A .Net ULID implementation
A GUID/UUID can be suboptimal for many use-cases because:
- It isn't the most character efficient way of encoding 128 bits
- It provides no other information than randomness
A ULID however:
- Is compatible with UUID/GUID's
- 1.21e+24 unique ULIDs per millisecond (1,208,925,819,614,629,174,706,176 to be exact)
- Lexicographically sortable
- Canonically encoded as a 26 character string, as opposed to the 36 character UUID
- Uses Crockford's base32 for better efficiency and readability (5 bits per character)
- Case insensitive
- No special characters (URL safe)
PM> Install-Package NUlid
Or simply use the Nuget package manager GUI in Visual Studio.
Creating a ULID:
// Create a ULID
var myulid = Ulid.NewUlid();
// Print ULID
Console.WriteLine(myulid);Output:
01ASB2XFCZJY7WHZ2FNRTMQJCT
Parsing a ULID:
// Parse ULID:
var myulid = Ulid.Parse("01ASB2XFCZJY7WHZ2FNRTMQJCT");
// Print time-part of ULID:
Console.WriteLine(myulid.Time);Output:
4-8-2016 15:31:59 +00:00
You can also convert from/to GUID/UUID's, get the byte-representation of a ULID, create a ULID with specific timestamp and you can even specify an IUlidRng to use for generating the randomness (by default NUlid uses the CSUlidRng but a SimpleUlidRng is also provided, as well as a MonotonicUlidRng). The ULID is implemented as a struct with (operator) overloads for (in)equality, comparison etc. built-in and is, generally, very much like .Net's native Guid struct. An extensive helpfile is provided in the Nuget package and the testsuite also serves as a (simple) demonstration of NUlid's features.
Below is the current specification of ULID as implemented in this repository.
01AN4Z07BY 79KA1307SR9X4MV3
|----------| |----------------|
Timestamp Randomness
10 chars 16 chars
48bits 80bits
base32 base32
Timestamp
- 48 bit integer
- UNIX-time in milliseconds
- Won't run out of space till the year 10895 AD (this .Net specific Ulid implementation limits this to DateTimeOffset.MaxValue).
Randomness
- 80 (Whenever possible: Cryptographically secure) Random bits
As of v1.4.0 monotonic ULID's are supported (see below).
When generating a ULID within the same millisecond, it is possible to provide some guarantees regarding sort order (with some caveats). When you use the MonotonicUlidRng and a newly generated ULID in the same millisecond is detected, the random component is incremented by 1 bit in the least significant bit position (with carrying). For example:
// Create monotonic rng
var rng = new MonotonicRng();
// Create ULIDs, assume that these calls occur within the same millisecond:
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MCX
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MCY
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MCZ
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MD0
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MD1
Console.WriteLine(Ulid.NewUlid(rng)); // 01DBN5W2SG000DCBVYHX4T6MD2By default the most significant bit of the random part is set to zero; this ensures you can generate enough ULID's after the initial one before causing an overflow. Some implementations simply pick a random value for the random part and increment this value, however, there's a (very small) chance that this random part is close to the overflow value. If you then happen to generate a lot of ULID's within the same millisecond there is a risk the you hit the overflow. By our method we ensure there's enough 'room' for new values before 'running out of values' (overflowing). It is, with some effort, even possible to 'resume counting' from any given ULID.
Crockford's Base32 is used as shown. This alphabet excludes the letters I, L, O, and U to avoid confusion and abuse.
0123456789ABCDEFGHJKMNPQRSTVWXYZ
The components are encoded as 16 octets. Each component is encoded with the Most Significant Byte first (network byte order).
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| 32_bit_uint_time_high |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| 16_bit_uint_time_low | 16_bit_uint_random |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| 32_bit_uint_random |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| 32_bit_uint_random |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
ttttttttttrrrrrrrrrrrrrrrr
Where:
t is Timestamp
r is Randomness
Based on / inspired by alizain/ulid.
Below measurements are based on an Intel(R) Core(TM) i9-10900X CPU @ 3.70Ghz:
BenchmarkDotNet=v0.13.2, OS=Windows 11 (10.0.22621.963)
Intel Core i9-10900X CPU 3.70GHz, 1 CPU, 20 logical and 10 physical cores
.NET SDK=7.0.101
[Host] : .NET 7.0.1 (7.0.122.56804), X64 RyuJIT AVX2
DefaultJob : .NET 7.0.1 (7.0.122.56804), X64 RyuJIT AVX2
| Method | Mean | Error | StdDev | Gen0 | Allocated |
|------------------------------------- |----------:|---------:|---------:|-------:|----------:|
| Guid.NewGuid() | 58.14 ns | 0.334 ns | 0.296 ns | - | - |
| Ulid.NewUlid(SimpleUlidRng) | 50.32 ns | 0.193 ns | 0.171 ns | - | - |
| Ulid.NewUlid(CSUlidRng) | 117.07 ns | 0.862 ns | 0.720 ns | - | - |
| Ulid.NewUlid(SimpleMonotonicUlidRng) | 65.02 ns | 0.124 ns | 0.110 ns | - | - |
| Ulid.NewUlid(CSMonotonicUlidRng) | 65.58 ns | 0.119 ns | 0.099 ns | - | - |
| Guid.Parse(string) | 193.62 ns | 0.813 ns | 0.721 ns | 0.0095 | 96 B |
| Ulid.Parse(string) | 226.54 ns | 0.299 ns | 0.250 ns | 0.0181 | 184 B |
| Guid.ToString() | 168.76 ns | 0.597 ns | 0.558 ns | 0.0095 | 96 B |
| Ulid.ToString() | 144.66 ns | 0.504 ns | 0.447 ns | 0.0079 | 80 B |
| 'new Guid(byte[])' | 10.96 ns | 0.031 ns | 0.024 ns | 0.0040 | 40 B |
| 'new Ulid(byte[])' | 12.55 ns | 0.024 ns | 0.021 ns | 0.0040 | 40 B |
| Guid.ToByteArray() | 65.32 ns | 0.183 ns | 0.152 ns | 0.0039 | 40 B |
| Ulid.ToByteArray() | 127.23 ns | 0.337 ns | 0.298 ns | 0.0038 | 40 B |
| Ulid.ToGuid() | 117.48 ns | 0.434 ns | 0.406 ns | - | - |
| 'new Ulid(Guid)' | 64.26 ns | 0.574 ns | 0.537 ns | - | - |