Gropius: Semi-Automated Identification and Management of Cross-Component Issues Considering Their Architectural Dependencies

The Gropius project, spearheaded by the University of Stuttgart's Institute for Software Engineering, addresses the growing complexities of modern software systems, particularly those utilizing microservices and component-based architectures regarding issue management. As these architectures grow in popularity due to their ability to manage complexity, they also bring about significant challenges, particularly with cross-component issues. These issues arise when problems in one component propagate and affect others, causing system-wide impacts that are difficult to trace and manage.

The core aim of the Gropius project is to develop semi-automated methods for identifying and managing these cross-component issues by integrating them with their architectural context. Traditional issue management systems (IMS) are typically limited to individual components and do not account for the dependencies and interactions between different components. This limitation often leads to inefficient and error-prone manual processes for issue tracking and resolution across multiple components.

Gropius proposes a novel approach by acting as a wrapper over traditional IMS, allowing for the synchronization and semantic linking of issues across different IMS projects, regardless of the IMS provider. This method enables developers to manage issues in a more integrated and context-aware manner, significantly improving the efficiency and effectiveness of issue resolution processes in complex software architectures.

1. Evaluation of Current Prototype and Prototype Enhancement: Conduct a pilot study to assess the efficiency and effectiveness of the existing Gropius method in an industrial setting. Based on the findings from the pilot study and integration tools, the Gropius prototype will be enhanced to address any identified gaps and improve its functionality.
2. Integration of Existing Projects: Given that many existing projects are deployed on Kubernetes, the project will develop tools to automatically recover the architecture and dependencies of these projects and transform them into Gropius models. This will enable seamless integration of Gropius into existing development environments, reducing the overhead of manual architecture documentation.
3. Identification and Analysis of Issue Propagation: Create approaches to detect and analyze issue relationships and propagations across components with NLP- and architecture-based approaches.
4. Comprehensive System Evaluation: The final phase involves a thorough evaluation of the enhanced Gropius system through an industrial field study.

• Ontological Metamodel: The Gropius method employs an ontological architecture model that allows the definition of custom component types and the relationships between these custom components. This sophisticated model supports various component types, including microservices, infrastructure services, libraries, and custom components, ensuring flexibility and adaptability to different project needs.
• Issue Relationship Detection: Utilizes advanced NLP techniques to semi-automatically identify semantic relationships between issues, facilitating more accurate and efficient issue tracking.
• Propagation Analysis: Detects potential issue propagations based on an architecture-based analysis, providing developers with critical insights into how issues might affect upstream and downstream components.

The project spans 24 months, starting on March 1, 2024.

• University of Stuttgart: Leading the research with a focus on software quality and architecture.
• Volkswagen AG: Serving as the industrial partner and providing real-world application contexts.

• Enhanced Gropius prototype capable of managing cross-component issues.
• Improved tools for semi-automated architecture recovery and integration.
• Validated methods for detecting and analyzing issue propagations and issue relations.
• Demonstrated scalability and practical benefits of the Gropius method in industrial environments.

This project (Grant No. 01IS23072) is part of the Software Campus initiative, supported by the German Federal Ministry of Education and Research (BMBF), aiming to foster future IT leaders and promote innovative research.