When the right people find each other, like one another, and agree on the issue at hand and how to tackle it, it is in the interest of all involved to join forces and go.
In 2002, just such a group of researchers converged around one goal: accessible high-performance computing (HPC) for scientific research in Atlantic Canada.
Since its founding in 2004, ACENET has grown from five universities to include 15 institutional members in the region. What began as a service for "hard science" fields such as physics, chemistry, astronomy and oceanography, has expanded to encompass humanities, social sciences, and even business. It has served over 4,000 clients and trained nearly 14,000 people, at no charge to researchers and students.
All this was achieved with an average of 15 people on staff since ACENET’s first hire in August 2004.
A clutter of clusters
That grand dream began as a humble necessity.
"We literally had it stuffed in the closet," recalls Ross Dickson, Lead Research Consultant at ACENET, referring to a computer cluster in the computational chemistry lab from his days at Queen’s University. The makeshift nature of that system was common in the late ‘90s and early 2000s. Scientists across disciplines resorted to building these stacks of consumer-level hardware as they grappled with demand for computing power, lacking funding, infrastructure, and real estate to house it.
Though cheap, these setups weren’t pleasant to have around. They had to be put somewhere, under a desk or in a closet, making the rooms hot and noisy. “You need dedicated space for it because you can’t sit in the same room with one hundred PCs,” says Peter Poole, a physics professor at St. Francis Xavier University and a member of ACENET’s founding group. The question became, why should all these labs be putting themselves through the same ordeal? A few years earlier, the Canada Foundation for Innovation had been created to invest in research infrastructure, but only for institutions. As a result, it made increasingly more sense for researchers to think bigger than their individual labs.
“So out of all that context, a few of us got together starting in 2002 to go down the road that created ACENET,” concluded Poole, citing the hard work of many people involved, including Mark Whitmore of Memorial University, Virendra Bhavsar of the University of New Brunswick and David Clark of Saint-Mary's University.
A culture of teamwork
That first successful team effort set the tone for the next two decades.
Early on, a culture had settled in of teamwork and mutual support among people motivated to get the job done. “We knew we needed to protect it,” says Greg Lukeman, CEO.
Many of the current staff have been with ACENET for 10 years or more. “There's this idea floating around,” says Dickson, "that technical professionals ought to change jobs frequently.” But ACENET seems to have found a recipe for long-term commitment. “It’s the people,” says Phil Romkey, Lead HPC Systems Administrator, explaining why he has stuck around since 2006. “We’ve got really good people.”
That culture has ensured that ACENET remains a steady presence within the evolving ecosystem of advanced research computing hubs in the country, first under the umbrella of Compute Canada, and then the Digital Research Alliance of Canada.
It has also made the organization a major contributor to national-level teams, infrastructure and services. For instance, the national partner’s help desk and customer satisfaction surveys were modelled after ACENET’s. The team that developed and deployed the national LDAP (lightweight directory access protocol) was led by ACENET systems administrator Craig Squires. Karl Vollmer, another long-time systems administrator, brought in the national provisioning system under Compute Canada.
As part of the 2017 Leadership Council for Digital Research Infrastructure, ACENET helped revise the federal funding model, pushed for new funding, resulting in the integration of previously separate functions of Digital Research Infrastructure under the Alliance.
“We may be small, but we punch above our weight,” writes Michele Fash, Director of Business Development.
The consortium’s focus on cooperation and exchange has been felt at a regional level as well by setting a precedent for pan-Atlantic cooperation. “We are typically smaller institutions and we’re able to amplify each voice by working together rather than continually competing for resources,” says Tana Allen, acting Vice-President (Research) at Memorial University and ACENET Board of Directors Chair.
Adapting to change
ACENET also broke ground when it became the first consortium to establish a cluster focused on helping industry. ACENET’s original five systems were set to be decommissioned in 2019. “We realized that we still wanted to have a presence of infrastructure in Atlantic Canada,” says Lukeman. With funding from the Atlantic Canada Opportunities Agency (ACOA), ACENET built the Siku cluster at Memorial University in 2019 with a new industrial mandate.
This was an opportunity to bring value into the region outside the academic space, with the co-benefit of supporting ties between industry and academia. “Sometimes we’re able to create those relationships because we’re seeing the skills and offerings on both sides,” says Lukeman. Industry concerns can also be different, such as cybersecurity requirements and more user-friendly interfaces, pushing ACENET to optimize and improve its services overall.
Empowering research
Though teamwork is the foundation of ACENET, its success also stems from the staff’s dedication to helping researchers. “It's not just a job,” says Lukeman. “It's sort of a vocation.”
“The great joy of this for me has been going to new research groups and new disciplines and figuring out their language,” says Dickson.
The Research Support team, led by Dickson, assists researchers in translating their specific jargon to the digital sphere, whether it’s for biochemists, engineers, librarians or any of the many and varied disciplines currently served by ACENET.
Romkey’s Infrastructure team ensures that the machines can handle the demands of hundreds of simultaneous users. But he also enjoys working directly with researchers when he can, whether to navigate the purchasing of equipment or inputting projects into complex systems.
The consortium began with the dream of making HPC accessible and free to any researcher. The organization has shown that it can reach the loftiest goals with clear vision, pooled resources, the right people, and an openness to reinvention. “ACENET is responsive,” says Allen, referring to the evolving nature of computing technologies. “I have very good feelings about the next 20 years.”
