DevID

The most useful, developer friendly ID or token generator.

Key Features

• Random enough to be used both as a random ID or a secure token
• Compact, just 24 characters (plus optional prefix, if you wish)
• Supports prefixes, providing clear context to developers and users alike
• Tamper-proof, with checksum embedded in the DevID that can be checked offline, without hitting a database
• Secure, with 2.8e+14 unique DevIDs per microsecond
• Human readable, by avoiding ambiguous characters
• Auto-lints inputs

Installation

npm install devid

Usage

importdevidfrom"devid";// Get a new DevID, without prefixconstid=devid();// type: DevIDconsole.log(id.toString());// Output: 0R2FWCHN822NKS5XGED6P5BC// Get a new DevID, with prefixconsole.log(devid("user").toString());// Output: user_0R2FWCHN822NKS5XGED6P5BC// Initialize a valid DevIDconstvalidId=devid("user_0R2FWCHN822NKS5XGED6P5BC");console.log(validId.toString());// Output: user_0R2FWCHN822NKS5XGED6P5BC// Initialize a valid, dirty DevIDconstvalidDirtyId=devid("user_or2fwndvlrk897784eotpz3k");console.log(validDirtyId.toString());// Output: user_0R2FWNDV1RK897784E0TPZ3K// Note that linting replaced o's with 0's and l with 1// Initialize a DevID with bad checksumconsttamperedId=devid("user_0R2FWCHN822NKS5XGED6P5BD");// throws a ChecksumError// Initialize a DevID with bad formatconstinvalidId=devid("user_0R2FWCHN822NKS5XGED6P5BCBCJE");// throws a FormatError

Technical Details

Prefix

The prefix must be less than 24 characters long. Anything beyond it will throw a FormatError.

Delimiter

The default delimiter is the underscore _, however you can change it to any of the following characters: _ - / \ ..

To do so, use setDelimiter:

import{devid,setDelimiter}from"devid";setDelimiter("-");console.log(devid("user").toString());// Output: user-0R2FWCHN822NKS5XGED6P5BD

Make sure you set it also when importing the module just to parse DevIDs.

DevID Format and Implementation

DevIDs are 15 bytes long.

The implementation of DevIDs includes:

• 7 bytes to store the generation timestamp, in microseconds
• 6 random bytes to introduce randomness within each microsecond
• 2 bytes for checksum, using CRC-16 (CCITT)

While a timestamp in microseconds would require 8 bytes to be stored, the first byte is dropped. This means that timestamps will reset to zero and duplication might occur approximately every 2,284 years.

The resulting 15 byte buffer is encoded to 24 characters using Crockford's Base 32 encoding, which allows to avoid ambiguous and special characters. The character space of Crockford's Base 32 is 0123456789ABCDEFGHJKMNPQRSTVWXYZ.

Based on the convention above, the first 11 characters of a DevID represent the timestamp and can be used to easily sort DevIDs. Note: Base 32 means that every character encodes 5 bits, so in reality the least significant bit of the timestamp is encoded in the 12th character of the DevID. However, given that the 12th character also encodes randomness, it cannot be used for sorting.

Sorting for the entire DevID cannot be guaranteed, however, due to the random bits used after the timestamp.

Here is an example of the process of generating a DevID:

1. Generation time: 2023-09-10T10:19:36.295761Z
2. Convert to timestamp in microseconds: 1694341176295761
3. Represent as a BigInt (Big Endian): 0x604FE8BF10151, or 00 06 04 fe 8b f1 01 51
4. Drop the first byte: 06 04 fe 8b f1 01 51
5. Generate 6 random bytes: 0x2ce2cc6e3fe3
6. Calculate the CRC-16 checksum of the 13 bytes obtained so far: 0x6438

############################################################################
# Timestamp # Random Bytes # CRC-16 #
############################################################################
# Bytes #
# 06 | 04 | fe | 8b | f1 | 01 | 51 | 2c | e2 | cc | 6e | 3f | e3 | 64 | 38 #
############################################################################
# Crockford's Base 32 #
# 0 R 2 F X 2 Z H 2 3 R J S R P C D R Z Y 6 S 1 R #
############################################################################