"There is a growing mountain of research. But there is increased evidence that we are being bogged down today as specialization extends. The investigator is staggered by the findings and conclusions of thousands of other workers-conclusions which he cannot find time to grasp, much less to remember, as they appear. Yet specialization becomes increasingly necessary for progress, and the effort to bridge between disciplines is correspondingly superficial."
- Vannevar Bush, 'As We May Think', published in The Atlantic, 1945.
Echoes of Bush's landmark essay 'As We May Think' seemed to resonate across many strands of eResearch NZ 2014 and the essay was actually quoted in the very last comment of the three-day conference's wrap-up session. Bush's post-war concern that the expansion of research data and scientific outputs was far outstripping our means to comprehend them was to be remedied, it was argued, by working towards the development of devices for aggregating, sorting, storing, and linking data and ideas. And though we have now realized in form a lot of Bush's visions, we also appear to have maintained, if not increased, the initial problems: mountains of data, more research than we can comprehend, and low levels of interdisciplinarity.
Many of the problems originally identified by Bush were engaged with at eResearch NZ 2014, which took place at Waikato University in Hamilton, New Zealand, from 30 June to 2 July. This was the fifth year of the conference, which once again brought together a wide range of researchers and high-performance computing (HPC) experts from across New Zealand, as well as from further afield. Consisting of over 25 workshops, more than 40 presentations, and multiple panel and open discussion sessions, the symposium was a national forum focusing on themes of real-world developments and collaborations across the sector; data science; and the need for new skills and career paths around advanced computing for research.
The conference was opened by Dave Turek, the vice president of exascale systems at IBM with responsibility for IBM's overall HPC strategy. Turek gave a presentation on the evolution of HPC and the design of HPC systems for a data-centric world. His brief history of HPC began at a time when power consumption and space weren't critical factors for design considerations, but in our current climate the financial and time implications of big data movement and storage are significantly outstripping compute needs. His warning, "divorce big data and analytics from HPC at your own peril", brought HPC into the current climate where big data is becoming a significant problem for HPC maintainers, rather than just a buzzword for marketers, politicians, and funding bodies. Turek also argued that the Top500 list doesn't tell us anything about how HPC performs against big data. He suggested instead that the real value over time would be in software exploitation rather than hardware, and that costs were already rapidly shifting in this direction.
Day two of the conference was opened by a keynote from Bryan Lawrence, professor of weather and climate computing at the University of Reading in the UK. Lawrence presented on the infrastructure challenges of performing climate science on a global scale and the implications of tackling this grand problem via cross-border and cross-disciplinary collaborations. He reiterated the opening keynote speaker's concerns about the "significant, looming data problem"; the hundreds of petabytes of data shared across the global climate community is taxing our current storage capacities, creating a "paradigm shift that has crept up on us". He argued for the development of more partnerships and e-infrastructures to be forged to address this quickly growing problem, as well as more investment in the analysis tools and automation needed to manage this type of data.
The New Zealand eScience Infrastructure (NeSI) was a gold sponsor of eResearch NZ 2014. NeSI director Nick Jones also spoke on day two about growth and development of future capabilities in HPC and e-research for the New Zealand research sector. New Zealand is quite new to having the HPC capabilities provided by NeSI as part of its research infrastructure, but the success of NeSI as an e-infrastructure and the investment that it has placed in people as much as platforms has been well received by researchers. Jones spoke of the desire at NeSI to "be able to grow people so that they can look after themselves, or address their development needs so they have the opportunities to approach more challenging problems." NeSI's support of the communities of research in New Zealand as they roadmap their future research plans was seen as key to this strategy of growth and development.
The tone for day three was set by a keynote from Kaitlin Thaney, director of the Mozilla Science Lab. Her talk, entitled 'Making the web work for science', addressed how openness and access to content, data, tools, and technology, could help researchers address some of the current bottlenecks experienced in science and research. Her push for "open research" was an appeal towards using the power of the web to influence the culture of science by demonstrating new and open ways of conducting research. Thaney spoke to me after her talk about her belief in the importance of facilitating a culture change for early career researchers, so as to enable them do more with their data and to do so in a more efficient and collaborative fashion. The work of Mozilla towards enhancing the social infrastructures of research communities and providing frameworks for the skills development of early-career researchers provided a good model for how to facilitate and scale communication and coordination across the research sector within New Zealand.
The conference closed with a spirited open discussion on all that had gone on over the past few days. The challenges and barriers to contemporary research were acknowledged, as was the equally amazing work that had been showcased by a range of speakers over the past days. The world of science and research was identified as one of great complexity and scale - this is the norm; but in reflecting back on Bush's essay, it was agreed that the encouragement of innovation, the development of emergent skill sets, and increased collaboration across a range of disciplines and institutions would yield the best results for society.
Read more about the eResearch NZ 2014 event in our recent interview with NeSI director Nick Jones: 'Supporting research with HPC in New Zealand'.