Using AI to Automate the Modernization of Legacy Software Applications

Abstract

Translate Article

The task of modernizing legacy software has gained increasing attention in recent years. Old code is prone to security vulnerabilities, and is difficult to maintain and upgrade. Manual approaches to modernize legacy software involve intensive human effort and are challenging to scale up. Thus, there is an urgent need to develop automated techniques to modernize old code. In this proposal, we shall look at three aspects of this problem. The first is the conversion of legacy monolithic software architectures to modern microservice architectures. The second is the translation of code written in older programming languages like C, to code written in modern programming languages like Rust. The third is the detection of bugs that arise during modernization. We look at three prior papers (written by this author) that address each of these three aspects of application modernization. For each of these, we also present some ideas and directions for further research.