AWS has announced the general availability of AWS Transform, an agentic AI service aimed at accelerating and de-risking the modernization of enterprise mainframe and VMware workloads. The service leverages generative AI to automate these legacy systems’ assessment, planning, and transformation into cloud-based architectures. This follows AWS’s earlier introduction of similar generative AI-powered transformation capabilities for .NET applications within Amazon Q Developer, indicating a consistent strategy towards AI-driven modernization across various legacy platforms.
AWS Transform provides a multi-agent AI architecture that analyzes mainframe codebases, decomposes them into domains, and modernizes IBM z/OS applications to Java using orchestrated AI agents specializing in specific tasks. According to the company, this system leverages deep learning models and nearly two decades of AWS’s mainframe migration expertise. Users can interact with these agents to create customized modernization plans.
AWS Transform for Mainframe aims to significantly accelerate the traditionally multi-year mainframe modernization journey through capabilities such as comprehensive code analysis, automated documentation, and business logic extraction. These capabilities enhance understanding and facilitate modernization of mainframe applications. In addition, it also includes intelligent COBOL to Java refactoring and streamlined deployment solutions using infrastructure-as-code templates for efficient setup on cloud platforms.
(Source: AWS Migration & Modernization blog post)
Similarly, AWS Transform for VMware tackles the difficulties associated with migrating and modernizing VMware workloads, especially given the evolving VMware landscape. It automates application discovery, dependency mapping, migration planning, network conversion, and EC2 instance optimization. The process is structured in four phases: inventory discovery, wave planning, network conversion, and server migration, all managed through a collaborative web interface.
(Source: AWS News blog post)
However, Corey Quinn pointed out in a Last week in AWS blog post that discovering a new service term has tempered the enthusiasm surrounding AWS Transform’s capabilities. According to Quinn’s analysis of the updated AWS service terms, Section 50.14 for AWS Transform mandates:
By using AWS Transform, you agree to run the transformed workload on AWS for a minimum of 24 months from the date that the transformation is complete.
The promise of an AI agent tackling mainframe modernization, a notoriously complex endeavor, has also been met with skepticism within the tech community. As Artur Schneider, a Principal Cloud Consultant at Skaylink, commented in the LinkedIn post from Luc van Donkersgoed:
Huge props to anyone brave enough to let an AI agent give their mainframe a makeover, because what could possibly go wrong? Right? Right???
This sentiment echoes Quinn’s concerns about the potential for unforeseen issues and the lack of guarantees in the transformation process.
Lastly, AWS Transform for mainframe is available in the US East (N. Virginia) and Europe (Frankfurt) Regions. AWS Transform for VMware offers different availability options for data collection and migrations. AWS states that they currently provide core features, including assessment and transformation, at no cost to AWS customers.
