CUCCIO is an organization of university of CIOs. I was invited as a "token researcher" to talk at the Digital Infrastructure Summit 2012.

Jonathan Schaeffer Opening

Schaeffer opened the summit talking about the challenges faced by researchers. Administrators like to complain about researchers who get a grant and then talk to you, but we have to understand what they face - researchers have to talk to all sorts of people to put a project together. Can we find a way to make it easier for researchers. Can those administer research infrastructure understand and support researchers?

John Hepburn: The Perspective of a VP Research

Hepburn talked about the perspective on infrastructure from the administration. He quoted a truism "If brute force isn't working you aren't using enough of it." We are limited by:

Budgets
Power lines
Cooling

Computers are not the only devices getting faster and bigger. We can collect more and more data now. The data explosion is mirrored in everything. We need computer infrastructure for everything we do. When we buy it, it is almost instantly obsolete. Data centres are becoming a problem because of heat and electricity.

We need the stuff, it is expensive to maintain, there should be a national strategy. But there isn't. We have a bunch of competing interests. We don't have a comprehensive system for any large scientific infrastructure.

I can't help asking if we really need the big stuff?

He gave examples of successful projects like TRIUMPH and the Canadian Light Source project. The key to the Light Source project was that it was developed by researchers. All the researchers, once it became clear that the Saskatchewan proposal was the best one, got behind it.

He made some comments about the lack

As for digital infrastructure we have a long way to go. We don't have a national consensus. There is a need for a cultural transition. People will only get behind centrally controlled facilities when they have to - when they can't do research on their own or in smaller projects.

When people can run their own stuff they do so and the culture becomes "if I can't see the computer and touch it, I'm not happy." The workstation and linux box became important. We haven't made the transition to thinking together about digital infrastructure.

When CFI forced the HPC folk together to create Compute Canada they wanted us to create a national system. Alas what was created was a system of regional projects. The challenge is whether we can produce a national system? What are the challenges to getting people to buy into a national system:

It can't benefit only some folk
It should allow individual researchers to be able to invest in
You don't need to be near the infrastructure to use it
Who would pay for it in this time of cuts
Why bother when there are industry solutions for big data - why not contract out?

What do we have to do:

We have to define what is national digital infrastructure. We need to define what is national for those things that are too expensive for individuals.
The users/researchers have to determine what they need. Perhaps the users don't need petaflop machines. Users should be involved in the defining process.
We need someone central who would embrace this and be identified as the person responsible.

We are behind some other countries, but not that far behind.

Thursday, June 14th

Reality Check Panel

I spoke at the Reality Check Panel on �Supporting Humanists with Digital Infrastructure" along with Michel Dumontier. Rick Blunt moderated. For this talk I prepared a set of questions for auditing whether a university has good support for digital research in the humanities. See Check IT Out!.

Ann Borde: The Road to Integration

Dr. Borde talked about eResearch in Australia. eScience (UK), cyberinfrastructure (US), and eInfrastructure (EU) and eResearch in Australia all have similar features of transcending borders. She talked about the Fourth Paradigm: Data-Intensive Scientific Discovery which is seeing changes:

Observational/experimental
Theoretical/analytical
Computational

Researchers are extreme information workers. "Data sets are the new instruments of science."

She showed a very interesting slide about research patterns. How do different disciplines disseminate data sets.

She talked about the stack of service layers:

Digital Labs (Virtual Research Labs)
Research Data Commons
Shared Access Methods (networking, better HPC modelling, security)

The digital labs face out to the researchers in a particular problem area. The data can be found and used. This is supported by underlying infrastructure.

NeCTAR (National eResearch Collaboration Tools and Resources) is the latest project. See http://www.nectar.org.au . The idea is to develop tools and virtual labs with the research community.

HuNI is the Humanities Networked Infrastructure project, see http://www.huni.net.au .

She talked about eResearch centres or Virtual Research Environment (VRE) - how they can help coordinate needs across institutions for a research community.

Some of her closing thoughts:

Every interaction crosses boundaries
What is the role of service providers
What new skills are needed?
We need to understand the patterns of use
Researchers like to have their own stuff

In the UK there was a good study of uptake called eUptake that showed that training is often patchy.

Scott Tomlinson: The IPY (International Polar Year) programme

The goal of the International Polar YearIPY was to gather valuable research about the North Pole. Before IPY a lot of the data was in boxes under desks. A lot of what is out there is for specific disciplines and is hard to use outside of a discipline.

IPY developed a data policy and invested in a Polar Data Catalogue. They invested in a Canadian IPY publications database.

There are a lot of issues around access and preservation. They created a network of data assembly centres. They issued a call for proposals. They brought six centres together and explained their vision them and got the centres to work together.

Technology is the easy part, the socio-cultural part is difficult. A lot of time is spent coaching researchers.

Some of the lessons learned include:

Get early agreement on standards
Have a secretary to coordinate
Rationalization and harmonization of overlapping policies

David Barnard: A President's Perspective

Dr. Barnard gave us the perspective of a university president. Universities used to be leaders. Now we try to keep up with industry and with our students. While infrastructure is supposed to support research it should also keep the public good in mind. We need to keep the larger budget issues in mind. Cyberinfrastructure will be expensive, how can we justify substantial investments in this time of restraint. We need to develop not just a sense of what to do, but a model and case for convincing people who represent the public. We should seek the common good. We can't make special claims to universities. There are lots of lots of special needs that need to be synthesized.

From the president's perspective the dilemma is that there are too many good claims. How to optimize among the good ideas. How do governments and presidents optimize. We don't get the collective good by sticking to what we want ourselves. Compromise and standardizing are ways to optimize. This is a political issue.

How do we move forward? Barnard suggested we may need to be unified, compelling, have a vision, and perhaps go somewhere different.

Robert Cook: An Overview of Progress Toward an Integrated Digital Solution

Book Cook then surveyed the situation and discussed the documents we had.

We were given an hour to develop a vision statement and four principles to achieve that vision. My group came up with the following vision:

Enabling communities of understanding to create, access, and utilize shared digital information.

I was thrilled to see that important to the group was inclusiveness and the preservation of data. Alas we got stuck trying to edit the vision in a room of 75 people. The resulting vision was something like:

Canada has a digital ecosystem that accelerates leading research and innovation to benefit society.

Steven Liss: Leadership Council for Digital Infrastructure Integration

Liss talked about the creating a Leadership Council for Digital Infrastructure Integration. The idea is that we need a council to provide leadership, but that this should put researchers first. There was a certain amount of discussion about how to involve researchers. Who should lead? Should it be the researchers or the other stakeholders (SSHRC, NSERC, CFI, CARL, CUCCIO ...)

I wondered about the idea of a couple of pilot projects. I can imagine a series of audits where you take a real research challenge in different disciplines and follow them through.

Jonathan Schaeffer

Jonathan had the last session talking frankly and calling for action. He commented on how a lot of this is "deja vu all over again." We have had reports on the need for cyberinfrastructure for a while. What has changed? Jonathan mentioned CUCCIO is a change - a real group that can get things going. What can be done? Much of what was suggested seemed to me to be calls for champions, calls for government leadership and not calls for action. My sense is that this group (CUCCIO) could commit to actually doing something - a prototype of some sort. Jonathan ended by suggesting that getting researchers to sign on to the vision and some principles would be a step forward.

CUCCIO 2012