Supporting User Communities

In this posting I argue that providing support on ICT topics for research infrastructure user communities is different from supporting individual researchers or research groups. This is something that computing centers, such as CSC, need to take into account to maintain their close relation with the users.

Researchers and experts at or working with a research infrastructure form a community. The research communities span typically many organizations and countries. Often they include also research from many disciplines. Such communities are organized in the sense that they have some sort of a governance structure in place, necessary for operating a research infrastructure.

That the communities are organized have certain implications for e-Infrastructure providers as well. For example, the research communities can prioritize their thematic requirements for e-infrastructure services, which is not always the case with separate labs working on same research questions. The community has also more weight towards service providers with their demands than more scattered needs of unconnected research groups. On the other hand, the computing center can rest assured that services provided for a larger user community can reach a wide user base and very often target top level research.

Many research infrastructure user communities encompass wide variety of topics, and similarly need ICT solutions for various fields. Not only more traditional computing power and computationally efficient software, but also research data management, virtual research environments as well as high-speed networks etc. It is convenient for the community if most of the services can be reached through a single service provider. Also, as the research infrastructures reach operational phase, they start to require more and more professionally provided e-infrastructure services, implying business models, service level agreements and continuous operations.

So, in many aspects serving research infrastructure communities with e-infrastructure solutions is differing from the traditional operations of computing centers, which has been based on computing capacity in one flavor for all users, provided on “best-effort” conditions. The change includes both challenges and opportunities. The trend is not totally new, of course, and most large centers have already embraced the change. However, I am sure all best practices have not been tested yet, and improvements are possible. A good bet is to increase contacts between e-infrastructure providers and research communities as this usually reveals both bottlenecks and simple improvements.

Finally, it is good to mention that all this does not mean that from now on all users of computing or research data resources should join a user community. The traditional way of supporting individual research groups continues to be important also in the future. But instead of one-size-fits-all supercomputing a variety of  ICT environments is needed, all with their roots on the common IT platforms.

EUDAT 2nd Conference in Rome

EUDAT 2nd Conference was held last week in Rome. There were more than 200 participants present which shows that the topic of Research Data is highly relevant for many communities. The packed programme of the conference consisted of workshops, training sessions, keynotes and parallel sessions.

In its quest to develop sustainable European Collaborative Data Infrastructure EUDAT aims to work in close contact with the user communities. That is of course a natural but often in practice overlooked approach to succeed in building something that will actually be used by the community. In my opinion EUDAT is doing good job in this respect, as also exemplified by the parallel sessions on EUDAT services in the conference. In these sessions both available and new services were presented and as joined one of these sessions (on the B2Share a.k.a. Simple Store service) there was genuine interaction between the representatives of the user communities and the developers.

The keynotes speakers included high-profile names from research but also leaders from international eScience projects. iSGTW online newsletter gives good overview on these and many other presentations of the conference. From the keynotes, however, I want to highlight a presentation (slides in pdf format) given by William (Bill) Michener, Professor and Director of e-Science Initiatives for University Libraries, University of New Mexico & DataONE Principal Investigator. DataONE is a cyberinfrastructure - as they say in the U.S. - for Earth observational data which has put in place an impressive set of data services for the earth science researchers.

In summary the EUDAT project seems to be going strong and has shown its ability to roll out concrete services for research data management. Together with its links to international activities (including Research Data Alliance RDA) and the user community involvement the initiative stands in excellent position to build a sustainable data infrastructure in Europe.

Impressions from CLARIN Annual Meeting

Last week Prague hosted the second CLARIN Annual meeting since establishment of CLARIN ERIC (CLARIN = Common Language Resources and Technology Infrastructure). The annual meeting collects together representatives from CLARIN centers from member countries and from countries interested in joining in near future, as well as others involved in one way or another. Altogether there were this year more than 100 registered participants. The Finnish community FIN-CLARIN was represented by a group of four participants.

The CLARIN ERIC, by the way, is only the second ERIC established, having received this status on 29th of February 2012.

Many plenary talks were given at the annual meeting describing the status of the CLARIN infrastructure from different angles: technical, integration, services, and organizational. Additionally the event included a number of showcases and demos/posters (in a so called points of pride session) giving an overview of new developments from around the CLARIN centers in different countries.  I liked this combination of different types of presentations as it gave good overview of the research infrastructure, both its many strong sides and also the challenges. Some of the presentation slides are available at the conference web site.

What I find attractive in the CLARIN ERIC approach to running a multi-national distributed research infrastructure is that there is plenty of freedom delegated on the national level, especailly in selecting which services and resources to provide for the European users. On the other hand the European level has put on place standards and has selected technologies that the national level outputs need to match. This seems like a good balance between top-down and bottom-up approaches.

In addition CLARIN has set up an assessment process that national centers need to pass in order to be accepted as official CLARIN Centers. The second assessment round is currently ongoing, and the third round is planned for spring 2014. Receiving the status of a CLARIN Center is naturally the ambition of the Language Bank of Finland as well - the Language Bank of Finland being the common name for Finnish CLARIN services. The third assessment round seems like a suitable target, although reaching the required level by spring takes some serious work. In any case, I announced Finland’s preliminary interest for this round in the combined Clarin Center Committee and Clarin Assessment Committee meeting.

