The year was 1991 when my mentor told me that if I wanted to be a DBA, I needed to pay attention to details, because that was the number one quality for this role. And little did I know how right he would prove to be.
I started the DBA journey during a time when we installed Oracle from floppy discs and magnetic tapes, with an opportunity to experience Oracle version 2 and the reel-to-reel life for UNIX backups. We won’t even talk about Oracle version 1, because those that know the story, know it.
In those days, you needed to factor in the BPI (bytes per inch) of your storage medium, which would determine the high-capacity data storage. On top of that, you would need to have more than a passing knowledge of the art of compression, to determine how much data you could store on an IBM 3590 data cartridge – it was all about the details, and there were so many.
Back then, those were critical details to attend to, and today we are faced with a new plethora of details to consider, in the ever-changing role of the DBA.
The DBA role expands
Over the years, we’ve seen the appearance and rapid evolution of the Apps DBA (Application DBA), a role responsible for patching, upgrading, and managing the application processes along with the database itself. To this day, I rank this rather high on the complexity scale of the DBA skillset. This is probably due to my experience with the Oracle EBS suite, which required thorough knowledge of the application modules, the conflict resolution manager, and how this impacts the database.
This included little intricacies such as the fact that if data is entered in the Purchasing module, then some of that data may be shared with the Account Payable module, causing database monitoring and the architecture of tablespaces to become an even more detail-oriented task.
We also have the creation of the Performance DBA role, which is always on high demand, so much so that there is always a shortage of professionals of this capacity. This is intensified by the fact that it takes years to accrue the knowledge required to perform this role, as it takes deep knowledge of the database that truly comes only with extensive hands-on experience. This senior level DBA easily gains respect and can not only tune and optimize, but also architect your system for the highest levels of availability and replication. This professional may have the most challenging role among DBAs, constantly facing challenges due to the people and systems around the environment they are tuning. They are usually called in when the problem has escalated or after the lack of tuning of the surrounding legacy system has affected the companies’ main path to profitability.
Part of the work involves sometimes facing developers who may have their guard up with an unwillingness to examine or take criticism for their coding skills. Skills that may be very good when creating programs run in isolation, but that when joined with other programs, may be the causing factor to bring a system to a slow crawl.
The Performance DBA must also understand the details of the operating system tuning and optimization adjustments that are critical to their database processing. This includes details of the kernel, memory, and swap allocations. All these elements need to be considered as a part of their performance plan, which should include the tuning methodology and approach used.
Data recovery and database security
During the most crucial situations, we see the need for the Recovery DBA, an expert who usually works as a part of a team, and specializes in restoring a down production situation with the intent to restore all data. It’s wise to consult this DBA specialist before things become extreme. They can ensure you are archiving with appropriate log rotation and placement of adequate replicas within each data center. The Recovery DBA is also in charge of testing various recovery scenarios and will try to understand the type of transactions to guide them as they understand the RTO (Recovery Time Objective) and RPO (Recovery Point Objective) targets. This should be tracked as a part of their backup and recovery plan.
In recent times with the thirst for more data, the need for the Security DBA has emerged. It’s easy to say that each DBA should know how to enforce security, but having a professional on the team specifically assigned this task can be a true asset. This specialist can create a database map that focuses on access management, complete with privileges for all objects. This requires full understanding of the segregation of duties as it relates to the business need, with control of any auditing measures needed for compliance and reporting. Some large corporate companies may even have a Schema DBA, which is the person who is most closely involved with the development teams. They will create tables and apply all modifications to maintain the schema, while focusing on the integrity by way of verifying entity relationships as they are reviewed by the Data Modeler. Often this can be found as a role within the Security DBA’s responsibility.
Enters the cloud
Next on the list we have the Cloud DBA. At the beginning of the cloud era, companies believed they didn’t need a DBA, but that has been proven incorrect, which led to the creation of this specialization of the DBA role. In the land of IaaS, the level of control is much like that of an on-prem environment, allowing the user full installation, creation, and management of the database. Yet, DBaaS will take away that control and simplify the database deployment. The challenge will be catering for the differences presented by the various cloud vendors, so the DBA must learn the navigation and nuances of each, as well as the tooling they provide that will make the database interaction easier. Most DBAs specialize in one database, but the Cloud DBA must be versatile and may need to learn the basics of MySQL, MongoDB, and PostgreSQL, as they are available in the major clouds. The cloud has planted new roots on the farm.
- Failover - Automatic switchover from primary to a replica
- Availability - Access to the system, Uptime, MTTR (mean time to recovery), MTBF (mean time between failure)
- Redundancy - Zero data loss, multiple replica options
- Monitoring - View of daily activity, system health, and resource impact
Learning about compute resources to assist with the farm can help with optimization and recovery strategy making use of the proper availability zones.
Vector databases create demand for yet a new type of DBA
We once thought the cloud would ride us into the future, but as we enter a new dimension in the world of data and the speed of technology, we are presented with the Vector DBA. This is the DBA that must take us into the future, where attention to detail will be needed even more than so far, due to the rapid changes and massive amounts of data being consumed and stored in the database. Not to mention that the object maintenance is at a new level, with data being stored in multiple formats within a single environment. Today, we live our lives on timelines from our favorite social media apps reminding us of the history while we post for the future.
The Vector DBA is much similar to the Apps DBA, as both need to understand the applications being used. From BI reporting to the increasing AI explosion, there are many reasons that require this DBA to obtain skills on LLMs (Large Language Modules) and familiarize themselves with anything they can learn on chatbots and ChatGPT. The vector database itself is packed with information that calls for detailed analysis, as it allows for the storing of images, audio, and text in vector embeddings which provides quick retrieval. Samples of vector data can be found where you find the integration of tools like PostGIS, QGIS, or ArcGIS in use, with a collection of geometric data that includes lines, point-to-point references, and polygons.
The vector database has quickly risen to the top of the charts, and so has the need for Vector DBAs, mainly due to the business need for consuming data quickly in various formats. Data scientists kicked off the start of the era, and now generative AI making use of vector databases to analyze data will take us into the future. Vector DBA may even be called Future DBA.
What lies ahead
With the honor of having DBA in your job title comes the responsibility to own the role and what comes with it, including the challenge of expanding your knowledge as technology changes. Yes, the traditional tasks of tuning, optimization, and integrity maintenance will always be a part of the role, but a true DBA needs to expand their skillset so that we only have one type - the Modern-day DBA, no matter what day it is.