This was a cool talk. Well structured and very clear, and did a good job of contrasting how different DB types solve different problems. Nice stuff, Steinkamp
Before relational databases, we worked on network and hierarchical databases. The indexes were designed into the database. And by that, if we wanted to add or remove an index, we would have to unload and reload the database until 3rd party tools filled that gap. The hierarchical database tables could link to only one parent where network databases could link to multiple parents.I remember Object Oriented databases appearing in the 80s - a little too early? Great talk! I learned a lot.
For over 99% of companies PostgreSQL and the extensions available are FOSS and can do anything they need. Relational, document, vector, timeseries, graph, KV, etc. Postgres has you covered. Want row store, compressed row store, column format storage or even parquet storage? Postgres has that covered. Want scaleout? Postgres has that covered too. It’s all free and you can deploy it anywhere. If you want to start with something start there. If you then realise that you need something that is a little bit better because it is very specialised then change. However, I cannot think of anything where there is an order of magnitude difference between Postgres + extensions and more specialised publically available DB that more than at most 1% of companies might need.
Great talk but she breathes out from her nose at the end of a phrase very often, and she needs a water bottle that stays open so she doesn’t have to constantly twist it open and closed.
This was a cool talk. Well structured and very clear, and did a good job of contrasting how different DB types solve different problems. Nice stuff, Steinkamp
Loved the talk
Before relational databases, we worked on network and hierarchical databases. The indexes were designed into the database. And by that, if we wanted to add or remove an index, we would have to unload and reload the database until 3rd party tools filled that gap. The hierarchical database tables could link to only one parent where network databases could link to multiple parents.I remember Object Oriented databases appearing in the 80s - a little too early? Great talk! I learned a lot.
For over 99% of companies PostgreSQL and the extensions available are FOSS and can do anything they need. Relational, document, vector, timeseries, graph, KV, etc. Postgres has you covered. Want row store, compressed row store, column format storage or even parquet storage? Postgres has that covered. Want scaleout? Postgres has that covered too. It’s all free and you can deploy it anywhere. If you want to start with something start there. If you then realise that you need something that is a little bit better because it is very specialised then change. However, I cannot think of anything where there is an order of magnitude difference between Postgres + extensions and more specialised publically available DB that more than at most 1% of companies might need.
Interesting. Thanks.
Great talk, interesting and informative. Thanks for that.
(Btw, you are still very young!)
6:49😂
09:55 😂😂
Gee, what an unfortunate timing.
unfortunate in what way?
@@xybersurfer InfluxData nuked a few of its datacenters and clients did not know about it a few days ago.
Great talk but she breathes out from her nose at the end of a phrase very often, and she needs a water bottle that stays open so she doesn’t have to constantly twist it open and closed.