The Responsibilities of an Enterprise Data Architect: - Continued.....
Several groups have responsibility for aspects of data in the enterprise, including application teams, data warehouse managers and the integration competency center, as well as business people. The EDA focuses on data that moves through the enterprise, and must coordinate work with these groups and other stakeholders. Various teams in the IS organization and the business will execute the more-granular tasks; the EDA will provide overall leadership for driving the enterprise toward an overarching data architecture vision.
The EDA's responsibilities include:Mapping data sources:
An understanding of where data is stored and maintained is essential to data architecture. The data map should include descriptions of the business meaning of the data, its uses, its quality, the applications that maintain
it and the database technology in which it is stored. Documentation of a data source must describe the semantics of the data so that the occasional subtle differences in meaning are understood. Because business processes increasingly involve business partners, some data sources will be outside the enterprise. These data sources should not be ignored.
Although the application teams are responsible for documenting the details of application data, the EDA must know where data is maintained and accessed throughout the enterprise. Application enhancements that result in the maintenance of additional data subject areas are of particular interest, because these have the potential to increase data fragmentation and redundancy. As described in 'The Architecture Engagement Process', a process must be in place to evaluate the impact of project proposals on enterprise architecture. This process ensures that the EDA assesses the impact of these changes on the architecture.
Documenting interfaces and data movement:
Having mapped its sources, the next step in understanding enterprise data is to record how it is moved around the 'virtual' enterprise. This includes the frequency of movement, the source and destination of each step, how the data is transformed as it moves, and any aggregation or calculations. The application team, the integration competency center and the data warehouse team play vital roles here, with the EDA coordinating their efforts and fostering consistency.
Designing the movement of data through the enterprise:
Having documented the status quo — the sources of data and how the data is moved around — the EDA can then look at how this movement can be improved. Some changes might be obvious, such as eliminating unnecessary movement. Others might be more complex, involving brokering agreements between business units about sharing interfaces. Again, groups such as the integration competency center and the data warehouse team do much of the detailed work, with the EDA identifying and coordinating reuse opportunities.
Defining integrative views of data:
These views will draw together data from across the enterprise. Some views will use a database of extracted data — others will bring together data in 'near real time'. The EDA works with business people and application designers to identify and model these integrative views and determine the quality of service requirements — data currency, availability, response times and data volumes.
Designing canonical data views:
Because the format and the semantics of data differ from application to application, data must be transformed as it moves from its source to its destination. Without a canonical view of data, each interface will perform a unique point-to-point transformation, with the number of these transformations proliferating exponentially and becoming an enormous burden. A shared, canonical view makes these transformations more manageable, with the data from the source being transformed once to the canonical view, and once more from the canonical view to the target view. If an integration competency center does not exist, the EDA must define the canonical views. If it does exist, the EDA must work with this group to ensure that the canonical views assist as broad a range of transformations as possible.
The Responsibilities of the Enterprise Data Architect - Continued.....
Articles Index Click here to submit your Article.