Microsoft's official blog describes why the construction of a new Json parser rather than continue to use industry guidelines Json.Net
Microsoft blog address: https://devblogs.microsoft.com/dotnet/try-the-new-system-text-json-apis/
In the official Github, but also a detailed description of this problem: https://github.com/dotnet/corefx/issues/33115
There is briefly described a .Net Core Span <T> class, which is similar to the array, but it is not hosted heap, thus providing a high-performance memory operation, which will serialization and deserialization provide higher performance .
At the same time a point UTF-8 and UTF-16 is Microsoft considered.
I began to learn more about it:
class Program { static void Main(string[] args) { Console.WriteLine("Hello World!"); string person = "[{\"name\":\"fanqi\",\"age\":25,\"phone\":[\"10086\",\"10010\"]},{\"name\":\"zhangrong\",\"age\":24,\"phone\":[\"10000\",\"10010\"]}]"; List<Person> personList = JsonSerializer.Deserialize<List<Person>>(person); string json = JsonSerializer.Serialize(personList, personList.GetType()); JsonDocument document = JsonDocument.Parse(person); Console.ReadKey(); } } class Person { public string name { get; set; } public int? age { get; set; } public List<string> phone { get; set; } }