Finally, it is improbable that all of an organization’s databases will move to the cloud, at least initially. That means DBAs must be able to work in a hybrid environment, maintaining their traditional skillset for on-prem data, while embracing and extending their capabilities to manage the cloud data.
Supporting Different Types of Usage
Most DBAs are skilled at managing traditional batch and transactional workloads in terms of the three core DBA responsibilities: performance, availability, and recoverability. But there are many additional types of usage that must be supported by DBAs.
More and more workloads are analytical in nature. DBAs support such use cases differently in terms of performance requirements, latency, and so on. Today, traditional batch, online transactions, and data warehousing all remain important workloads that DBAs must continue to support, but additional use cases have been added.
Data scientists use data much differently than traditional applications. They collect datasets, some of which come from databases, and spend much of their time working with this data to transform it, wrangle it, and better understand it to produce insights and actionable recommendations.
Although data scientists sometimes build programs, using languages such as R and Python, they frequently deploy Jupyter Notebooks to create and share documents that contain code, equations, visualizations, and text. Jupyter notebooks are used for many of the typical tasks of the data scientist such as data cleansing and transformation, machine learning, numerical simulation, statistical modeling, data visualization, and so on. DBAs must understand how this new constituency works differently with data, as well as the different tools being used and the nature of what comprises “production” data and work for a data scientist.
Containerization
The use of containers as a form of operating system virtualization makes it easier to build, test, deploy, and redeploy applications on multiple environments. A single container might be used to run anything from a small microservice or software process to a larger application. The container includes all the necessary executables, code, libraries, and configuration files for a specific use case; however, containers do not contain operating system images.
It is becoming common for database environments to be implemented in containers for development and production work. Containers can require fewer system resources than traditional hardware or virtual machines, can increase portability of an environment, can improve consistency, and can make it easier to support agile development teams. As a result, DBAs must be capable of deploying and working with containers and container management and orchestration tools such as Kubernetes.
DBAs Are Becoming Data Experts
The traditional job of the DBA has been to manage the database implementation and structure but not really to understand and manage the data content. Today, it is becoming much more common for DBAs to be responsible for having an understanding of the actual data. These are duties that used to be performed by data administrators (when those still existed), data architects, and other subject matter experts.
But as cloud deployments and automation remove some of the more routine “drudgery” work of database administration, DBAs are beginning to take on more forward-looking, strategic responsibilities. This can be a struggle for some DBAs because it means adopting a more business-focused role than the traditional technology-focused DBA role. Acquiring a deep understanding of the business and its data requires working closely with business experts and users to ensure that policies are being enacted to appropriately govern data usage, ensure privacy, adhere to industry and governmental regulations, and ensure access.