Facts / Statistics about DATABASE DOWNTIME

Here are the questions I need answered:

  Need to know frequences and downtime durations for:

  1) Overall database downtime numbers 

  2) Database downtime numbers across different industries (US/world).

  3) Breakdown by database vendor (Oracle vs. DB2, etc)

  4) Downtime cause (operatior error caused vs. application error caused downtime).

I have no hard numbers, but I can tell you from several years'
experience that Oracle is rock-solid if the underlying environment is
stable, and it's dead-easy to cluster it on commodity hardware
guaranteeing nearly 100% uptime in the face of hardware failures.

Also, in my experience, 90%+ of downtime is due to user error.  "Um. 
I think I accidentally dropped your table."  "Oops.  I accidentally
kicked out the power cord."...  that sort of thing.

Thanks for the comment - having similar experiences, what I really
need is a credible references and numbers across industries, platforms
and database vendors.

I looked around on this question myself, spent quite a bit of time to
see if anyone did 'downtime' stats, but as mwallace pointed out,
downtime is really a reflection on the DBA and not the Database
itself. Most downtime is OS and the User anyway (80%) and the rest is
messing up with an 'oops'.

You could look at stress testing, but this doesn't reflect on the down
time rate at all, because if your database is going down due to
stress, then this again is a reflection on your Database Administrator
not understanding the limits of his/her database, not the Database
itself.

With that, there are no reliable stats to work with. To do the test
you would have to have Admin's versed in the databases being tested
and rule out 'downtime ' caused by situations such as OS, hardware and
power outages. In this setup, with the major databases all of them
would have 100% uptime, over the course of 30 days, or even a year.
So, what's the point really?

webadept-ga