What is the role of DBAs in Managed Database Platforms?

Consider the scenario of using a database for your Enterprise application. In the on-prem world, this starts by acquiring the right hardware, deploying the database software, configuring for high availability, backups, and the whole nine yards. These activities were driven by Database Administrators (DBAs), that is considered a specialized role within IT teams. DBAs had a deep understanding of the 100s of DB configuration parameters, query language nuances, KPIs to monitor, etc.

In the Cloud-age, applications are increasingly being provisioned in the cloud. There are three ways to provision a database:

• Spin-up your own (similar to on-prem, but minus the need to manage hardware and OS installation.
• Managed Database Service: As the name suggests, the focus is on optimizing the application usage of the database. All the other activities are offloaded to the provider. While the cloud provider does the heavy-lifting, the Enterprise IT team is responsible for the specifying the policies for security, availability, durability, auto-scaling, etc.
• Serverless Managed Service: This takes the managed Database service model to the next level. The idea is to think in terms of jobs/tasks. It further relieves the IT team from tracking utilization in the context of cost optimization (an increasingly important piece of the equation especially as Enterprises move all-in to the cloud). In the serverless model, compute resources are spun-up on-demand in the context of the job/task.

Given the ease of consumption of managed database platforms, the Enterprise consumption model has shifted from a one-size-fits all database, to selecting the right database for the job.

Given these shifts, the role of the DBA has been clearly evolving along the following three dimensions:

• Instead of being deep-in-the-weeds of single database technology, being able to pick the right database for the given application requirement.
• Instead of learning to tune 100s of knobs, focussing on policies available for database services and how to automated APIs (infra-as-a-code).
• Instead of assuming high quality data being ingested within database, focusing on the data pipeline aspects including data-sources, data quality, etc.

In summary, the role of the DBAs is undergoing a transformation — the North Star for this transformation is to avoid undifferentiated heavy-lifting. DBAs will continue to play an important role in efficiently transforming Application SLAs into Database Management policies. Given there is no standard language for specifying policies across cloud providers, this will continue to be an evolving topic. Overall, given the DBA’s proximity to data, I would expect the role to evolve into parts of the Data Scientist focussing on data pre-processing, cleansing, etc.