Eastern Canada Research Grid

Eastern Canada Research Grid
Name	Eastern Canada Research Grid
Type	Research infrastructure
Established	2010s
Headquarters	Halifax, Nova Scotia
Region served	Atlantic Canada, Quebec, Ontario

Eastern Canada Research Grid

The Eastern Canada Research Grid is a distributed high-performance computing and data infrastructure serving universities and research institutes across Atlantic Canada, Quebec, and Ontario. It links compute clusters, storage systems, and campus networks to enable simulations, data analytics, and multidisciplinary projects involving partners such as Dalhousie University, McGill University, and the University of Toronto. The Grid supports collaborations that include climate modelling, genomics, oceanography, and materials science with connections to national and international initiatives.

Overview

The Grid connects institutional nodes at institutions including Dalhousie University, Memorial University of Newfoundland, Saint Mary’s University (Halifax), Université Laval, McGill University, University of Toronto, and Queen’s University at Kingston. It interoperates with national platforms such as Compute Canada and international frameworks like European Grid Infrastructure, XSEDE, and PRACE. Core services include batch scheduling provided by systems like Slurm (software), distributed file systems influenced by Lustre (file system), and middleware patterned after Globus (software), OpenStack, and Hadoop. Network interconnectivity relies on regional backbones such as CANARIE and provincial research networks comparable to BCNET and ON2.

History and Development

Early discussions involved stakeholders from Dalhousie University, Memorial University of Newfoundland, Acadia University, and St. Francis Xavier University alongside provincial research councils and entities like Mitacs and the Natural Sciences and Engineering Research Council of Canada. Pilot deployments in the 2010s referenced architectures used by Compute Canada and drew technical guidance from projects such as GridPP and Open Science Grid. Subsequent expansion phases aligned with funding rounds from agencies analogous to Canada Foundation for Innovation and partnerships with commercial vendors including Dell Technologies, Hewlett Packard Enterprise, and Cisco Systems. Collaborative governance models referenced precedents at TeraGrid and NorduGrid.

Infrastructure and Technical Architecture

The Grid’s architecture comprises heterogeneous clusters at universities, high-capacity storage arrays modeled on EMC Corporation solutions, and virtualization layers using KVM (kernel-based virtual machine), Docker, and Kubernetes. Authentication and identity management mirror federated approaches like Shibboleth and InCommon, while data transfer leverages tools inspired by GridFTP and rsync. Inter-site networking uses fiber links coordinated with CANARIE and regional exchanges, and monitoring and orchestration adopt toolsets similar to Nagios, Prometheus (software), and Ansible. Scientific workflows are executed using engines comparable to Pegasus (workflow management), Nextflow, and CWL (Common Workflow Language).

Research Applications and Collaborations

Researchers in climate science at Université du Québec à Montréal and oceanography at Dalhousie University use the Grid for runs of models like MITgcm, WRF, and ECMWF-compatible tools, often collaborating with groups at Fisheries and Oceans Canada and the Canadian Meteorological Centre. Life sciences teams at McGill University and University of Toronto perform genomics analyses with pipelines based on GATK and BLAST and collaborate with hospitals like Montreal General Hospital and regulatory bodies such as Health Canada. Materials researchers at McMaster University and Queen’s University at Kingston run density functional theory codes like VASP and Quantum ESPRESSO. Cross-disciplinary projects include partnerships with environmental NGOs, provincial ministries, and consortia linked to ArcticNet and Ocean Networks Canada.

Governance and Funding

Governance involves boards and committees with representation from member institutions such as Dalhousie University, Memorial University of Newfoundland, Université Laval, and provincial research offices resembling Nova Scotia Research and Innovation Trust. Funding sources have included competitive awards from agencies like the Natural Sciences and Engineering Research Council of Canada, infrastructure investments informed by the Canada Foundation for Innovation, provincial grants, and equipment procurements with vendors such as Lenovo and Hewlett Packard Enterprise. Operational models borrow from cooperative consortia exemplified by SURAgrid and funding partnerships akin to Horizon 2020 consortia.

Performance, Scalability, and Security

Performance tuning uses benchmarking suites inspired by HPC Challenge and LINPACK while scalability strategies include elastic bursting to cloud platforms such as Amazon Web Services, Google Cloud Platform, and private clouds built on OpenStack. Security practices adopt standards comparable to ISO/IEC 27001 and identity federation protocols like SAML and OAuth 2.0, and incident response coordinates with national cyber centres similar to Canadian Centre for Cyber Security. Data governance follows principles adhered to by research data management frameworks at Launchpad-style platforms and institutional repositories like OCAD University Library and university libraries across member campuses.

Impact and Future Directions

The Grid has accelerated research outputs at partner institutions such as Dalhousie University, McGill University, and Université Laval, supporting publications in journals associated with societies like the Canadian Mathematical Society and the Geological Association of Canada. Future directions include tighter integration with federal initiatives, expanded collaboration with international projects such as PRACE and EuroHPC, adoption of accelerators from vendors like NVIDIA and AMD, and enhanced open data services following FAIR principles advocated by organizations such as the Research Data Alliance and CODATA. Continued evolution will likely involve partnerships with provincial innovation agencies and technology transfers to regional industry clusters in Halifax, Montréal, and Toronto.

Category:Research networks Category:High-performance computing in Canada