monoton

README

Monoton

Highly scalable, single/multi node, predictable and incremental unique id generator with zero allocation magic.

Installation

Via go packages: go get github.com/mustafaturan/monoton/v2

API

The method names and arities/args are stable now. No change should be expected on the package for the version 2.x.x except any bug fixes.

Usage

Using with Singleton

Create a new package like below, and then call Next() or NextBytes() method:

package uniqid

// Import packages
import (
	"fmt"
	"github.com/mustafaturan/monoton/v2"
	"github.com/mustafaturan/monoton/v2/sequencer"
)

var m monoton.Monoton

// On init configure the monoton
func init() {
	m = newIDGenerator()
}

func newIDGenerator() monoton.Monoton {
	// Fetch your node id from a config server or generate from MAC/IP address
	node := uint64(1)

	// A unix time value which will be subtracted from the time sequence value.
	// The initialTime value type corresponds to the sequencer type's time
	// representation. If you are using Millisecond sequencer then it must be
	// considered as Millisecond
	// If we want to init the time with 2020-01-01 00:00:00 PST
	initialTime := uint64(1577865600000)

	// Configure monoton with a sequencer and the node
	m, err = monoton.New(sequencer.NewMillisecond(), node, initialTime)
	if err != nil{
		panic(err)
	}

	return m
}

func Generate() string {
	m.Next()
}

func GeneateBytes() [16]byte {
	m.NextBytes()
}

In any other package generate the ids like below:

import (
	"fmt"
	"uniqid" // your local uniqid package from your project
)

func main() {
	for i := 0; i < 100; i++ {
		fmt.Println(uniqid.Generate())
	}
}

Using with Dependency Injection

package main

// Import packages
import (
	"fmt"
	"github.com/mustafaturan/monoton/v2"
	"github.com/mustafaturan/monoton/v2/sequencer"
)

func NewIDGenerator() monoton.Monoton {
	// Fetch your node id from a config server or generate from MAC/IP address
	node := uint64(1)

	// A unix time value which will be subtracted from the time sequence value.
	// The initialTime value type corresponds to the sequencer type's time
	// representation. If you are using Millisecond sequencer then it must be
	// considered as Millisecond
	initialTime := uint64(0)

	// Configure monoton with a sequencer and the node
	m, err := monoton.New(sequencer.NewMillisecond(), node, initialTime)
	if err != nil{
		panic(err)
	}

	return m
}

func main() {
	g := NewIDGenerator()

	for i := 0; i < 100; i++ {
		fmt.Println(g.Next())
	}
}

Features

Time Ordered

The monoton package provides sequences based on the monotonic time which represents the absolute elapsed wall-clock time since some arbitrary, fixed point in the past. It isn't affected by changes in the system time-of-day clock.

Please refer to ADR 01 - Time for details and consequences.

Initial Time

Initial time value opens space for time value by subtracting the given value from the time sequence.

Readable

The monoton package converts all sequences into Base62 format. And Base62 only uses ASCII alpha-numeric chars to represent data which makes it easy to read, predict the order by a human eye.

The total byte size is fixed to 16 bytes for all sequencers. And at least one byte is reserved to nodes.

Please refer to ADR 02 - Encoding for details and consequences.

Multi Node Support

The monoton package can be used on single/multiple nodes without the need for machine coordination. It uses configured node identifier to generate ids by attaching the node identifier to the end of the sequences.

Extendable

The package comes with three pre-configured sequencers and Sequencer interface to allow new sequencers.

Included Sequencers and Byte Orderings

The monoton package currently comes with Nanosecond, Millisecond and Second sequencers. And it uses Millisecond sequencer by default. For each sequencer, the byte orders are as following:

Second:      16 B =>  6 B (seconds)      + 6 B (counter) + 4 B (node)
Millisecond: 16 B =>  8 B (milliseconds) + 4 B (counter) + 4 B (node)
Nanosecond:  16 B => 11 B (nanoseconds)  + 2 B (counter) + 3 B (node)

Please refer to ADR 03 - Byte Sizes for details and consequences.

New Sequencers

The sequencers can be extended for any other time format, sequence format by implementing the monoton/sequencer.Sequencer interface.

Benchmarks

Command:

go test -benchtime 10000000x -benchmem -run=^$ -bench=. github.com/mustafaturan/monoton/v2

Results:

goos: darwin
goarch: amd64
pkg: github.com/mustafaturan/monoton/v2
cpu: Intel(R) Core(TM) i5-6267U CPU @ 2.90GHz
BenchmarkNext-4        	10000000	       108.7 ns/op	       0 B/op	       0 allocs/op
BenchmarkNextBytes-4   	10000000	        99.88 ns/op	       0 B/op	       0 allocs/op
PASS
ok  	github.com/mustafaturan/monoton/v2	2.194s

Contributing

All contributors should follow Contributing Guidelines and ADR docs before creating pull requests.

Credits

Mustafa Turan

License

Apache License 2.0

Copyright (c) 2019 Mustafa Turan

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Documentation

Overview

Package monoton is a highly scalable, single/multi node, human-readable, predictable and incremental unique id generator

Time Ordered

The monoton package provides sequences based on the monotonic time which represents the absolute elapsed wall-clock time since some arbitrary, fixed point in the past. It isn't affected by changes in the system time-of-day clock.

Initial Time

Initial time value opens space for the time value by subtracting the given value from the time sequence.

