Re:How to select the most appropriate DB for your project?
by uniqueuid on Tuesday November 30, 2021 @04:27AM
(#29390415)
Personal experience tells me to go by the following rules
- Pick what you know
- Pick what's boring (i.e. settled technology)
- Pick what others know and understand
- Don't be afraid to change schemata or DBs - in fact, get comfortable with the idea from the beginning. This prevents mental lock-in
- You don't need a specialized database (i.e. graph, time-series). If you need one, you already know it.
- Always, always have backups and test them.
And some more subjective takes:
- Sharding is easier than distributed DBs. If you can design your data flow to be largely unidirectional and compartmentalized (i.e. user profiles etc.), you benefit massively
- Multi-Master replication is often a huge cost in terms of complexity and risk. Avoid it wherever you can. If you can't, get someone who has battle scars to help you.
- Pick what you know
- Pick what's boring (i.e. settled technology)
- Pick what others know and understand
- Don't be afraid to change schemata or DBs - in fact, get comfortable with the idea from the beginning. This prevents mental lock-in
- You don't need a specialized database (i.e. graph, time-series). If you need one, you already know it.
- Always, always have backups and test them.
And some more subjective takes:
- Sharding is easier than distributed DBs. If you can design your data flow to be largely unidirectional and compartmentalized (i.e. user profiles etc.), you benefit massively
- Multi-Master replication is often a huge cost in terms of complexity and risk. Avoid it wherever you can. If you can't, get someone who has battle scars to help you.