Trying out new trie-based memory in memory storage engine

Apologies if this is in the wrong topic.

I’m testing out FoundationDB to get an estimate of how expensive it would be for us to run, and it seems like the trie based engine would be a huge win for our use case (~80% of our data is in key prefixes shared by 100s to millions of keys). Is there a way to try it out? Looking through the source code, it looks like I could use ssd-redwood-experimental to test out the Redwood backend, but I don’t see an equivalent for the new in memory store.

The radix tree engine isn’t merged to master yet, but the configuration option is memory-radixtree and the code is in @mengranewo’s branch here:

And here is the pull request:

1 Like