A deep dive into non-relational core databases, rather than just using them as caches.
When your company is still small/medium it’s not always easy to see future use cases. We mistakenly chose MongoDB once only to realize few years later that it doesn’t fit well.
What made it a bad fit?
Our data ended up being very structured, so we needed cascading updates and deletes, which is not straightforward with Mongo.
Structure and cardinality (important if you have tennants) are the poison pill for any NoSQL db. Can be solved with strong design constraints.
Sounds like you had a strong use case for a traditional relational system. Lots of on-prod-DB analytics involved?
When your company is still small/medium it’s not always easy to see future use cases. We mistakenly chose MongoDB once only to realize few years later that it doesn’t fit well.
What made it a bad fit?
Our data ended up being very structured, so we needed cascading updates and deletes, which is not straightforward with Mongo.
Structure and cardinality (important if you have tennants) are the poison pill for any NoSQL db. Can be solved with strong design constraints.
Sounds like you had a strong use case for a traditional relational system. Lots of on-prod-DB analytics involved?