Today we have a guest editorial from Kathi Kellenberger as Steve is on vacation.
When I first began working with SQL Server twenty years ago, I had one book that I kept next to my computer. It was about four inches thick and contained just about everything I needed to know about SQL Server. Back then, I thought a 500 MB database was large, but I was working with relatively small organizations at the time, so my experience was limited. My primary job back then was as a developer, and SQL Server was really just the place where data was stored as far as I was concerned.
After becoming a DBA in 2002, the hot skill at the time was failover clustering. It was not easy to set up, but once in place and working, it was a thing of beauty. Today, the more flexible Always On Availability Groups are often used instead for high availability (and disaster recovery), but many folks still use failover clustering and it’s easier to configure than it was back then. Security, maintenance, patching, scripting, and testing restores were also important day-to-day skills for DBAs, and still are today.
In 2005, the data management views and objects (DMVs) were added which gave DBAs a way to “look under the hood” and better manage SQL Server. It seems to me at least, that was when interest in internals by many folks began. More DMVs are added with each version of SQL Server, and in 2008, Extended Events were introduced. Many of us are still being dragged “kicking and screaming” to Extended Events even though the feature is now ten years old.
Over time, SQL Server became more than just an RDMS platform as business intelligence components were added. Ten years ago, SSRS, SSIS, and SSAS were sought-after skills as more organizations built data warehouses for analytics and reporting. Over time, even more components have been added to SQL Server such as Master Data Services and Polybase. Today, you can run R and Python in SQL Server, and many database professionals are learning about machine learning and data science.
Even from a hardware perspective, things have changed a lot. At one time, virtualizing SQL Server was just not done – now it’s unusual to see a SQL Server on physical hardware. Direct attached storage was replaced by SANs which, in turn, have become faster as SAN storage has moved to SSDs. Processors are faster and contain multiple cores. (This seemed great until Microsoft changed the licensing model of SQL Server from per processor to per core in 2012.) Virtualization and SAN storage, while making SQL Server easier for the organization, adds complexity for the DBA. Just when we thought we had things figured out, now we can run SQL Server in Linux and Docker containers.
Column Store indexes, In-Memory OLTP, JSON support… The list of new things to learn is endless. Instead of a four-inch thick book next to my computer, I need an entire wall of books. Oh wait, that is one more thing that has changed. Now I read books on my phone.
