THE ROLE OF THE DATA ARCHITECT

THE ROLE OF THE DATA ARCHITECT

(Part 2 of 4)

By W H Inmon

The primary responsibility of the data architect is to build, maintain, and protect the foundation of data on which the corporation does its analytic processing. To this end there are many diverse activities that make up the job of the data architect. While the data architect may become involved in production systems, it is usually the role of the data base administrator to care for and tend production systems. The data architect has a strategic perspective while the data base administrator has a more tactical perspective. Of course, there may be overlap between the roles on occasion.

That foundation for analytic processing started from data thrown off by the execution of applications. Then the artifact data turned into the legacy environment. Ultimately the data ended up as the foundation of data on which to do analytic processing. The data that was thrown off was never designed to be a cohesive whole. Thus, the need for care and tending by the data architect.

In no particular order here are some of the activities and responsibilities of the data architect –

1)????? Ensure that the foundation of data supports the strategic objectives and directions of the organization.

2)????? Make sure that the foundation of data is available and open to anyone who needs it.

3)????? Be able to find data in the foundation when needed, in assistance to analysts who are lost and/or intimidated by the foundation.

4)????? Be available to resolve issues in conflicts and anomalies of data. This means that the data architect must explain and/or resolve issues that arise in the usage of data. (This can be a tough one.)

5)????? Ensure completeness and timeliness of data in the foundation. It does no good to have incomplete data or data that is not current on which people want to base decisions.

6)????? Ensure that data with a high probability of access is immediately available and that data with a low probability of access is stored inexpensively but still be able to be accessed if needed. This includes the periodic movement of data from high probability of access of data to low probability of access of data.

7)????? Understand the data model and the ramifications of the data model to the foundation of data.

8)????? Understand the sources of the data in the foundation and the algorithms that are used to pull the data into the foundation.

9)????? Ensure that when fast response time is needed that the data needed to support fast response time is available.

10)?? Monitor the foundation of data on an as needed basis, looking especially for the growth and usage of the data in the foundation.

11)?? Add new types of data on an as needed basis. Make sure that the new data is integrated and documented. Review any tools brought into the organization.

12)?? Review any new software and its output to determine the fit or lack of fit with the foundation of data.

13)?? Care for and maintain a catalog of where foundation data can be found. Among other things the catalog needs to contain the lineage of data.

14)?? Hand code data amendments occasionally.

15)?? Make adjustments to the corporate data model occasionally.

16)?? Understand the rate at which data is flowing into the foundation.

17)?? Be aware of end user data activities and sources that may affect the foundation.

18)?? Include structured data, analog data and textual data in the foundation in a cohesive manner.

And this is just the short list.

Some of these tasks are immediate and tactical. Some of these tasks are strategic.

Strategic tasks include the alignment with corporate objectives. The most strategic is the separation of low probability of access data from high probability of access. Most corporations do not make note of this distinction. And those corporations who do not accommodate this distinction –

?? Spend far more money on their IT infrastructure than they have to

?? Make life difficult for the analyst trying to use the data in the foundation.

?? Create unnecessary obstacles in the usage of data.

?? Burn major amounts of unnecessary machine cycles

Another strategic concern is that of the creation and maintenance of the catalog of where data resides. The foundation of data is disorganized and has grown in a random fashion. You need to have a guide as to what data is there and what data isn’t there.

So there is a lot to becoming a data architect and doing the job well.

?

Bill Inmon lives in Denver with his wife and his two Scotty dogs – Jeb and Lena. Today is a typical winter day in Denver. It snowed 6 inches this weekend and is bright and sunny today. Jeb has been rooting around the back yard, enjoying the sun. When he came to the door his face was full of snow.