Walrus: A High Performance Storage Engine built from first principles
A little bit of intro, Walrus is an embedded in-process storage engine built from first principles and can be used as a building block to build these things right out of the box like Timeseries Event Logs, Database WAL, Message Queue, mini Kafka, key value stores, task queues, and much more..
the recent release outperforms single node apache kafka and rocksdb at the workloads of their choice (benchmarks in the repo)
repo: https://github.com/nubskr/walrus
If you're interested in learning about walrus's internals, these two release posts will give you all you need:
1. v0.1.0 release post:https://nubskr.com/2025/10/06/walrus (it was just supposed to be a write ahead log in the beginning..) 2. v0.2.0 release post: https://nubskr.com/2025/10/20/walrus_v0.2.0
I'm looking forward to hearing feedback from the community and the works of a 'distributed' version of walrus are in progress.
(it wouldn't let me post the links at the top for some reason)