I'm looking for a technology very similar in function to count-db.

Use case

You want to use count-db if you need to write and read billions of counts very efficiëntly from a Java program. Use cases are logging large amounts of user data, counting n-grams for language models or building an index to search texts. You don't want to use count-db if you require transactions or my-sql style querying.

I need to be able to do these things with it:

  1. Store data whose logical structure is similar to the Map<Integer,Integer> data type
    • At the minimum, 200 entries in size each
  2. Atomically increment a subset of the Map's values by 1 given a list of ~15 Integer keys
    • Assume this is done 10,000/s on average, but the tech recommendation should have the
      capability of scaling out
  3. Get an entire Map's values based on an Integer key
    • So structurally, the data store should be organized like Map<Integer,Map<Integer,Integer>>
    • Assume getting an entire Map is done 1,000/s on average

count-db captures the spirit of what I'm trying to do -- strictly counting and retrieving integers. The reason I'm not using it is because it doesn't seem to offer storing Map data types, so getting an entire Map would require doing 200 key-value pair lookups for something that should logically require 1. And also, the only documentation there is about it is its homepage on github -- a single, short Java example. There's not even comments in the source code.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.