MongoDB vs. ScyllaDB
bbigford last edited by
Let me tell you, we had been loving MongoDB, but this makes me feel SO much better deciding to move to ScyllaDB recently.
I'll have to take a look into ScyllaDB. Obviously, it's not a 1-to-1 replacement for MongoDB, and maybe that's a great thing. Got any pros/cons while I'm doing some research on it?
tonyshowoff last edited by tonyshowoff
Aside from MongoDB's licensing, it also has a habit of losing data due to the fact it doesn't guarantee writes, even though ScyllaDB is not ACID compliant either it doesn't suffer from this disappearing data problem to the same extent it seems. This on top of retrieving data, you can get unexpected or not-expected results such as requesting the same information twice but the second time being older data than the first one. This has been a design problem for years.
It also is only single master so your scaling is limited to reads rather than writes, you'll have a write bottleneck always. Further due to the single master problem, if you want constant uptime then MongoDB is not what you're looking for because depending on data and network size, it can take up to a minute before one of the slaves becomes the new master. Plus you have to write in extra code in your application to automatically understand this or deal with another external MongoDB service.
If you know SQL at all then you'll have less to learn since ScyllaDB has SQL access where as MongoDB uses JSON(B) which can get clunky sometimes, though if you're writing you're app in node.js it's pretty straight forward.
scottalanmiller last edited by
ScyllaDB participates here. Thats huge, too.
Everything depends on your specific use-case, MongoDB has it's upsides and Scylla is, well, absolutely awesome :smiling_face_with_halo:
If you describe what you are trying to achieve, you might get some more constructive input. As for Scylla specifically, please feel free to join the Slack channel and ML if you have questions
jmoore last edited by