Have you ever read those “white papers” distributed over the internet on total costs of database administration? One claiming that “Microsoft SQL Server required significantly less effort to install and maintain than Oracle”, another one saying that “DBAs can perform typical administrative functions in 41 percent less time when using Oracle Database compared to Microsoft SQL Server”.
You can just stare in disbelief in their “facts” and “arguments”. The former quote above is from ComputerWorld “The Total Cost of Administration Champ: Microsoft SQL Server 2005 or Oracle Database 10g?”, the latter is from TechRepublic “Oracle Database 11g vs.
Microsoft SQL Server 2008″.
If Oracle is the Lexus of relational databases, is then your Oracle DBA cost as high as the service cost of an expensive car? Should you consider switching to SQL Server? Nobody is saying that IBM’s DB2 is cheap but very often the comparison is between Oracle and SQL Server.
You might have seen Don Burleson’s article Oracle 10g v. SQL Server. I agree with Don that reducing the number of expensive DBAs is hardly an area of opportunity for the IT manager to save money. But when you often show managers figures like the ones below:
Managers will sooner or later start believing some of those “research figures”. Some of them at least. You see, somebody claims that one DBA can administer 9.9 Oracle databases/31.2 Microsoft databases , another one mentions 15.2 Oracle databases/65.4 Microsoft databases. Probably the figures are not far from reality although I have seen that with some automation the numbers can grow much higher (100-150 databases per DBA) although I can’t take seriously studies giving me numbers about users supported by DBA. Sorry, what users? Internet users, database users/schemas? Come on..
It is not up to the number of databases a DBA handles or the gigabytes per DBAs or even the brand. It is up to the skills and experience of person doing the database administrations, it is up to environment and the DBA team handling the work. It is much about the set-up of the databases and their complexity. Of course the automation involved in the DBA work is a significant factor as well. But generalizations like the ones above and below are just mildly insane. If someone tries to convince you that you will save a lot of money by putting your databases in “the Cloud”, just turn and run :-)
About 5 years ago, Forrester did the following forecasts, decide for yourself if predictions were TRUE or FALSE. At least I disagree with the performance tuning curve.
1. DBA to databases ratio:
2. Database administration cost is declining:
3. Database administration challenges:
Database administration has been and will remain the most complex part of IT. The new features added by vendors like Oracle, IBM and Microsoft to the relation software only make administration more difficult, complex and hard to learn and master from scratch.
P.S. There is a book written in 1954 called “How to Lie with Statistics”. You can get it now for 5.99 from amazon.com