Speaker
Description
Providing a remote experimental environment for facility users and utilizing advanced computational resources of cloud environments for instrument control are challenging issues for scientific user facilities such as J-PARC MLF. Therefore, we have developed a hybrid cloud-based instrument control system by modifying IROHA2, which is the standard instrument control software framework for neutron scattering experiments in J-PARC MLF. IROHA2 comprises four core server components, namely, device control, instrument management, integrate control, and sequence management, each of which serves as a web interface.
In the newly developed system, the server components for device control and instrument management are deployed in the local computing environment of instruments, while the front-end server components are deployed in the Amazon Web Services (AWS), one of the leading cloud environments, thereby creating a hybrid cloud-based distributed system. Moreover, a loosely coupled distributed system is realized by introducing event-driven asynchronous communication between computing environments via a message broker service in AWS. This prevents communication and system/service failures from spreading to the entire distributed system, improving fault tolerance and availability of the distributed system. Furthermore, Pub/Sub messaging via message brokers is highly compatible with microservice architecture, enabling flexible and easy integration of our system with various services and computing resources in cloud environments (e.g., for automatic measurement and feedback control).
Using the developed system, users can access the front-end of the cloud through a web browser to remotely control instruments, including data collection and control of sample environment equipment.
In this presentation, we discuss the features of our hybrid cloud-based system and further present future development plans.
| Abstract publication | I agree that the abstract will be published on the web site |
|---|
