Year: 2026
Type: conference-paper
Abstract: Architectural performance models (APMs) achieve a high-level view on source code, enriched with performance-related annotations. This allows performance predictions to explore design alternatives. As a result, the APM can be adapted to gained insights. Since iterative software development practices (e.g., DevOps) embrace frequent code changes, keeping APMs and their source code consistent can become challenging and laborious. Thus, we want to investigate the continuous round-trip engineering between APMs and source code in DevOps. In this paper, we report on our current work in progress in this direction, where we build upon the Continuous Integration of architectural Performance Models approach, which handles the code-to-APM direction. In this paper, we present an approach for the direction APM to code, which obtains changes on an APM, preprocesses them, and (semi-)automatically propagates them via incremental, rule-based model transformations to the source code. We prototypically implemented parts of the approach and provide an initial evaluation. In the future, we plan to extend the implementation and evaluation and to investigate the role of LLMs.
Institution: Karlsruhe Institute of Technology
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: Deutsche Forschungsgemeinschaft
OpenAlex ID: W7203481641
