Database maintenance, key to good data governance

Who is in charge of database maintenance in a data governance program? The solution is that it depends on the scope of that program.

One of the principles of data governance is that company data does not belong to people. Data is an asset that belongs to the company. Even so, These data must be managed and some institutions assign responsible personnel to whom they make the data owner and who is in charge of maintaining the database, while other companies deviate from this concept of data ownership.. Let's look at these two different ways to manage databases.

Databases with assigned personnel

Through this approach we have people who are given the responsibility of being able to work with the data.. Roles are assigned to them and one of them is designated as the data owner. This data owner helps coordinate all responsibilities for that data.

In some institutions, this approach is used to solve a critical part of database maintenance, access management. These data owners have the power to choose who can have access to company data.. What they do is that once a person requests access to the data and is given permission for that access, someone technically grants physical permission to enter the app, database, any other type of data store that contains the data. that person wants to enter.

In many institutions, this access management function is carried out through data governance. But before I tell you to use the term data owner in data governance, you need to understand your organization's approach to entering management, if the concept of data owner exists and what responsibilities do you have.

If you decide to use the concept of data owner, it is essential that you clearly define the responsibilities of data owners versus data stewards.

Database maintenance with federated responsibility

Company data, by its own nature, flow through the organization, touching many technical and commercial processes and being stored, moved and transformed by many IT systems, so in the end a lot of reports and data coexist. data sources.

In small companies or institutions with simple data flows, it may be feasible to trace some types of data from creation and across all data streams. In these cases, Responsibility for this data can be assigned to a person who would be responsible for ensuring that all steps are carried out safely and enforcing quality rules and any other types of standards and rules associated with maintaining the database..

Despite this, this is not feasible for most large or complex companies. For this, the concept of data owner may not be useful. What they do in these cases is that they have another approach: federated responsibility for data.

With this approach, first documents the path the data takes since its creation on a specific system. After, data responsibilities are assigned for a manageable number of segments. But this approach brings us several challenges:

1. Documenting data flow can be very tough work. What if it is not finished on time? Sometimes you have to assign responsibilities first and that includes the investigation.
2. Assigning Responsibilities Can Be Difficult. If these responsibilities are in addition to other existing responsibilities, people may feel overexploited.
3. Managing and coordinating group responsibilities takes time and effort. You will need documentation to assign roles and responsibilities for the data.
4. It may be necessary to designate a lead person. This person should be familiar with database maintenance, should be able to discover who is assigned responsibilities for each segment of that flow and should be willing to involve resources in the investigation, problem solving and impact analysis.

This approach to database maintenance with federated responsibility has a disadvantage: is more complex. But it also has a significant advantage: it works.