Technology Deployment Plan
On July 30th, 2015, Canada Foundation for Innovation’s (CFI) announced $30 million funding, from their Cyberinfrastructure Initiative; this investment is being managed by Compute Canada.
Compute Canada is currently leading this broad transformation of Canada’s advanced research computing platform, replacing many aging compute and storage systems with modern systems designed to meet today’s science needs. We already store approximately 15 petabytes (PB) of valuable research data and will expand this capacity to more than 60 PB.
These investments are addressing urgent and pressing needs and are setting the stage for creation of a national platform that can continuously meet the needs of research and innovation in Canada. These facilities will represent highly concentrated investments, located at four university data centres: the University of Waterloo, the University of Toronto, the University of Victoria, and Simon Fraser University.
These new systems will be larger and more capable than the sum of the systems they replace, and will provide operational cost savings and economies of scale. Local support – at all of Compute Canada’s member institutions – will remain a hallmark of Canada’s advanced research computing. Consolidation of systems will yield larger and more capable systems, and distributed support already yields better capabilities for researchers to do their work.
Compute Canada is evolving from regional areas of expertise and local teams at institutions, towards a model where the most capable personnel from across all member institutions work together as national teams. This is necessary to effectively run the consolidated systems, since they will have a broader mandate and higher service expectations.
Seamless access to data resources will be available to facilitate more complex and robust workflows, while making it easier to share data without unneeded duplication. Features include data isolation for multiple tenants of data resources, object storage access mechanisms, georeplication and backups of datasets, as well as hierarchical storage management, which will appear to most users as unlimited capacity.
Compute Canada is growing its cloud deployment, and integrating cloud services – including storage and job workflows – with larger clusters. Cloud services will host the customized virtual machines needed by some researchers.
Within the advanced research computing environment Compute Canada will provide, researchers will be able to specify the mix of resources they need – different storage pools and data access mechanisms, long-lived Web or database services, computing capabilities from a single CPU to an entire cluster, visualization and analytic services, and other components provided. Distinctions that are important today, such as whether a file is local or remote, or the characteristics of a particular high performance computing system, will not be impediments to a researcher to create the system he or she needs.