I’ve been working with SQL Server since 1996. As it happened, I started with SQL Server 6.0 on a month-long engagement for a point of sale project. Immediately following, I worked with one of the very early SQL Server versions – 4.2 — for a salt company’s inventory reporting system. That gig lasted only a few weeks but I learned to code stored procedures and got a good overall foundation in client/server architecture. Later that year I was invited by some Microsoft folks to the release party for the Back Office Suite in Las Vegas. So began a love affair that’s lasted nearly 20 years.
Prior to SQL Server, I’d been working on the Unix side of things: Postgres, Progress, Ingres and lots of xBase code (dBase, FoxBase, FoxPro, Clipper et al). Just like everyone else I had to make a choice. The choice was either the upcoming SQL Server product and client/server or stay with what I knew.
At that time I believed the sun was setting on many Unix products, as was the old xBase languages. I figured the future belonged to distributed computing and good systems would be based on server-based data systems. I believed then (as I do now) that the future belongs to us (and Futureman!).
Back then I worked for small to medium sized businesses that couldn’t afford Oracle, Sun, IBM or anything in the bigger sense of data. To be sure, some did have multiple platforms and on occasion I’d get a crack at wring PL/SQL and DB/2 and seeing how the other side lived — and I’m very thankful for those opportunities.
In 2011 and 2012 I focused on Hadoop and how to extend SQL Server using Cloudera’s implementation of the popular Open Source platform. The problem I was solving was finding the right way to create a data lake and leverage it from SQL Server. Mainly this was to get around the need for Federation for scale out deployments on very large systems. Federation definitely is still an answer but it’s woefully complex, expensive and not always a well performing approach.
What I love about Hadoop is the built-in toughness on cheap hardware. For the most part it’s easy to manage and in many circumstances – blindingly fast. It’s also a small return to my Unix roots which is fun! I’ve also explored HBase, Cassandra and CouchDB as well.
In the coming weeks I’m going dig into cool, new features of SQL Server 2014 such as:
- In Memory OLTP
- In Memory Data Warehouse
- Updateable Column Store Indexes
- AlwaysOn integration with Azure
At some point this year I’d also like to continue my SQL Server + Hadoop proposition.
Have a wonderful new year!