I was recently thinking about how SQL Server has changed over the past 20 years, and how the hot, in-demand skills have evolved as well. I started working with SQL Server 6.5 in 1998. I don’t remember much about working with that version, but it must have been a painful experience. Databases were quite small, and DBAs were concerned about security and backups.
When SQL Server 7.0 was released, Microsoft promoted it by saying that you no longer needed DBAs to run SQL Server. This was far from the truth, however, but this release did introduce tools we rely on today like color-coded syntax and graphical execution plans. This was the first “modern” version of SQL Server supporting up to 8 processors and 64-bit operating systems. DTS, the precursor to SSIS, and OLAP Services, the precursor to SSAS, were both new in 7.0. This was also the beginning of 8K pages, parallelism, and cost-based locking.
Many shops skipped over 7.0 and upgraded to SQL Server 2000 when it was available. My favorite new feature was multi-instance support, but there were many improvements in features and performance. Failover Clustering was the hot skill of the day.
SQL Server 2005 was a game changer, and the first “enterprise-ready” version of SQL Server, in my opinion. There were lots of changes to T-SQL which, unfortunately, delayed many upgrades from 2000 for years. There was more focus on business intelligence with the SSIS, SSAS, and SSRS components. DBAs could now understand more about what was going on “under the hood” with the new system views and DMVs (dynamic management views), and understanding internals was an important skill. Many shops also took advantage of the new Database Mirroring feature for high availability and disaster recovery.
Encryption, compression, and Extended Events were the exciting features of 2008. It was around that time that some shops started thinking about virtualizing SQL Server. There was resistance at first, but now it’s unusual to see SQL Server running on dedicated hardware. Releases of on-premises SQL Server have appeared at least every two years after 2008. There have been a dizzying number of features added or improved since then. The 2008 R2 version, released in 2010, is still used in many shops, but the most compelling feature was probably PowerPivot.
New T-SQL features, Column Store indexes, and Always On Availability Groups were exciting features of SQL Server 2012, while In-Memory OLTP was the hottest feature of 2014. SQL Server 2016 supported the R language giving the ability to run advanced analytics applications without importing data to the client. Many database professionals were starting learning R and data science. DBAs also started thinking about data in the cloud as this version supported a feature called Stretch Databases that allows storing a portion of a table’s rows in Azure SQL Database. Query Store, JSON support, PolyBase, Mobile Reports and more proved that 2016 had a lot to offer, and database professionals had a lot to learn.
SQL Server 2017 is the latest and greatest version available with 2019 on the horizon. Query tuning is evolving as the first adaptive query processing features were released in 2017 with even more coming in 2019. Today’s DBA needs to learn about containers and (gasp!) SQL Server on Linux. Understanding cloud computing is even more critical for DBAs as many companies take advantage of Azure and AWS.
SQL Server has changed quite a bit over the years, and database professionals have changed as well.