KEK Computer Center
Generated by GPT-5-miniExpansion Funnel Raw 78 → Dedup 0 → NER 0 → Enqueued 0
| KEK Computer Center | |
|---|---|
| Name | KEK Computer Center |
| Established | 1971 |
| Location | Tsukuba, Ibaraki, Japan |
| Type | Research computing facility |
| Parent | High Energy Accelerator Research Organization |
KEK Computer Center The KEK Computer Center provides high-performance computing and data services for particle physics and allied sciences at the High Energy Accelerator Research Organization in Tsukuba, Ibaraki. It supports experiments at facilities such as the KEK Belle II experiment, the Japan Proton Accelerator Research Complex, and international collaborations including CERN and Fermilab, integrating storage, networking, and middleware for large-scale data analysis. The center collaborates with universities and institutes like the University of Tokyo, Osaka University, and the National Institute of Informatics to enable distributed computing and grid services for projects spanning accelerator physics, materials science, and astrophysics.
Overview
The Computer Center operates as a regional hub within the global ecosystem of research computing, interfacing with projects such as LHC experiments like ATLAS and CMS, neutrino programs such as T2K and Super-Kamiokande, and synchrotron facilities like SPring-8. It provides services including batch processing for simulation campaigns tied to the International Linear Collider studies, data preservation for long-term collaborations with KEK partners, and authentication federations connected to identity providers like Eduroam and the Global Research Identifier. The center’s role aligns with national initiatives involving the Japan Science and Technology Agency and multilateral frameworks exemplified by agreements with JAXA and the RIKEN computational projects.
History and Development
Founded in the early 1970s to serve experiments at the National Laboratory for High Energy Physics site, the center evolved alongside landmark projects such as the TRISTAN collider and the KEKB accelerator. During the 1990s and 2000s it upgraded to support grid middleware from collaborations with European Grid Infrastructure and Open Science Grid, adopting protocols tested by experiments like BaBar and Belle. Strategic modernization efforts paralleled computing shifts driven by institutions such as IBM, Fujitsu, NEC, and vendors used in deployments for projects like SuperKEKB. The center’s development reflects influences from computational paradigms advanced by Google and Microsoft Research as well as standards from WLCG and the International Virtual Observatory Alliance.
Facilities and Infrastructure
The site houses server farms, tape libraries, and high-throughput networks linking to infrastructures such as the SINET backbone and international exchanges used by ESnet and GÉANT. Core hardware includes clusters with CPUs from vendors like Intel and AMD, accelerators from NVIDIA and Intel Xeon Phi, and storage solutions inspired by architectures from Hitachi and NetApp. Cooling and power systems comply with standards influenced by projects at CERN and data center research by the National Institute of Standards and Technology. The facility provides virtualized environments adopting platforms like OpenStack and container orchestration frameworks pioneered by Kubernetes to support reproducible workflows for collaborators at Kyoto University and Nagoya University.
Research and Services
Services include large-scale Monte Carlo simulation support for experiments such as Belle II and ATLAS, data reconstruction for detectors like Time Projection Chamber installations, and software distribution for analysis frameworks including ROOT and Geant4. The center supports development of machine learning pipelines using toolkits from TensorFlow and PyTorch for applications in experimental particle identification, detector alignment, and anomaly detection in data streams from instruments such as COMET and Hyper-Kamiokande. It provides identity and access management, cataloging via metadata schemas influenced by the Dublin Core community, and persistent identifier integration compatible with ORCID and DataCite for research outputs from groups at Tohoku University and Hokkaido University.
Projects and Collaborations
Key collaborations span multinational experiments and national laboratories including Belle II experiment, T2K, J-PARC, CERN, Fermilab, and SLAC National Accelerator Laboratory. The center participates in grid and cloud initiatives with WLCG, EGI, and national partners such as the University of Tsukuba computing centers. Collaborative software and data management projects involve contributions to DIRAC, dCache, Rucio, and workflow systems used in analyses for Astro-H and LIGO follow-up studies. Partnerships extend to industry through joint efforts with NEC, Fujitsu, and Hitachi to prototype energy-efficient computing informed by research at KEK accelerator projects like SuperKEKB.
Operations and Management
Operations follow practices common to large research facilities with shift-based system administration, incident response coordination modeled after procedures at CERN and ESnet, and security policies aligned with recommendations from NIST and national cyber frameworks. Governance includes oversight by the High Energy Accelerator Research Organization and advisory links to university consortia such as The University of Tokyo and national funding agencies like the Ministry of Education, Culture, Sports, Science and Technology (Japan), ensuring alignment with research priorities exemplified by programs at JAXA and the Japan Science and Technology Agency. Training and outreach engage with communities at conferences such as CHEP, SC Conference, and Supercomputing Conference to disseminate operational experience and foster collaborations.