https://github.com/yankeguo-deprecated/idy
64bit ID generation service
https://github.com/yankeguo-deprecated/idy
Last synced: about 2 months ago
JSON representation
64bit ID generation service
- Host: GitHub
- URL: https://github.com/yankeguo-deprecated/idy
- Owner: yankeguo-deprecated
- License: mit
- Created: 2017-08-14T16:55:56.000Z (almost 8 years ago)
- Default Branch: master
- Last Pushed: 2017-08-14T17:02:52.000Z (almost 8 years ago)
- Last Synced: 2025-02-14T16:58:31.978Z (3 months ago)
- Language: Go
- Size: 27.3 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Idy
[](https://travis-ci.org/yanke-guo/idy)
Idy is a unique random id generation service, it starts with 0 and iterates over all uint64 numbers.
Idy supports clustering by pre-defined sharding pattern.
Idy has a database format easy to read and does not grow.
# Install / Upgrade
```base
go get -u github.com/yanke-guo/idy/cmd/idyd
```
# Usage
## Start
```
Usage of ./idyd:
-b string
HTTP service bind address (default "127.0.0.1:8865")
-d string
location of data directory (default "data")
-p string
comma seperated pool names, only ^[0-9a-zA-Z._-]+$ is allowed (default "user")
-s string
sharding partten of this instance, see README.md (default "1:1:2048:4096")
```
For example
```bash
idyd -b 0.0.0.0:8866 -d /data/idyd-data -p users,orders,comments -s 1:1:4096:9999
```
Idy will create database if not exists.
## HTTP API
Get a new id in hex format
```
curl http://127.0.0.1:8865/user/_hex
1e
```
Get a new id in decimal format
```
curl http://127.0.0.1:8865/user/_dec
24
```
# How It Works
Idy can provide multiple pools, each pool is a independent unique random number service.
You can use one pool for user id, another for order id.
For each pool, Idy creates a slice of continious UInt64 numbers, shuffles it with Fisher-Yates algorithm and fetch numbers from it.
If current slice drained, Idy move to next one, until max possible UInt64 number.
Idy database file is basically a json file, Idy will update database file automatically.
```javascript
{
"version": 1, // version of database file
"shard": "1:3:2048:4096" // sharding pattern
"seed": "1234567890998", // seed of current slice, used for shuffling, string as int64
"start": "0", // start of current slice (inclusive), string as uint64
"index": 0, // lates used index in slice, int
}
```
* `version`
`version` should be and only could be 1.
* `shard`
`shard` is sharding pattern, it describes how slices are created.
For single instance, first two numbers should be `1:1`
For 3 instances cluster, first two numbers should be set to `1:3`, `2:3` and `3:3`.
For last two number, `2048:4096` means each slice have 4096 numbers and 2048 of them will be used.
This will allow Idy skip some numbers, making your id unpredicatable.
* `seed`
`seed` is the random source of shuffling in current slice, regenerated and saved each time Idy moves to a new slice.
* `start`
`start` will be the start of this slice across the cluster.
That means, if `start` is `0`, `node-1` will use `[0,4096)`, `node-2` will use `[4096,8192)`, `node-3` will use `[8192,12288)`.
And when current slice drained, the `start` of next slice will be set to `12288`.
* `index`
the position of last used id in current slice.
Idy uses a local copy of golang `math/rand` package, once `seed` is determined (persisted in database file), a sequence of pseudo-random number is determined.
Thats why Idy can resume a whole suffled slice from `seed` and `start`, this will make database file small and no grow.
# Credits
Yanke Guo , see LICENSE file for MIT License.