Trends in database administration

Trend in the database industry is rapid DBMS versioning. It seems like we just start to get a handle on the latest and greatest version of our favorite DBMS and then -- wham! -- the vendor releases a new version. I mean, people are still digesting Oracle9i and then Oracle unleashes 10g on the world. And how about IBM? On the mainframe side of the world most folks are still using DB2 V7, even though V8 has been available since March 2004. Who wants to be the first to encounter all those bugs that you just know are in that new version? So, most people wait until the bleeding edge implementers have slogged through the upgrade until starting to plan their upgrade path. That means we have two or three versions that are considered "current" (at least by the users) in the field. In such situations it becomes very difficult to keep up with which features are available and which are not.

Which leads me to the next big trend: complexity. Our database systems are getting more and more complex. This trend is driven by all of the new features and functionality in the new versions, along with the fact that most organizations have heterogeneous database implementations. Let's tackle feature bloat first. Today's modern DBMS offers functionality never dreamed of by the database pioneers of the 1970s. The DBMS is sucking features into it that previously required you to purchase additional software. For example, today's DBMS products usually offer analytical/OLAP and ETL features. In the 1990s these were common products, but today they are common features. So the DBMS becomes more complex. Additionally, we add new data types (BLOBs, CLOBs and so on), code (UDFs, triggers, stored procedures) and XML support. These make the DBMS of today more functional and, yes, much more complex.

Heterogeneity is an issue, too. Most organizations have multiple DBMS products installed and managing corporate data. Rare is the shop that only has to worry about Oracle or DB2 or SQL Server. Instead, they have all three and maybe MySQL and IDMS, too. Now how do you ensure data integrity when the same data is spread across each of these products? And maybe in Access and Excel, too?

Another consistent trend is the on-going Web enablement of our data. And the DBMS is changing to better support this trend. So, too, are DBAs changing. The biggest challenge for Web-enabled data is availability. If data is not available, the e-business is not functioning. This will impact sales, profitability and ultimately stock price and valuation. Being prepared to eliminate and reduce planned and unplanned outages is the biggest job of the eDBA. This requires more automation and better tools, as well as forethought, planning and vigilance. A modern DBMS allows more changes to be made without requiring the database to be brought down. Furthermore, the modern Web-enabled DBMS supports Java, .NET, XML and other Web technologies. It is better integrated to Web servers and application servers. And it is supported by tools that enhance availability by enabling on-the-go maintenance tasks.

Another trend is industry consolidation. This one has started to hit the DBMS market with IBM acquiring Informix (a few years ago) and Oracle acquiring PeopleSoft just a couple months ago. This trend is actually industry-wide and is likely to continue. The smaller DBMS vendors may get swallowed up by the larger ones. And as when DBMS vendors acquire application vendors, we all get to sit and wait to see how well competing DBMS products are supported. I mean, does anyone really believe that Oracle will still be supporting DB2 and SQL Server in the PeopleSoft/Oracle applications five years from now? Maybe, but, then again, maybe not.

Of course, many other trends exist in the industry that impact the database market, including open source, data growth, ERP, CRM, data warehousing and business intelligence, outsourcing and so on. The only way to stay on top of things is to allot time to study, learn and research. Good luck keeping your databases up and running!