.Net 9.0 is scheduled for official release next month. Recently, as the second release candidate (RC2) of .NET 9.0 has been released, attention has been focused on the newly added features and improvements.
On the 10th, .NET 9.0 RC2 version was released.
According to the document introducing new features of .NET 9 that was reviewed on the 17th, .NET 9 includes a runtime, a library, and a software development kit (SDK). There are changes in ML.net, etc. It also includes ‘.NET Aspire’, a simplified cloud-native technology stack that was officially released last May.
.NET 9 comes after .NET 8 and focuses on native app development and cloud performance. .NET 9 is expected to be officially released in time for the ‘.NET Conf’, which will be held online from the 12th to the 14th of next month. Technical support is provided for 18 months with the version according to the standard support period (STS).
Microsoft.Extentios.AI library concept diagram
.NET 9 rc2 creates the Microsoft.Extensions.AI library. Microsoft described it as “an integrated C# abstraction layer for interacting with artificial intelligence (AI) services.” It is a middleware that can easily connect AI services such as small and large language models (SLM, LLM) and insert them into applications developed based on .NET.
“AI capabilities are evolving rapidly, and common patterns are emerging for features such as ‘chatting’, nesting, and device notifications,” said Luis Quintanilla, program manager at Microsoft. “It is necessary to do this,” he explained.
“This middleware can add valuable functionality without burdening the producer, providing an immediate benefit to the user,” he said, “It aims to create attractions that can be placed the functionality across multiple services, tailored to a specific provider’s API release services,” he said.
Microsoft.Extensions.AI provides a unified API abstraction for AI services, similar to logging abstractions and dependency injection (DI). The plan is to implement standards for caching, telemetry, device notifications, and other common functions that work with all providers.
By using the abstraction layer for integration, you don’t need to modify or write new code separately every time you connect multiple types of AI services to a .Net application. You can maintain a single API across your applications and experiment with multiple packages using native abstractions.
Currently, reference implementations such as OpenAI, Azure AI Inference, and Ollama are provided.
.NET 9 includes version 4.0 of ‘ML.NET’, an open source cross-platform framework that can integrate machine learning models into .NET applications. Added support for tokenizers like Tiktoken and additional tokenizers from modules like Llama and CodeGen.
.NET Aspire Package
The officially released .NET Aspire is a simplified native service development stack for building distributed container-based applications suitable for the business environment. Includes tools, templates, and NuGet packages to help you build distributed applications in .NET more easily. It covers the entire stack, including the developer dashboard, orchestration, components, service discovery, and deployment.
In addition, .NET 9 has more performance and quality improvements. Apart from the C # 13 version, most of the focus is on improvements. They write that .NET MAUI, a multi-format app UI, aims to improve quality. The goal seems to be to persuade a conservative developer base by focusing on quality improvement rather than causing anxiety by aggressively introducing new frameworks and features.
to write Byline Network
yong2@byline.network
