Speaker
Description
Recent advancements in cloud-based architectures are transforming control systems in large-scale scientific facilities. At J-PARC MLF, the instrument control framework IROHA2 has been widely adopted for neutron experiments across many instruments. Although it has served as a standard framework for over a decade, its tightly coupled, synchronous communication model limits flexibility, making it difficult to adapt to evolving experimental requirements.
To address these challenges, we propose a next-generation instrument control framework as a post-IROHA2 system based on a hybrid cloud architecture, currently in the prototype phase. Emerging use cases such as hybrid cloud-based operation, multi-instrument integration, and data-driven and AI-based technologies require a more scalable and flexible approach. The proposed framework adopts a distributed architecture, where instrument and device control remain on the instrument side, while user-facing components and a Redis-based data store for managing experimental information are deployed in the cloud.
This architecture enables secure remote experiments and facilitates integration with advanced cloud-based AI services. At the same time, physical separation between components introduces additional latency and network constraints, which remain important considerations in system design. In this contribution, we present the hybrid cloud-based framework architecture, including performance evaluation, and outline its current status and future development plans.
