AWS Transform for.NET
AWS Transform for .NET: Agentic AI Service Transforms Massive-Modification of Scale.NET Applications
AWS Transform For .NET, the first agentic AI solution created especially to speed up the modernization of.NET applications at scale, is now generally available, according to Amazon Web Services (AWS). The time, expense, and complexity often involved in moving legacy .NET Framework programs to more contemporary, cross-platform .NET versions that can function well on Linux are intended to be greatly decreased by this new service.
It may be quite difficult to update .NET Framework apps to the open-source, cross-platform .NET. Detailed analysis, incompatibility identification, manual code adjustments during porting, and extensive validation are all necessary for this labour-intensive and error-prone procedure. This difficulty is exacerbated for big businesses that may be in charge of hundreds of.NET Framework apps. In order to address this, AWS Transform For .NET uses generative AI and agentic AI Large Language Models (LLMs), particularly using knowledge from Amazon Bedrock LLMs, to automate a number of these intricate processes.
AWS claims that the service may up to four times speed up the transformation process. Reducing transformation costs across analysis, planning, refactoring, and tooling, as well as cutting running expenses by up to 40% by eliminating Windows licensing requirements, are some of the main advantages mentioned. Improved code quality, speed, and security are anticipated from modernized apps. On Linux, they might be able to manage increasing workloads with 50% greater scalability and operate 1.5–2 times quicker with enhanced performance. Additionally, compared to equivalent x86-based instances, it is claimed that running these updated programs on AWS Graviton-based instances offers superior pricing performance.
The service provides two unique user experiences: an extension for individual project and solution porting within the Visual Studio IDE, and a single online experience designed for the large-scale modernization of possibly hundreds of applications across many teams. A .NET domain-expert agent communicates with users, frequently using natural language.
The agent may access source code repositories (supporting GitHub, Bitbucket, and GitLab) via the web interface, analyse them for supported project types, versions, and package dependencies, and provide a report on the pre-transformation evaluation. The agent suggests customized modernization approaches based on the codebase, dependencies, and specified business objectives. Users have the option to manually pick repositories or modify the suggested strategy.
In a safe, network-isolated setting, the agent automates the transformation procedure after user consent. For a variety of supported project types, including MVC (including Razor views), WCF, Web API, Console applications, and unit test projects, this includes code conversion and refactoring. The agent fixes build issues iteratively and manages third-party packages.
The agent’s ability to independently run unit tests after a successful transformation and confirm Linux readiness is a noteworthy feature. The original source code is preserved when the modified code is committed to a new branch in the user’s repository. Summaries of natural language transformations are produced for convenient change review. Updates and deep links to redesigned repositories for evaluation are sent via email alerts.
New Capabilities
- There is now support for projects that depend on private NuGet packages. Users can submit missing packages, or the service can identify and automatically change the source code of private NuGet packages if it finds them.
- Following the completion of the change, unit test execution is now supported.
- Project-level natural language summaries and comprehensive JSON reports are among the enhanced transformation visibility and summary reporting options offered.
- The UI layer of MVC apps may now be supported by automatically porting MVC Razor views to ASP .NET Core Razor views.
- Integration with Bitbucket, GitLab, and GitHub repositories is made possible via expanded connector support.
- Dependent repositories are chosen for transformation based on the automated detection of cross-repository relationships.
- Users may obtain comprehensive evaluation results for private NuGet packages and recognised repositories.
With a focus on cross-platform .NET 8, the service presently supports translating apps created with.NET Framework versions 3.5+,.NET Core 3.1, and .NET 5+. Libraries, console programs, Web APIs, Business Logic Layers of SPA backends, MVC apps (including front-end Razor views), Windows Communication Foundation (WCF) services, and unit test projects (NUnit, xUnit, MSTest) are among the application types and C# code projects that are the subject of this course. But without core compatible libraries, it doesn’t change WebForms (.aspx), WinForms, Blazor UI elements, or Win32 DLLs.
Additionally, it doesn’t change apps that are already in .NET 8.0+. The service does not change the original repository branches; instead, it writes to a distinct target branch. In order to upload missing packages, validate plans, set up connections, and evaluate and accept changed code, human assistance may be required.
Significant outcomes have been reported by early adopters:
- According to Thomson Reuters, AWS Transform seems like an extension of their team, enabling them to work more efficiently and speeding up not just modernisation but everything.
- According to The Hartford, it sped up their modernisation efforts and enhanced speed to market, reducing the time it took to adapt old code from months to weeks.
- GTI successfully completed a .NET Framework to.NET 8 port upgrade with a 70% savings in effort.
- Signaturit used the service to cut a project that was supposed to take six to eight months down to a few days.
- Over 143,000 lines of .NET Framework code were converted to cross-platform.NET by the Caribbean Examinations Council in less than two days, saving an estimated 270 development hours.
- 90% of a codebase was modified by Planet DDS, which finished a three-month project in less than a week. The tool was found to properly handle intricate modifications, such as dependency injection setup.
- Experian demonstrated a 47% productivity improvement with 80% automation across 687,600 lines of code, resulting in a 49-sprint decrease in total effort across seven.NET apps. Additionally, each upgrade project work was reduced from 15 to 8 sprints.
Currently, AWS Transform For .NET is accessible in the US East (N. Virginia) and Europe (Frankfurt) regions. The use of AWS Transform itself is free of charge. Only AWS resources generated or utilised by the transformation’s output are subject to standard pricing.
To begin using this agentic AI service, You may update your .NET apps to be cross-platform compatible by using the AWS Transform For .NET agent. Refer to this feature as “.NET modernization.” You may start a AWS Transform For .NET modernization task after configuring your workspace in AWS Transform.
