Let's imagine a situation that needs to start a development process because of compilance.
The dml SQL that solves the problem can run within 50 minutes in the small development database.
This would be 50 hours in the full size testing database. Moreover, it has to be run 5 times
because the development is not perfect at first.
So we can avoid to pay a 50M HUF fine because we can fulfill the deadline.
It is difficult to express in monetary terms, because
it is many variables problem:
the savings rate can be very different by companies and by time moments
(what does the fact mean for example to operate 5 test environments instead of 1
of a production environment)
the time is more important than money:
depend on the skill of colleagues and the maturity of the organization
the saved time can be "convertable" to more time+money
The following should be considered.
These appear about the testing databases:
1: should be produced
2: should be operated by infra and OS side
3: should be operated by DBA side
4: should be operated by application side
5: should work with (end users and applications)
The value of the saving can be characterized by (using PrDaRe):
-2: have to work a lot on it
-1: have to work on it
0: not have to work on it but not save anything
1: saving
2: lots of saving
So the figure below can show our opinion and experience about the small databases
(producing and using):
-2
-1
0
1
2
1
x
A little bit more manuality but ready soon
2
x
Less CPU, RAM, DISK, NETWORK BW needed!
3
x
Database administration is faster
4
x
Application administration tasks can be done faster
5
x
By orders of magnitude faster little databases!
We can multiply the above if we would like to produce not just 1 small database.
There are companies doing this 3 time and there are other companies reducing 300 databases,
this can be very different by companies. We apply no limit on the number of reducable databases.
The processes above are done almost with starting these so the safe saving will come in the short term.
The extent of this can be different by discussed above.