Restore Confidence in Compute Canada
The following is an open letter to the Member Institutions and Board of Compute Canada to express the concern of HPC stakeholders in Canada regarding recent events and actions by the board (such as the firing of the CEO) and the negative effects these are having on HPC and research competitiveness in Canada. The letter will be delivered to the board with names and institutions (but not email) of signatories. If you have questions about recent events, we recommend that you contact your regional computing representatives. Questions submitted here will be answered as time permits.
Dear Members and Board of Compute Canada,
As members of the Canadian research community who are critically dependent on high- performance computing (HPC), we are writing to express our dismay and disappointment at the current state of Compute Canada. Access to reliable, sustained and predictable HPC infrastructure is central to global competitiveness in a wide range of disciplines - it is Compute Canada's responsibility to ensure that this environment is maintained and enhanced. This responsibility includes: establishing Compute Canada as a credible authority with governments, funders and other digital infrastructure organizations and ensuring that the provision of service to our community from the regional HPC organizations is not disrupted. Compute Canada has now been in a state of flux for over a year and we are not confident that recent actions have done anything to help achieve these goals and, indeed, feel that they have left the organization in a perilous state. It is critically important that Compute Canada moves forward in a way that will re-establish the confidence of all stakeholders.
We have the following specific concerns and suggestions:
* The board has shown reluctance to engage the community in a meaningful dialogue and to create a transparent and accountable organization. Not only are key board minutes and resolutions, which might beneficially be made available to the formal members and community, kept secret but the board has made clear that strategy will be developed by the board with little opportunity for the community to guide the process in a meaningful way. Significant input had been provided to the former CEO during his tour of major sites in Canada. His termination was not only alarming but it removed someone with understanding of research needs and leaves uncertain the fate of our input to the strategic planning process. It is the norm both in Canada and elsewhere that the research community is directly involved in planning and strategy because of the way in which researchers use advanced computing infrastructure. We feel that the to- be-established National Advisory Council with membership recommended by the key community disciplines is an appropriate vehicle to drive or have significant input to the planning process. This council should be established immediately.
* The termination of the CEO appears to have been undertaken without a clear interim plan for managing Compute Canada before a new CEO is hired. Instead, the board, with very limited time or expertise in HPC or research is itself assuming the executive functions of Compute Canada. This confusion of roles, evident even while the CEO was in place, is concerning. An interim solution would appear to be to engage with and leverage the expertise of the regional organizations to ensure that the ongoing business of Compute Canada does not suffer.
* The insistence of the board in dealing directly with the institutions has delayed the flow of critical operating funds to the regions to a truly alarming degree. This sort of brinkmanship is clearly not an effective way to run a critical resource for research in Canada, more importantly it has created undue and unnecessary stress on a highly qualified staff built up over the past decade and more.
* We understand the need for - and welcome - the evolution of Compute Canada: this can greatly strengthen the chance of achieving the goals outlined above. It is simply not sensible, however, to ignore the significant successes that the regions have already achieved to collaborate and inter-operate in many areas. Further, the regions are serving the ongoing, on-the-ground research needs of several thousand users: any new direction must be implemented without disrupting this critical activity. This will require a willingness on the part of the board to engage in detailed and meaningful discussions with the regional organizations and to recognize the equal role and interest the provinces have in the provision of HPC resources.
The research community is the key stakeholder in Compute Canada. We are dependent on HPC resources to carry out our research and, because of the strong desire of funding agencies to focus all research computing beyond the desktop into a single platform, Compute Canada must be a trusted and responsive organization. For many of us, particularly those who are more junior, our careers depend on access to these resources. For Compute Canada to succeed in enabling forefront research in Canada, we believe that it is essential that the board undertake to engage in an open dialogue with the community and regional organizations in order to rebuild credibility in the organization, to manage Compute Canada effectively in the short term and to engage in a supportable and realistic strategic planning process.
This Letter in PDF Format