I'm guessing nobody chooses to work with Oracle anymore for reasons or in situations that we would consider reasonable. It's probably either governments contracts, with or without corruption, companies already locked in, contracts made by executives that don't really understand technology, that sort of thing.
I worked as a contractor for the Wisconsin state government and they had hundreds of Oracle databases that they were consolidating on the Oracle EXADATA11 servers. Insane having hardware that can only run Oracle but the Oracle DBA said that the Exadata was dozens of times faster than Oracle on VMware VMs.
Lies. Fucking lies. We were a three environment shop until we moved to Exa and the compute/$ ratio is so bad that we had to cut it down to two.
But we're talking about Oracle here so that's par for the course.
I didn't make any claims about performance per $, just relative performance compared to VMs. I hate Oracle as much as anyone but the EXADATA is impressive hardware. It has lots of RAM and Infiniband networking. It can push query predicates to the storage controllers to reduce the data that had to be transferred.
It is impressive. But for the same cost you can get vastly better performance with Postgres and bigger hardware.
It does come with internal redundancy, but do you need that? Also the cluster nature of it can come with some surprises as compared to a single database.
Actually, it is mostly companies who are too reluctant to change. If it works, keep it as is, even if better technologies are the norm nowadays. Maybe this will help them move away from this obsolete Larry Ellison crapshot
If it works, keep it as is
That's a good principle though. It doesn't make the initial choice good today or even back then. But change is always a risk that may not be worth it, cause you have to make sure that the inevitable semi-chaos coming with it is at all times lower than what you have. And analyzing that may be hard.
Maybe this will help them move away from this obsolete Larry Ellison crapshot
This creates positive incentives, so yes.
Iow, everything probably goes as it should, really.
I somewhat agree. I think for tangible things (cars), you don't need to reinvent the wheel. But, tech moves fast. If a superior tech (for instance, more secure) is available but requires some discomfort (moving things around), then it is worth it to avoid this type of crap
Seemingly, most companies have a terrible ability to judge if a technology is superior, hell most of the time they lack the ability to judge if a technology is massively inferior. Companies may understand what they do but they commonly have no understanding of what they do in relation to the abstraction layers between them and said technology.
Often this technology has been in place for some time and the original creators are long gone for one reason or another. To migrate away from this system the business will need to spend a significant amount on contractors/consultants to understand both the system they have well enough, and the system they are moving to. It can be a huge expense and companies are very willing to push that off into the future.