Turing·DB

Turing·DB Logo

TuringDB is a database, written in Rust, that aims to be distributed and scaled horizontally. It aims to be a replacement for where you don't need a relational database or a schema.

The database is backed by Sled key-value store

Motivation

The motive behind this database is to have a key/value database with ACID properties, speed, type-safety, changefeeds without polling, multi-cluster queries and replication. Rust is the most qualified for speed, type-safety and compile-time checks. Furthermore, sled.rs has been used as the embedded key/value store for this database as it is lock-free, has fully ATOMIC operations, zero-copy reads, SSD optimized log-storage and is written in Rust, hence inherits all the sweet properties of the language.

Features

The database aims to be:

  1. Really simple to use document database
  2. Have lookup and range capabilities
  3. Be partition tolerant and consistent
  4. Offer real-time push capabilities without polling, inspired by RethinkDB changefeeds
  5. Offer simple joins
  6. Offer optional distributed capabilities backed by Raft consensus algorithm
  7. Offer optional multi-cluster queries
  8. Be small enough to use as embedded database
  9. Be small and fast enough to be used on embedded devices or large servers
  10. Be really fun to use

Features under development include

  1. Replication
  2. Multi-cluster queries
  3. Changefeeds without polling, inspired by RethinkDB
  4. JSON support

Server Usage

  1. Install from crates-io

    sh $ cargo install turingdb-server

  2. Start the server

    sh $ turingdb-server

  3. Create a new cargo repository

    sh $ cargo new my-app

  4. Edit Cargo.toml file

    ```toml

    [dependencies]

    turingdb-helpers = #add the latest version here bincode = #add the latest version async-std = #add latest version here anyhow = # add latest version custom_codes = #add latest vesion ```

    Alternatively you could use cargo-edit if it is already installed, instead of adding dependencies manually

    sh $ cargo add turingdb-helpers bincode async-std anyhow custom_codes

  5. Open src/main.rs file in an editor

    ```rust use asyncstd::net::TcpStream; use asyncstd::io::prelude::*; use serde::{Serialize, Deserialize}; use custom_codes::DbOps;

    const BUFFERCAPACITY: usize = 64 * 1024; //16Kb const BUFFERDATA_CAPACITY: usize = 1024 * 1024 * 16; // Db cannot hold data more than 16MB in size

    [derive(Debug, Serialize, Deserialize)]

    struct DocumentQuery { db: String, document: Option, }

    [derive(Debug, Serialize, Deserialize)]

    pub (crate) struct FieldQuery { db: String, document: String, field: String, payload: Option>, }

    [async_std::main]

    async fn main() -> anyhow::Result<()> { let dbcreate = "db0".asbytes(); let mut packet = vec![0x02]; packet.extendfromslice(&db_create);

    let mut buffer = [0; BUFFER_CAPACITY];
    let mut container_buffer: Vec<u8> = Vec::new();
    let mut bytes_read: usize;
    let mut current_buffer_size = 0_usize;
    
    let mut stream = TcpStream::connect("127.0.0.1:4343").await?;
    stream.write(&packet).await?;
    
    loop {
        bytes_read = stream.read(&mut buffer).await?;
    
    
        // Add the new buffer length to the current buffer size
        current_buffer_size += buffer[..bytes_read].len();
    
        // Check if the current stream is less than the buffer capacity, if so all data has been received
        if  buffer[..bytes_read].len() < BUFFER_CAPACITY {
            // Ensure that the data is appended before being deserialized by bincode
            container_buffer.append(&mut buffer[..bytes_read].to_owned());
            dbg!(&container_buffer);
    
            dbg!(bincode::deserialize::<DbOps>(&container_buffer).unwrap());
    
            break;
    
        }
        // Append data to buffer
        container_buffer.append(&mut buffer[..bytes_read].to_owned());        
    }
    
    Ok(())
    

    } ```

Current query methods supported by the database

  1. Repository Queries

  2. Database Queries

  3. Document Queries

  4. Field Queries

Warning

A document cannot hold more that 16MiB of data and if this threshold is exceeded, an error from the custom_codes crate DbOps::EncounteredErrors([TuringDB::<GLOBAL>::(ERROR)-BUFFER_CAPACITY_EXCEEDED_16MB])

Contributing

We follow the Rust Code of Conduct in making contributions

License

All code contributions to this project must be licensed under Apache license

Acknowledgement

All libraries used in this project are subject to their own licenses