See this (a bit older, I know) report of
a large Firebird deployment project.
- 24h operation on 6 or 7 days a week
- operation on a fail safe cluster system
- about 25 millions of blobs per year
- about 50 millions of records per year
- 100% logging and archiving of all data