Readable

The monoton package converts all sequences into Base62 format. And Base62 only uses ASCII alpha-numeric chars to represent data which makes it easy to read, predict the order by a human eye.

The total byte size is fixed to 16 bytes for all sequencers. And at least one byte is reserved to nodes.

Multi Node Support

The monoton package can be used on single/multiple nodes without the need for machine coordination. It uses configured node identifier to generate ids by attaching the node identifier to the end of the sequences.

Extendable

The package comes with three pre-configured sequencers and Sequencer interface to allow new sequencers.

Included Sequencers and Byte Orderings

The monoton package currently comes with Nanosecond, Millisecond and Second sequencers. And it uses Millisecond sequencer by default. For each sequencer, the byte orders are as following:

Second:      16 B =>  6 B (seconds)      + 6 B (counter) + 4 B (node)
Millisecond: 16 B =>  8 B (milliseconds) + 4 B (counter) + 4 B (node)
Nanosecond:  16 B => 11 B (nanoseconds)  + 2 B (counter) + 3 B (node)

New Sequencers

The sequencers can be extended for any other time format, sequence format by implementing the monoton/sequncer.Sequencer interface.

Example using Singleton

package uniqid

// Import packages
import (
	"fmt"
	"github.com/mustafaturan/monoton/v2"
	"github.com/mustafaturan/monoton/v2/sequencer"
)

const year2020asMillisecondPST = 1577865600000

var m monoton.Monoton

// On init configure the monoton
func init() {
	m = newIDGenerator()
}

func newIDGenerator() monoton.Monoton {
	// Fetch your node id from a config server or generate from MAC/IP
	// address
	node := uint64(1)

	// A unix time value which will be subtracted from the time sequence
	// value. The initialTime value type corresponds to the sequencer type's
	// time representation. If you are using Millisecond sequencer then it
	// must be considered as Millisecond
	initialTime := uint64(year2020asMillisecondPST)

	// Configure monoton with a sequencer and the node
	m, err = monoton.New(sequencer.NewMillisecond(), node, initialTime)
	if err != nil{
		panic(err)
	}

	return m
}

func Generate() string {
	m.Next()
}

// In any other package unique ids can be generated like below:

package main

import (
	"fmt"
	"uniqid" // your local `uniqid` package inside your project
)

func main() {
	for i := 0; i < 100; i++ {
		fmt.Println(uniqid.Generate())
	}
}

Index

• type MaxByteSizeError
  • func (e *MaxByteSizeError) Error() string
• type MaxNodeCapacityExceededError
  • func (e *MaxNodeCapacityExceededError) Error() string
• type Monoton
  • func New(s sequencer.Sequencer, node, initialTime uint64) (Monoton, error)
  • func (m Monoton) Next() string
  • func (m Monoton) NextBytes() [16]byte

Examples

• Monoton.Next
• New

Types

type MaxByteSizeError

type MaxByteSizeError struct {
	ByteSizeSequence     int
	ByteSizeSequenceTime int
	ByteSizeTotal        int
}

MaxByteSizeError is an error type with sequence & time byte sizes

func (*MaxByteSizeError) Error

func (e *MaxByteSizeError) Error() string

type MaxNodeCapacityExceededError

type MaxNodeCapacityExceededError struct {
	Node    uint64
	MaxNode uint64
}

MaxNodeCapacityExceededError is an error type with node information

func (*MaxNodeCapacityExceededError) Error

func (e *MaxNodeCapacityExceededError) Error() string

type Monoton

type Monoton struct {
	// contains filtered or unexported fields
}

Monoton is a sequential id generator

func New

func New(s sequencer.Sequencer, node, initialTime uint64) (Monoton, error)

New inits a new monoton ID generator with the given generator and node.

Example

s := sequencer.NewMillisecond() // has 4 bytes free space for a node
n := uint64(19)                 // Base62 => J
t := uint64(0)                  // initial time (start from unix time in ms)

m, err := monoton.New(s, n, t)
if err != nil {
	panic(err)
}
fmt.Println(m.Next()[12:])

Output:

000J

func (Monoton) Next

func (m Monoton) Next() string

Next generates next incremental unique identifier as Base62 The execution returns the following Bytes (B) for the known sequencer types:

Second:      16 B =>  6 B (seconds)      + 6 B (counter) + 4 B (node)
Millisecond: 16 B =>  8 B (milliseconds) + 4 B (counter) + 4 B (node)
Nanosecond:  16 B => 11 B (nanoseconds)  + 2 B (counter) + 3 B (node)

For byte size decisions please refer to docs/adrs/byte-sizes.md

Example

s := sequencer.NewSecond()     // sequencer.Second
n := uint64(19)                // Base62 => J
t := uint64(time.Now().Unix()) // initial time (start from unix time in s)

m, err := monoton.New(s, n, t)
if err != nil {
	panic(err)
}
fmt.Println(len(m.Next()))

Output:

16

func (Monoton) NextBytes

func (m Monoton) NextBytes() [16]byte

NextBytes generates next incremental unique identifier as Base62 16 bytes array The execution returns the following Bytes (B) for the known sequencer types:

Second:      16 B =>  6 B (seconds)      + 6 B (counter) + 4 B (node)
Millisecond: 16 B =>  8 B (milliseconds) + 4 B (counter) + 4 B (node)
Nanosecond:  16 B => 11 B (nanoseconds)  + 2 B (counter) + 3 B (node)

For byte size decisions please refer to docs/adrs/byte-sizes.md