Data Model Standardization
Uploaded by
Michael CorselloData Model Standardization
Uploaded by
Michael CorselloRF
Corsello Research Foundation
Data Model Development
Future Concepts for Evolving Data Models
Purpose of a Standard
A standard exists as a reference point. A standard is a defined set of guidelines or best practices. The purpose of any standard is to achieve a uniformity that can be relied upon across some domain. Any individual or group that could have use for a standard is a domain of that standard.
Corsello Research Foundation
RF
Data Model Implied Requirements
Upgrade Path from previous versions of the model Extensibility of the standard Customizability of the implementations derived from the standard Mappable by business domains / communities Non-repetitive Comprehensive
Standards compatible for existing external standards
Translatable by third party vendors Automation capability for tool support Integration points for connectivity to external systems and standards Vendor neutral
RF
Corsello Research Foundation
Foundations
RF
Corsello Research Foundation
Foundations for a standard
Meeting the requirements requires more that a data model. Several dependent standards need to exist to support the overall model.
RF
Corsello Research Foundation
The Pieces
Logical Data Model - The logical data model is the logical representation of the entire set of data entities modeled in the standard. This includes the definition of all data primitives (geometries, numeric, logic, text, date, etc.), logical types (table-like structures representing real-world things), and other extended primitives such as restricted domain values of other primitives (e.g. pick lists). Business Domain Perspectives (Profiles) - Business domain perspectives are logical aggregations of entities from the LDM that have the greatest relevance to a specific functional business domain (community of users).
Logical Associations - Logical associations define which entities are permitted to be connected via relationships. This is theoretically similar to a relational data model which defines the relationships that are permitted.
Rules - Rules are a combination of simple textual documents and machine interpretable data that define a specific rule set related to the standard set. Data type conversion rules define the algorithmic rules for converting between data types (e.g. 64-bit signed integer converted to a 32-bit unsigned floating point type). Encoding Standards - Encoding standards define the structure of physical implementations of all elements of the standards.
RF
Corsello Research Foundation
Fitting the Standards Together
The LDM defines the data entities.
The business domain perspective creates an aggregate view into the LDM.
The table associations define which entities can be related. Finally, the vendor mapping allows a tool to create the physical implementation from the LDM.
RF
Corsello Research Foundation
Vendor Specific Data Conversion
To exchange data between different vendor specific implementations, there are 2 primary steps: Conversion from vendor 1 physical implementation to the LDM vendor neutral implementation Conversion from the LDM vendor neutral implementation to the vendor 2 physical implementation
RF
Corsello Research Foundation
Logical Example of Physical Implementation
User Selection LDM
Custom Extension
Physical Output
RF
Corsello Research Foundation
The Logical Model
RF
Corsello Research Foundation
Logical Data Model (LDM)
The LDM is broken into logical packages.
The foundation package contains the primitive descriptions that define core data entity types that the remainder of the standard will reference.
the root LDM is broken into packages defining functional areas that encompass responsibilities over physical phenomenon to be mapped.
RF
Corsello Research Foundation
Physical Model Description
In order to permit the dynamic generation of physical implementations of the LDM, a set of guiding standards also need to exist. This includes the rule sets that:
Describe how to convert between data types Associations on how to describe the permitted table associations Grammar for describing how a specific toolset will implement the LDM as a physical structure Grammar for describing custom data entities
RF
Corsello Research Foundation
Backtrack
Requirements Coverage Conclusions
RF
Corsello Research Foundation
Requirements Coverage
By developing the model through this methodology, all of these requirements can be met effectively and efficiently:
Upgrade Path from previous versions of the model
Extensibility of the standard Customizability of the implementations derived from the standard Mappable by business domains / communities Non-repetitive Comprehensive Standards compatible for existing external standards Translatable by third party vendors Automation capability for tool support
Integration points for connectivity to external systems and standards
Vendor neutral
RF
Corsello Research Foundation
Conclusions
While the suggested concepts for developing a standard model are but one potential approach, the authors and supporters of this concept believe it is a well-defined, sustainable, affordable approach to evolving the standard. Comments, questions, suggestions for improvement and general feedback on this concept are strongly encouraged by the author.
RF
Corsello Research Foundation
Questions
RF
Corsello Research Foundation
