This article is more than 6 years old

Better research through better infrastructure

A new report from the Royal Society showcases the range and value of UK research infrastructures. David Kernohan takes us through the issues, and highlights the need for investment in data infrastructure.
This article is more than 6 years old

David Kernohan is Deputy Editor of Wonkhe

Research is an expensive business. In many disciplines, being on the cutting edge requires access to world-leading technology, resources and facilities, with local access to everything a researcher might need being a dream that no single institution can afford.

There are an estimated 400 research infrastructures – defined as facilities, resources and services used by researchers who are not located at the infrastructure or responsible for maintenance – based in the UK, according to a report from the Royal Society. These range from the huge pieces of equipment like the Joint European Torus to enormous datasets such as those held by EMBL-EBI, and also encompass collections, archives and other electronic resources. Infrastructures can be found on a single site or distributed over multiple locations.

It’s a part of the sector that your average wonk, with an interest in the work of a particular institution and the way it is affected by policy, may not be aware of. Active researchers will perhaps have one or two that loom large in their lives. But these things are everywhere. Even the means by which you are most like reading Wonkhe today – the Janet network managed by Jisc – started off as, and remains, a key UK research infrastructure.

Who funds them?

The Royal Society has surveyed 135 research infrastructures (chosen because they were in receipt of at least some government funding) in order to understand better how they are commissioned, used, managed and funded. With such a wide definition, the survey includes everything from resources managed by a single institution and made available to others, up to huge pan-European projects involving a wide range of funders and participants.

The government – of course – is a key infrastructure funder, via the work of the Science and Technologies Funding Council (STFC) and other research councils. But the European Union has funded 84% of those facilities surveyed – 31% describe EU funding as “essential” to their ongoing sustainability, and a further 46% deem it “important” to their continued well-being. This is a particular issue for social sciences and humanities infrastructure – as a greater proportion of research grant funding in these disciplines comes from EU bodies.

Some infrastructure directly charges researchers and projects for access. In other cases, the cost is borne by grant funding from government, charities or business. The former is perhaps more sustainable, but may run the risk of pricing out research that could usefully benefit from access.

But there is also a role for individual institutions and consortia, both in funding and managing infrastructure. Federation, in particular, is an interesting model for electronic resources – a recent report by the Knowledge Exchange collaboration examined leading examples of the way such infrastructure is managed by consortia Europe-wide.

Who uses them?

Supporting researchers around the world (69% of UK-based infrastructure is international in scope) in gaining access to this stuff is no small task. Infrastructures have a lifecycle spanning from planning through construction, maintenance and decommissioning/archival and there is significant work to be done at each stage.

Over 4.3 million users access a sample of 87 research infrastructures annually. As above, most infrastructure-supported research is undertaken by a combination of internal and external researchers (40%), with the external use of a provided service, collection or database another key route. Only a little more than half of users come from the UK.

Infrastructures are just as international as research itself is – 77% of survey respondents reported they were a member of one or more international research project. This global outlook is reflected in the staff who work at these facilities – 32% are from overseas, 23% from elsewhere in the EU/EAA.

Some facilities have their own in-house researchers (on average permanent researchers are 35% of staff, temporary researchers are 27%) working both on their own projects and supporting the research of others. Technical staff are essential to the smooth operation of such services. On average 25% of staffing linked to a resource is permanent technical support, with a further 14% brought in on shorter-term contracts.

These are, of course, not just academic research resources: 93% of the sample reported some portion of research conducted with the commercial sector – for 14% of these this makes up more than 80% of their work.

What infrastructure do we have?

The bulk of existing infrastructure supports research in physical and engineering sciences, or in biological sciences – with much of the remaining third split between earth science, social sciences and humanities, and energy science.

The Royal Society report offers range of case studies which highlight particular infrastructures – but there are many more. Examples widely used in the UK include:

  • EMBL-EBI – this is a UK-based node – though not the UK node – in a Europe-wide project (ELIXIR) allowing for the sharing and reuse of experimental data in the life sciences.
  • VADS – a meta-resource, acting as a directory of collections of art and design resources held by individual institutions.
  • Meteorological Office DataPoint – a collection of real-time data feeds derived from the work of the Met Office in monitoring and predicting weather and climate.
  • Historical Texts – a Jisc-supported meta-resource, covering a range of digitised printed material across multiple disciplines.
  • ADRN – an access brokerage service supporting carefully monitored research access to detailed public sector datasets.
  • The Harwell campus – this is where much of the UK’s (and the world’s) serious research hardware lives – huge physics facilities like the Diamond Light Source and the ISIS pulsed neutron and muon source sit alongside a huge high performance computing facility and shared laboratories for medical research.
  • Zenodo – part of a EU Horizon 2020 project (OpenAire, which was part-funded by RCUK), allows researchers from any discipline to share their own research data and discover data from elsewhere.
  • The RRS Sir David Attenborough – a polar research vessel operated by the British Antarctic Survey, which will be in use from 2019.

Data: the future of research infrastructure

The inexorable rise of data driven methods, and the parallel rise of open research practice, mean that accessing and sharing huge amounts of data is inevitably going to be a major part of research in the future. In one sense this is a huge opportunity to lower the cost and raise the quality of research – more accessible data means that much more can be learned from a single experiment, and the ready availability of data from peers around the world means that findings can be cross-checked and replicated without having to generate new results.

More and more historical source materials are being digitised and shared through global and regional initiatives – more archives are emerging from library stacks and storage boxes to online databases and image galleries.

But resource storage and archival is a huge expense – both in terms of the raw cost of many terabytes of server and hard-disk space, and the expense of maintaining and updating records to aid discovery (there’ll be a continued marketing and awareness cost too). Current infrastructure provision is piecemeal and variable by discipline. Research funding generally comes as a time-limited support for a specific project – and there’s a very difficult point of inflexion between project funding and a sustainable resource. Difficult curatorial decisions need to be made – as great as it would be to keep every part of research infrastructure and every byte of data to aid replication, the expense of doing so would be incredible.

A Universities UK-led group is currently preparing a report on support for infrastructure linked to data, which will make a recommendation to the minister later this year. Its initial report examines the nature, scope and maintenance of a data-specific subsection of research infrastructure.

Investment in research is often seen as being a matter of channelling project funding to academics and institutions. But the infrastructures that support research – and preserve resources – are the essential underpinnings of these projects, and sustained funding to support these into the future is a lot harder to find. Brexit, with the likely loss of at least some funding access, will greatly increase these pressures.

Leave a Reply