MODA26 is held in conjunction with ISC-HPC in Hamburg, Germany.
The web site for ISC-HPC is at:
https://isc-hpc.com
The Monitoring, Observability, and operational Data Analytics Workshop (MODA26) invites original contributions on monitoring and analyzing operational data in High Performance Computing (HPC) systems and data centers. We welcome submissions on ways to collect, store, visualize, interpret, and leverage large-scale system data, as well as the use of machine learning and AI techniques to enable proactive system control and optimization. New this year, the workshop explicitly invites contributions on observability and explainability of HPC system behavior. MODA26 also encourages contributions with regard to monitoring for integrated Quantum-HPC systems, and solutions that contribute to the successful co-design, procurement, and operation of next-generation HPC systems.
Workshop Goals
Establish common frameworks and standards to guide more consistent and effective MODA practices, and encourage work that closes the gap between simply collecting data and using it effectively to achieve real improvements in HPC operations.
Bring together experts to share practical solutions, discuss challenges, and explore new ideas for improving how we gather, analyze, and leverage operational data.
Identify current trends, highlight critical gaps, and shape the evolution of MODA, influencing the design, planning, and procurement of next-generation systems.
Scope and Topics
Collecting and analyzing operational data in HPC and data centers at scale
State-of-the-practice monitoring tools, methods, and techniques
AI/ML approaches to understand system behavior and improve operations
Critical evaluations of AI/ML approaches to ensure practical improvements for MODA
Integrating MODA into system software, runtime environments, and resource management
Solutions to increase observability and explainability of HPC systems
Data-driven strategies for predictive maintenance, scheduling, and energy optimization
Guidelines, tools, and best practices for energy efficiency and reporting
Approaches to ensure FAIR data practices, compliance, and trusted multitenancy
Successful real-world MODA deployments, case studies, and work-in-progress
Integration of monitoring and analysis for Quantum Computing and HPC
Monitoring, Observability, and Operational Data Analytics as drivers for digital twins of supercomputers
Contributions focused solely on application performance modeling, compiler analysis, debugging, or programming models are out-of-scope for the MODA26 workshop.