toydb VS sled

This is great, but you might want to consider a different name. There's already a Rust project called ToyDB, and it's a distributed database with a Raft log, SQL, disk persistence, ACID transactions, etc. It's under active development (though the developer now works at Cockroach Labs), and has 5K stars on GitHub, so I think they have the right to the name.

Build something you’d like to learn about. Things I’ve considered replicating: A distributed database (see https://github.com/erikgrinaker/toydb), an interpreter (crafting interpreters is a good book), a Ray tracer (http://raytracerchallenge.com/), an RPC compiler and framework, a simpler neural network framework ( https://github.com/pjreddie/darknet)…

Hilariously, I was trying out ToyDB on the Lichee-RV recently. While it does compile and run the five-node example setup (and memory usage is surprisingly low, which is a plus considering the 0.5GB of RAM), performance is three orders of magnitude lower than on a desktop x86 PC. Some of that is due to just having a single core run 5 nodes, some is due to the lower clock speed and slower memory, and some is due to slower storage (SD card). I don't think that explains everything, so I may investigate that later.

toydb

Well I think if you could replicate this https://github.com/erikgrinaker/toydb anybody would hire you.

a few times, seems interesting. The author's also built a lot of other cool concurrency primitives for Rust as well.

[0] https://github.com/spacejam/sled

- Dropbox [3] uses a similar approach but they talk about it a bit more abstractly.

Sans-IO is more documented in Python [4], but str0m [5] and quinn-proto [6] are the best examples in Rust I’m aware of. Note that sans-IO is orthogonal to deterministic test frameworks, but it composes well with them.

With the disclaimer that my opinions are mine and mine alone, and don’t reflect the company I work at —— I do work at a rust shop that has utilized these techniques on some projects.

TigerBeetle is an amazing example and I’ve looked at it before! They are really the best example of this approach outside of FoundationDB I think.

[0]: https://risingwave.com/blog/deterministic-simulation-a-new-e...

[1]: https://risingwave.com/blog/applying-deterministic-simulatio...

[2]: https://dropbox.tech/infrastructure/-testing-our-new-sync-en...

[3]: https://github.com/spacejam/sled

[4]: https://fractalideas.com/blog/sans-io-when-rubber-meets-road...

[5]: https://github.com/algesten/str0m

[6]: https://docs.rs/quinn-proto/0.10.6/quinn_proto/struct.Connec...

Sled uses bw-tree actually https://github.com/spacejam/sled/wiki/sled-architectural-outlook

There is a valid argument to be made for threads over async in a large percentage of use cases where async is considered the default. If this is what you are referring to however, I don't think they ever referred to async as completely useless: https://github.com/spacejam/sled/issues/1123.

Have you looked into other pure-Rust databases as well, such as sled or GlueSQL which has an SQL interface on top of sled? I wonder how those would compare to Persy.

I'm not sure what you need. Are these of any use? https://github.com/meilisearch/heed https://github.com/spacejam/sled

There is also Sled but as I understand it that is being reworked to use the author's new DB core Marble

There are some databases like sled/FlashDB designed to be embedded to other applications even bare metal microcontrollers. But I do doubt the potential bussiness value of a pure embedded database.

https://github.com/spacejam/sled

To add transaction support, you probably need a good understanding of how the memtable works in Log Structured Merge trees: