Can someone redpill me on NoSQL? Why is it so "powerfull"?

>redpill
Fuck off back to where you came from

>>59881434
I thought the analogy of the pills, as the blue being the current status quo and the red leading to a new way of understanding was most adequate for this case.

Didn't mean to disrespect senpai.

>>59881407
Depends on what your data is and how you store it.

A relational database will prove advantageous if you rely on the database to do the work (lots of joins, structured data). If you have data that doesn't fit well in that model, or you are primarily doing read calls, then a more simplistic key->values storage might be preferred in which case you would want to use nosql.

>>59881434
ooooo, someone got triggered hard.

>>59881545
I keep reading that NoSQL is the next evolution of DBs, not an alternative to SQL.

>>59881808
Personally, both products wouldn't exist if one was superior. AWS (and GCE, and Azure) all know that some workloads work best with certain storage/DB mechanisms.

Nothing is a clear winner in computer science ever.

>>59881808
That's probably because those people don't know what they're talking about and/or jump on every single new technology like it's going to save us.

NoSQL is good for its problem domain, but there are still areas where it's not going to come close to an RDBMS.

>>59881545
This is true.
Basically you de-normalize the database (imagine no lookup tables) and take up more storage (because it's cheap), in exchange, it can easily scale up to a huge size and can be very performant.

Simple system for simple applications.

For dynamodb in particular, as with most cloud things, it boils down to a tool you can throw a lot of money at to make a problem go away for a while.

>>59881808
It goes in cycles.

But the more complex your application gets, the more reverse-lookup tables or odd semantics you add, and eventually you end up with a hack of a relational system on top of the NoSQL system.

The Dynamo paper is nice tho.