Click here to learn more about author Tejasvi Addagada.
Information Architecture essentially encompasses data and technology architecture. There is enterprise value that is realized by including Information Architecture as a discipline in Data Governance. Data Architecture is more about having to manage the business meaning of data in a consistent way. While Technology Architecture defines strategy and design of the technology components that work in tandem to achieve these data capabilities.
There is a necessity for business to put forth a common understanding of data and its attributes for Data Architecture Governance to succeed. This should be embedded naturally while Metadata Management service in Data Governance would ensure that common definition of business terms, data elements and their relationships.
Data Architecture further governs physical implementation associated with Technology Architecture. Data Architecture also reduces the risk of not using the right data to satisfy the business requirements. It also reduces the effort spent by Analysts and SMEs in understanding current and new data capabilities.
- Generally, the first activity I suggest is to identify logical and physical data domains and datasets. For example, the logical classification for any industry would be Customer and Finance.
- Then, classify the data into identified Logical and Physical domains and datasets identified.
These activities will further simplify the disparate data landscape, while creating clear boundaries for classification and understanding ownership at high level. We will look at having governance activities that are defined by policy definition around architecture later in the blog.
Semantics refer to the adoption of precise, shared, and consistent business meaning of data across the enterprise. This is what governing data as a meaning stresses on. The challenge for organizations have been the inability to harmonize disparate data across an enterprise or a business function as the meaning of data has never been standardized. This primarily has been due to the in-organic growth of data in organizations due to nonexistence of planning, Data Management, mergers, acquisitions, etc. Semantics is the discipline of assigning unambiguous meaning to data throughout its lifecycle.
Semantics are usually associated with the business architecture where business objects represent the significant informational and conceptual elements in business context. A passive entity that represents business, products or service in its business context is a business object.
Techniques like Semantic Models, Business Metadata Elicitation, Logical Classification, and Data Entity views can be used to capture precise consistent classifications and definitions, while aiding the simplification of the disparate data landscape.
Conceptual Level describes data at its highest level, identifying the critical data objects needed to satisfy a business objective while also defining their relationships to one another. The data objects are captured in the views of the application architecture.
The passive counterpart of the business object in business architecture is the Data Object in the application architecture level.
Logical Level is a fully attributed conceptual model that has been abstracted from the physical implementation. The logical model represents the business requirements in terms of what is needed to satisfy the objectives of the business function or service. The Logical Data Model can also be associated with the Data Objects in the application architecture. Techniques like Logical Models, Business and Operational Metadata can be used to capture precise, consistent definitions of data objects.
Physical Level is the instantiation of the meaning, relationships, and attributes of data into a physical implementation. The physical models are usually attributed to the technology architecture.
A physical piece of data is called a Data Artifact as is produced, consumed, and stored by the software application process or a software service. It models typical messages – extracts scripts, database tables, and documents. Example – Lead campaign extract.
Let us further understand the Lineage and relationships between various levels of data in an enterprise.
We will talk more about Metadata Management and Content Management as services in future blogs that will essentially cover governance guidelines associated with Data Architecture. The service definition would also define the right operating models, stakeholders, processes, and frameworks to embed the policy and guidelines into the organization.
Technology Architecture is usually associated with the definition of strategy and implementation of physical architecture as informed by Data Architecture. The guidelines for architecture ensure that the tools, platforms, and associated infrastructure components are realized by business services to achieve maximum efficiency in operations. Guidelines to govern technology architecture include acceptable database platforms, tool stack, storage guidelines, and risk assessment and management. These activities should be embedded organically and with ease into the project lifecycle processes of the organization while also including the Governance Committee in the strategy analysis, requirement, and design phases. Organizations should ensure that the Architecture Review Group takes an active role in enforcing Information Architecture Governance.
There is a need for a consistent representation of the current technology architecture to aid decisions associated with defining the strategy and the future roadmaps that come with new business capabilities. Towards the left, is a simple representation on the existence of data and relationships at various levels of architecture. I prefer to use Archimate standards to document the views related to data.
There are other techniques for representing data at various levels described in this blog by me on dataassociation.net.
Below is a snapshot of the business, application, and technology layers, the various data components (marked in green) in existence, and the relationships defined among them.