Automating Terraform Code Migration through Provider Evolution Knowledge
This program is tentative and subject to change.
Terraform is a popular Infrastructure-as-Code (IaC) tool for managing cloud environments, but its providers evolve rapidly, introducing frequent breaking changes. These updates pose migration challenges due to fragmented documentation and limited support, leading to delayed upgrades and accumulated technical debt. We present \textbf{TerraMod}, a framework that automates Terraform code migration across provider versions. It leverages external sources — changelogs, API schemas, and deprecation links to construct a migration context that guides LLMs towards effective upgrades. Evaluated on real-world breaking changes from the AWS Provider, \textbf{TerraMod} outperforms standard prompting techniques, enabling reliable upgrades. The dataset will be released upon publication.
This program is tentative and subject to change.
Fri 17 AprDisplayed time zone: Brasilia, Distrito Federal, Brazil change
14:00 - 15:30 | AI for Software Engineering 25Journal-first Papers / Research Track / New Ideas and Emerging Results (NIER) / Demonstrations at Europa II | ||
14:00 15mTalk | ArtifactSync: Automated Repository Synchronization through Hierarchical Change Impact Analysis Demonstrations Ebube Alor Concordia University, João Pedro de Souza Olivo Tardivo Universidade Estadual do Paraná, SayedHassan Khatoonabadi Concordia University, Emad Shihab Concordia University | ||
14:15 15mTalk | Introducing Phylogenetics in Search-based Software Engineering: Phylogenetics-aware SBSE Journal-first Papers Daniel Blasco SVIT Research Group. Universidad San Jorge, Antonio Iglesias Universidad San Jorge, Jorge Echeverria Universidad San Jorge, Francisca Perez Universitat Politècnica de València, Carlos Cetina | ||
14:30 15mTalk | Automating Terraform Code Migration through Provider Evolution Knowledge New Ideas and Emerging Results (NIER) Pranjal Gupta IBM Research, Pooja Aggarwal IBM Research, Brent Paulovicks IBM Research, Prateeti Mohapatra IBM Research, Rong Lee IBM Research, Vadim Sheinin IBM Research | ||
14:45 15mTalk | Replacing Training with Reasoning: Reinterpreting Classic ML Pipelines with LLMs New Ideas and Emerging Results (NIER) Marco Alecci University of Luxembourg, Jordan Samhi University of Luxembourg, Luxembourg, Tegawendé F. Bissyandé University of Luxembourg, Jacques Klein University of Luxembourg | ||
15:00 15mTalk | NB2P: Generating Data Science Pipelines from Computational Notebooks Research Track Haotian Gao National University of Singapore, Singapore and NUSRI Chongqing, China, Quang Trung Ta National University of Singapore, Tien Tuan Anh Dinh Deakin University, Australia, Nhut Minh Ho National University of Singapore, Zhiyong Huang National University of Singapore, Beng Chin Ooi National University of Singapore, Singapore | ||
15:15 15mTalk | Multi-Location Software Model Completion Research Track | ||