This means that many DBAs are working in teams with developers instead of in teams of other DBAs, at least for periods of time when development projects are very active.
Regardless of the pattern, the clear trend is that DBAs are more application-savvy. And this can be a good thing because the DBA can better administer the database for an application’s needs as it progresses from development to test to production.
Cloud and On-Prem Skills
Perhaps the most significant computing trend these days is the adoption of cloud computing. When cloud services are involved, the role of the DBA will change. The first significant change is that DBAs no longer need to plan for the on-prem hardware required for each new database instance.1 Cloud deployment changes this because the requirements are sent to the cloud service provider (CSP), who provisions the database instance on its hardware. This does not mean the DBA does not have to be involved in capacity planning, just that it changes how and when it is performed. Failing to plan and monitor for capacity spikes can result in large bills from your CSP.
When the database is delivered as a cloud service, it becomes easier to scale because you do not need to physically procure additional resources such as storage or servers. Another DBA responsibility that becomes easier is backup, which is typically managed by the cloud provider. Of course, DBAs still must understand recovery time objectives (RTOs) and ensure that the cloud provider is backing up such that recovery can be achieved as required by the business. You cannot simply say, “Oh, the cloud provider is backing things up so we don’t worry about that anymore.” In other words, the DBA must understand where failures can occur—both on prem and in the cloud—and be prepared to remediate problems that impact data access.
Furthermore, availability can become less of a concern because the cloud can replicate databases across multiple geographical locations. Keep in mind, too, that cloud DBaaS providers manage the maintenance, upgrading, and the application of fix packs to the DBMS. But, the DBA still must work with the CSP to know when maintenance is being applied and if there are any impacts (feature deprecation, new keywords, etc.).
The exact impact of cloud use on your DBA team depends on the specific cloud database services that will be deployed. Is it a full DBaaS, are you doing IaaS and managing the DBMS yourself, etc.? There is no cookie-cutter answer, so the DBA team must understand the services being provided and ensure that they conform to the needs of the organization’s applications and end users.
A fully managed DBaaS in the cloud can significantly minimize the DBA role for day-to-day management and many routine tasks, but there is still a need for strategic data management and database expertise that can be provided by DBAs. Indeed, freeing DBAs up from having to perform manual tasks will enable them to think more strategically and to be an organization’s custodian of the relationship with the DBaaS cloud provider.
Additionally, a DBA team must be able to take on new responsibilities in terms of dealing with the database in the cloud. The first challenge is how to move data from on-prem to the cloud. It usually is not as simple as unloading from a local database and reloading to a cloud database.
Another aspect of the DBA’s job that should take on additional importance with cloud deployment is security. Most cloud providers offer secure services and mechanisms to control access to data, but the proper application of data security remains the responsibility of your organization.
The DBA must also understand and participate in the budget management process for the cloud database. Just because it is easy to scale up using cloud database services does not mean there is no cost, and organizations need an expert who understands the impact of additional workloads on your DBaaS contract. Additionally, the DBA must have knowledge of the impact of database and application design and architecture decisions and how they impact costs.
1 Of course, with virtualization, the requirements may not have been physical hardware, at least not initially.