Skip to content

A "blazingly" fast key-value pair database without bloat written in rust

Notifications You must be signed in to change notification settings

BarbossHack/firefly

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

55 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Firefly

A fast key-value pair in memory database. With a very simple and fast API. At Xiler it gets used to store and manage client sessions throughout the platform.

CLI help menu

Installation

Cargo: $ cargo install ffly

AUR: $ paru -S ffly

Docker image: arthurdw/firefly

Performance comparison

Database ops
Firefly 167k
Skytable 143k
Redis 67k

(push_it scripts can be found in ffly-rs/examples/)

Query Language

Firefly only has three data operators, NEW, GET, DROP. We'll walk over each one.

Important note on values

ALL values must be valid ASCII, if not the server will reject the query.

Defining query types

The server may only accept one query type per TCP connection type. The default type is a string query. But this can be changed by using the QUERY TYPE keyword.

QUERY TYPE 'STRING' | 'BITWISE';

String queries

The simplest way to query something is by querying it using a string query. But because the parsing, and bandwidth of this is more than that is necessary we also provide bitwise queries.

String queries are very loosely defined, as it evaluates the query definition from left to right. And can derive that if you use e.g. GETV, you mean GET VALUE. This works for any query, identifiers are case-insensitive.

Create

You can create a new record by using the NEW keyword, the arguments should be wrapped within quotes. The first argument (after NEW) is the key, this should be unique throughout the db. If no TLL value is provided, the server will use 0 (aka no expiry).

NEW '{key}'
[ VALUE ] '{value}'
[ WITH TTL '{ttl}'];
Create examples
NEW 's2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y'
VALUE '86ebe1a0-11bf-11ed-aa8e-13602e2ad46b';
NEW 's2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y'
VALUE '86ebe1a0-11bf-11ed-aa8e-13602e2ad46b'
WITH TTL '604800';

Fetch

The GET keyword returns the value and TTL by default. But if you only want one of the two, you can specify this. You can only search by key!

GET [VALUE | TTL] '{key}';
Fetch examples
GET 's2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y';
GET TTL 's2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y';

Delete

Deleting one record is as straightforward as fetching one. You can only delete whole records. If required all records that have a value can also be deleted, but this action is very expensive and generally not recommended.

DROP '{key}';
DROP ALL '{value}';

Bitwise queries

Because string queries can consume more resources than what is required, there is a more efficient (less friendly) way to interact with Firefly. This is by sending the bits in a specific format. This section just describes the formats, if you want more information about the queries itself, then you can find it in the String queries section.

Please keep in mind that all the bitwise queries are very strict, and if a query is unrecognized it will discard it.

General notes

  • All values are delimited by a NUL character. (0x0)
  • The end of the query is assumed to be the last byte.
  • Queries start with their type, this is a numeric value
    • 0: NEW
    • 1: GET
    • 2: GET VALUE
    • 3: GET TTL
    • 4: DROP
    • 5: DROP ALL
    • 6: QUERY TYPE STRING
    • 7: QUERY TYPE BITWISE
  • The query type does not need to be delimited

Bitwise create

A with TTL must always be provided. If you don't want a TTL set this to 0.

0{key}0x0{value}0x0{ttl}

These are the two same create examples from the string queries: 0s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y0x086ebe1a0-11bf-11ed-aa8e-13602e2ad46b0x0 0s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y0x086ebe1a0-11bf-11ed-aa8e-13602e2ad46b0x0602800

Bitwise fetch

1s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y 2s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y 3s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y

Bitwise delete

4s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y 5s2.8eqYursP2McHeQvHB2bauyE6n3vptOj8M96PxmGAQMDfimeZ31WAzP3hSw5Ixv5Y

About

A "blazingly" fast key-value pair database without bloat written in rust

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Rust 97.5%
  • Shell 2.2%
  • Dockerfile 0.3%