We've been developing a new technology that we believe could benefit the Java community: a Redis-compatible in-memory data store created entirely in Java. This isn't just a cache; it's built to manage large datasets in RAM with full persistence while avoiding dependency on JVM garbage collection. Some notable advantages over Redis include significantly lower memory usage for typical datasets, ultra-fast snapshot speeds, and support for a wide range of data types and commands. The project is currently in an MVP stage and we plan to open-source it under the Apache 2.0 license if there's sufficient interest. I'm curious to know if an embeddable, Redis-compatible Java in-memory store would be valuable to you and what potential use cases you see for it, such as embedded analytics or memory-intensive applications that require consistent latency and efficient storage. I'd appreciate your feedback, suggestions, questions, and thoughts!
4 Answers
I think it's a good idea to gauge community interest before going full steam ahead. Getting feedback at this stage can save you time in developing documentation and features down the line.
I'm interested! An embedded, Redis-like Java store would definitely help with testing and could work great for our distributed batch processing applications. Just to clarify, would it operate across JVMs and support clustering?
Yes, that's one aspect we want to address with it. The goal is to create something that can scale and integrate smoothly into existing architectures.
What sets this apart from tools like Hazelcast? It looks like you have several key points such as a more compact memory use and a custom storage engine that could make a real difference. Does Hazelcast offer a similar API or commands?
The original poster mentioned using herded compression for RAM efficiency, which is intriguing. I wonder how that compares to Hazelcast's offerings.
Yes, very much interested in this! It could simplify things for prototypes and also offer a low learning curve compared to something like Hazelcast. Have you thought about a release timeline?
It would be awesome if it could be released soon. I’m sure many in the community would be eager to test.
Absolutely! It may even help in shaping the final product based on real user needs before investing time into that extensive wrapping up.