1

Seems like the use cases I’ve heard about for Cassandra have a lot to do with handling a massive volume of data or writes.

Is there a use case for Cassandra that doesn’t relate to how massively it scales? Like certain data structures or queries that are better suited for it (similar to how RDMSs are good for handling relational data with consistency)

Improve this question

1 Answer 1

Reset to default
3

From a pure data structure perspective, Apache Cassandra is a good match for requirements with:

  • Key-based read patterns.
  • Data that will be written once and never deleted or updated.
  • Log-based data sets that may grow over time.
  • Read patterns of many, small queries.
  • Time series.

The underlying log-based storage engine will definitely "play nice" with those types of use cases.

Improve this answer
8
  • This might be a silly question, but what’s the point of writing data if it’s never read?
    – Greg Thomas
    Commented Jul 19, 2022 at 13:48
  • @GregThomas Nah, not silly. Happens all the time in clusters that back logging or observability platforms. The idea is to write generously; read judiciously. Something like 70%-80% of all data written is never read.
    – Aaron
    Commented Jul 19, 2022 at 13:52
  • I mean, you'd rather have it in the event you needed it, right?
    – Aaron
    Commented Jul 19, 2022 at 13:53
  • 1
    @GregThomas Aaron's answer here is good (and succinct). The only addendum I'll add to what he said is "fast" is subjective. Most modern database systems write "fast-enough" for most use cases and people. Rarely do people actually hit the maximum throughput (which can be in the millions of rows+ per second written) of the database system they're using. Rather, oftentimes the bottleneck one hits is either from the hardware they provisioned or the way they architected their database. Most modern databases are within the same ballpark of read and write speeds, generally speaking.
    – J.D.
    Commented Jul 20, 2022 at 3:23
  • 1
    Thanks, nah I think it's too generic info (given the original post was Cassandra specific) to stand on it's own legs. But feel free to incorporate any of it in your answer if you'd like. I just like to throw my 2 cents in whenever I see an opportunity to make it clear that most database systems are on par with each other from a performance perspective (even when it comes to big data) despite what the marketing says or other misconceptions going around the internet are. What I liked about your answer is instead of talking speed, you gave clear objective use cases that Cassandra was built for.
    – J.D.
    Commented Jul 20, 2022 at 14:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.