January 30, 2026 at 12:00 am
Comments posted to this topic are about the item There Are a Lot of Databases
January 30, 2026 at 9:15 am
We do use Amazon Athena, which is Presto with a Hive metastore. Trinio is a fork of Presto.
Presto is a distributed SQL processing engine that lets you query many data sources as if they were one. It's an interesting idea, but can be expensive to run. I notice that ETL and BI use are considered anti-patterns.
Ab Initio had a concept they called SQL to graph. I believe that one was a case of Ab Initio using SQL to generate its graphs (ETL flow) and retrieve data. As Ab Initio can read files formats you've never heard of as well as the ones you have this was an interesting concept.
One of my friends ran a POC for VoltDB and Hazelcast while working for a company that needed near real-time turnaround of automotive telemetry. The problem they found was that, functionally, the systems did what they claimed. Operationally, they need a lot of DBA type intervention to keep them running. Especially given the SLAs that near real-time operation demands.
One of my ex-colleagues is very interested in typedb.com, which appears to be a graph database. There is lots out there. Too many toys, not enough time to play with them.
January 30, 2026 at 2:24 pm
I'd be curious how many of these obscure database systems are being chosen by developers instead of DBAs. I get the impression that these would be chosen by devs who want a specific function and are either unaware the major RDMS's can handle that function or are simply trying to get around the need for a DBA to manage said database.
Be still, and know that I am God - Psalm 46:10
January 30, 2026 at 2:43 pm
I'd be curious how many of these obscure database systems are being chosen by developers instead of DBAs. I get the impression that these would be chosen by devs who want a specific function and are either unaware the major RDMS's can handle that function or are simply trying to get around the need for a DBA to manage said database.
I've seen 3 edge case DB platforms in the wild. Two were chosen by CTOs, the 3rd was a ShadowIT thing.
The common themes were
Two of the DB platforms, on paper, were brilliant. I've got a soft spot for one of them, and it still exists.
The other desperately needed a sugar daddy investor and suffered from an absence of basic documentation; I can only find references to it up to 2009. It was an idea ahead of its time, but an evolutionary dead end.
January 30, 2026 at 7:24 pm
I think devs, or someone with previous experiences, make a lot of choices. They think a thing worked in the past, and it's a data store, so it should work here.
Might be true, might not.
I think some of the areas where a lot of systems are immature in is with HA and perf metrics. Often they don't support things well or they don't consider this when building. Even PostgreSQL, for as much work as has been done, is far behind MSSQL in a lot of areas. It's catching up, but until recently, I felt it was really missing some core things we depend on in the HA area, or in terms of details on what's happening on the server.
Viewing 5 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply