with the capability to identify and historically track business rule and domain value violations in an enterprise's data brought into the enterprise during transformations. This subject area also provides valuable relationships from domains and business rules to data quality metrics that may need to be further assessed. Measurement is performed against the quality of element values (single values, range specification or pattern) in data stores.
Next month, we will take a closer look at the IT portfolio management component area of the complete meta data model.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access