kruserr/rapiddb

A reasonably fast configurable embedded key-value sensor database

RapidDB

A reasonably fast configurable embedded key-value sensor database

Features

• Lightweight embedded database
• Simple key-value database interface
• Simple and flexible optional embedded REST API
• Memory first with on-disk persistence
• Memory Mapped Append-only Vector backing storage
• Bring your own database or API implementation
• Store sensor data inside a sensor database

Getting Started

Docker

Run database with docker

docker run -dit --rm -p 3030:3030 --name rapiddb kruserr/rapiddb:0.1

Further install options

Use the database with curl

Write to database with curl

curl -X POST localhost:3030/api/v0/test-0 -d '{"temp":4.00}'

Read from database with curl

curl localhost:3030/api/v0/test-0/latest

Explore the API with curl

curl localhost:3030/api/v0
curl localhost:3030/api/v0/sensors
curl localhost:3030/api/v0/test-0

Explore and customize the database

The database is highly customizable, if you use the database inside your cargo project. You can interact with the db object, and explore the IDatabase interface. You can also use warp::Filter to extend the API. You can also implement the IDatabase interface yourself, for your own database. Explore the docs to learn more, or look at the examples.

Documentation

Visit the Documentation.

Examples

Visit the Examples.

rapiddb's Issues

Add test code coverage

implement MemoryMappedHashMapDatabase

Implement IAsyncDatabase

Implement it such that there is a IDatabase and IAsyncDatabase, with corresponding high level implementations, such as MMAVDatabase and MMAVAsyncDatabase, while keeping the low level implementations the same or with small changes to support both high level implementations.

Remove MariaDB and Redis integration

MariaDB and Redis were exclusively used for benchmarking purposes, earlier in the project.
The way it was implemented it forced the end-user to download and build MariaDB and Redis dependencies even though they may not use this functionality.
This cleans up the codebase to better focus on the core functionality of the RapidDB project.

Redis integration may be revisited in the future.

use static dispatch in runtime

Theoretically should allow the compiler to further optimize code, which should give a performance improvement.

Emphasis on the fact that there is still no benchmark suite, such that we can not state that this is a real performance improvement but rather a theoretical one.

Enable builds without any webserver

• extract api into rapiddb-web
• use rapiddb-web in docker container
• publish rapiddb-web seperatly
• publish rapiddb without webserver
• update documentation