Honestly curious: Is there any objective reason for going with Oracle DB over PostgreSQL? Putting emotions aside :).
Or is it that Oracle runs on the fumes of legacy lock ins?
Or is it that Oracle runs on the fumes of legacy lock ins?
The answer is still: No, not really.
The price tag of Oracles database license is simply to high, compared to any benefit you might get. It's much better to try to redesign your application to work around any feature you might be missing in Postgresql, or MariaDB.
For many organisations, the licensing costs of an Oracle or Microsoft RDBMS are trivial compared to the costs of the projects they're supporting, and it's easy to find where the value might be. Everything that's difficult about running an RDBMS at scale (clustering, failover/replication, statistics collection, plan management, monitoring, online index rebuilding, table partitioning...) is easier with the features and management tools those companies provide. Add actually decent support to that, and you've got a very clear value proposition.
A Postgres instance will require a lot more hands on management, and a deeper understanding of how the database engine works. At smaller scale, these factors are much less likely to cause problems, and many organisations are unlikely to ever reach actually large scales (many TB instance, running hundreds or thousands of transactions per second). But if you want to start scaling an RDBMS application up, you're going to have to confront those costs in one way or another. After a point, licensing expensive features can start to look better than deploying more specialised labor.
I remember some colleagues rolling out a project to thousands of postal service branch offices. Oracle was very happy to provide 24/7 emergency DBA services in case anything went wrong with any of the instances (local to the branch offices).
Just an additional guarantee of "if one of them behaves weirdly, we will be right on it" (as a service).
Unless you get a Postgres-based consultancy to do the same for you (and ofc they're not first party), that support is just unbeatable (in terms of what you're selling to your customer, the post offices).
But it also seems to be uncontroversial that no good could plausibly come out of attempting to sue Oracle.
You have to live with them anyway?
You obviously don't try to sue them, their company is 90% legal staff [0]. (Sorry I can't find the original source)
[0] https://am21.mediaite.com/tms/cnt/uploads/2011/06/org-charts...
The support is pretty good although it's usually only needed because the products themselves are filled with obscure bugs, edge cases and combinations of features which don't work well together.
Mainframe is the CA of CAP.
Personally I chose SQLite sharded on S3 and dynamodb.
Why would they use AWS? They have their own cloud: https://cloud.ibm.com/catalog#services
This sounds interesting, would you mind sharing details?
But you can do a lot with Oracle XE and Apex (which are free). and you can link them up with Postgresql through HA or FDW.
Also some customers feel more safe when they know they've paid $10 000 or probably more for an oracle license.
Source: I once suggested to my boss that maybe we could save around $10 000 pr customer site and also save ourselves tons of work by using Postgres instead.
Firebird db is the only other database I know of that has package functionality. For a portable database, that is pretty impressive...
A TIMESTAMP WITH TIME ZONE data type that makes sense and is usable.
I'm fed up with support (it takes so much work to open a ticket) and with the documentation (excessively verbose and full of marketing).