Engineer | Lover of dogs, books, and learning | Dos XX can be most interesting man in the world, I'm happy being the luckiest. | I write about what I learn @ code-comments.com
Engineer | Lover of dogs, books, and learning | Dos XX can be most interesting man in the world, I'm happy being the luckiest. | I write about what I learn @ code-comments.com
Engineer | Lover of dogs, books, and learning | Dos XX can be most interesting man in the world, I'm happy being the luckiest. | I write about what I learn @ code-comments.com
Got it! For what it's worth, I was recently working on projects with ~10m records - I went with Mongo, some peers went with Cassandra. We had comparable query times (post indexing), but the set up of Cassandra was apparently a pain whereas Mongo is a breeze.
According to DZone, NoSQL is typically local:
Source
That said, this was the first time I had to look into it.
Pity. I remember looking into Couchbase and it supported global secondary indexes. Thought that MongoDB has caught up.
Can I ask, how much data are we talking about that you've got a distributed setup and need the global indices?
I don't really have one. I did quite a lot of research about Cassandra and wanted to have data there for a project but it never worked out.
Project was supposed to be a rewrite of legacy app that had to run in multiple regions and distributed database felt like a natural choice.
Cassandra, like MongoDB had only local Indexes at that time.
Got it! For what it's worth, I was recently working on projects with ~10m records - I went with Mongo, some peers went with Cassandra. We had comparable query times (post indexing), but the set up of Cassandra was apparently a pain whereas Mongo is a breeze.
Gee! Cassandra seems like a massive overkill for 10M records. Even MongoDB does, actually.
Also ye, Cassandra requires way more thinking in terms of data modeling and setting up infrastructure, but if done correctly - pays off.