Tim: Hello and welcome to the Computerworld Australia live podcasts. I’m Tim Lohman, editor of Computerworld Australia. Today, our guest is Peter Czeti, manager of infrastructure and client services at the CSIRO. Hi Peter.
Peter: G’day
Tim: Thanks again for joining us. You’re due to present at Cebit’s Future Proofing Your Data Centre Conference in Sydney on February 14. But before we hear a bit about that, can you perhaps give us some background on the CSIRO’s Data Centre environment?
Peter: CSIRO, as you can appreciate, has a very broad range of people that it participates with. So much of our data centre picture is actually a shared data centre picture because, of course, CSIRO being a research organisation participates in a range of federal and state driven research programs, and of course states like to have a share of what we do in the data centre space. So we share data centres across the country.
In Western Australia we have Ivec, in Tasmania we have Tepaz, in Sydney NSW we have Intersect. So all of our data centre considerations are a little bit muddled up with the fact that we have to deal with our partners and sometimes those needs don’t converge.
Tim: Okay, so how do you go about managing that?
Peter: A lot of them are tied around the actual research programs themselves. So in the case of the Queensland facility, CSIRO participated in the building of and the funding of that facility.
We manage it largely, but of course any changes that we make we need to consult with our partners. Nine times out of ten the changes are common sense because they’re going to be around capacities such as airconditioning and servers and those sorts of environmentals. The more difficult part is around the science application needs and sometimes they can diverge a little.
Tim: Can you give us a bit of a flavour of some of the IT initiatives underway at the CSIRO and how they may affect the way that you manage your data centre environment?
Peter: Yeah, look, it’s probably a little bit difficult to describe without pictures, of course, but as you can appreciate the way that science is undertaken now has changed dramatically over the course of the last 20 years. Particularly in the last five to seven years, where data has just become so much part of ... even the papers that a lot of the scientists publish. So if you publish a science paper in nature, or something like that, you’re actually expected to also provide data so that people can independently review and verify and comment on the data set that you’ve used to produce the publication.
Now, because of that need to be able to have that ... data available and accessible to a number of parties, we find that our biggest challenge is being able to provide enough accessibility to data across the community, so that people from different organisations can access it.
We have some initiatives under way. They’re national initiatives, not specifically CSIRO. Like the Australian Access Federation which tries to basically provide a group of well understood credentials across the science community that enable people to come in, look at data and manipulate it using the computer clusters that we have. But that’s still very much in its early stages, so for us the big challenge is making science applications and data available.
Tim: Your presentation at the Data Centre’s Conference focuses on undertaking our three- to four-year Data Centre Consolidations Program. Can you talk us through drafting that plan and some of the considerations in that process?