It will be interesting to compare this event to the 2nd EUDAT Conference that I am attending this week. EUDAT is an EC funded project aiming to build a Pan-European collaborative data infrastructure. Most international research infrastructures require various services for management of research data (as well as other e-infrastructure services). CLARIN is actually one of the user communities participating to the EUDAT project. These two initiatives have common ground that they are approaching from different sides, and it takes plenty of communication to find the optimal interface.

Finland to join European Research Infrastructures

International communication and collaboration has been part of scientific word for a very long time. Over the years multiple science communities have networked to great extent. More recently digitalization of science has proceeded quickly on basically all scientific areas. Together these trends have contributed to emerging need for coordinated international research infrastructures for the benefit of the research communities.

Currently, a large number of research communities in Europe are preparing to establish organizations to govern and operate research infrastructures in their field. Some communities have proceeded so far as to establish a legal entity specifically created for this purpose, commonly called as ERIC (European Research Infrastructure Consortium), and others are taking concrete steps to this direction.

Finnish researchers and research communities have been highly active in participation in designing these infrastructures. Clearly the benefits of being part of an international community is valued by the scientists themselves. Some of the benefits include access to leading scientific facilities, collections and services, for example, and on the other hand sharing the workload needed in development and operations. The RIs also facilitate the services, data collections and tools produced at Finnish universities to gain much broader audience.

The research policy-makers in Finland recognize the added value for the academia in being part of European research infrastructures. In particular the Research Infrastructure board FIRI at Academy of Finland has recommended Finland to join 6 new European RI organizations: BBMRI, EATRIS, CESSDA, CLARIN, ICOS and Infrafrontier. In one of these, the ICOS infrastructure (Integrated Carbon Observation System), the international head office will be established in Finland.

Even with the endorsement by the Academy the membership is not yet completed. Joining a RI organization actually requires the state of Finland to acquire the membership.  Consent of the Parliament is needed and thereafter the Government has to make a positive decision (in the case of ICOS head office a full parliamentary discussion will have to be passed).

Schedules for the variety of RIs may be different. There is also difference in that some of the RIs have already established governing organizations, such as CLARIN ERIC, and others are closing in on being established. In any case, many signs point to Finland having joined some of the European RIs as full members before the newly arrived snow has melted next spring.

More coordination?

Research infrastructures are acknowledged as having key importance for competitiveness and renewal of economy by enabling research breakthroughs and new innovations. Accordingly there are many ongoing national and international activities aiming to support construction, development and operation of Research infrastructures. At the same time strong need for prioritization between RIs is necessary as public funding is limited.

To tackle some of these questions nationally, a research infrastructure expert group has been established at the Academy of Finland (so called FIRI group). The FIRI group for example coordinates ongoing update of Finnish Research Infrastructure Roadmap, for which proposals were invited in early 2013.

This week the expert group organized a seminar on current state and future of RIs (presentations on Academy web site, in Finnish), discussing national RI policy as well as experiences drawn from planning and implementing certain emerging RIs. The discussion in the seminar stayed pretty well in topics of common interest (instead of lobbying for individual projects). Often mentioned topic was that how the different Research infrastructures can better collaborate and coordinate their actions. This is also a cost-efficiency question: the more RIs can share experiences, processes, platforms, tools etc., the higher impact RI funding has.

How to support this collaboration and coordination then? Simple answer is to increase discussion between people working on similar tasks at different initiatives. There are many organizational and legal issues, for example, that RIs are facing, such as joining officially an international governing body of the infrastructure; or preparing data policies safeguarding openness of research and possible requirements of funding partners.

Obviously, the commonalities extend to IT platforms as well. Most RIs require services for reliably storing, distributing and analyzing research data. Also identifying users and managing their authorizations to access the services or resources is a common need. IT solutions for such needs can often be generic. IT platforms do not make difference between bits representing particle physics information from bits representing genomics information.

Finland is relatively small country, which makes coordination achievable as long as there is willingness to share information. This is also what CSC aims to facilitate and encourage. Not only by developing excellent IT solutions, but seeking collaborations and promoting best practices. Not to mention representing Finnish stakeholders actively in international forum. In any case, regardless of structures and good intentions, collaboration is about people willing to share their knowledge and experiences with others.

What are Research Infrastructures?

In brief, Research Infrastructures can be defined as facilities, resources and related services used by the scientific community to conduct research  (see for example definition by the European Commission). Using the term research infrastructure, or shorter, RI for a facility or a resource implies also that it is accessible for wider scientific community to conduct research. Research Infrastructures also tend to be large so that it is not practical for a single department or university to fund and operate a RI without joining forces with other organizations. The users and contributors of an RI form a user community of the infrastructure.

Some common and well-known examples of major research infrastructures are CERN – European organization for Nuclear Research and ESO – European Southern Observatory. Often RIs are not centralized as the previous examples, but they are distributed and/or virtual. A distributed RI is for example a network of measurement stations – as in ICOS, which monitors greenhouse gases in the atmosphere through a wide network of stations as well as data collections. Virtual RIs are, for example, databases or archives that are used remotely from users’ workstations.

Additionally, often-used term together with Research Infrastructures is e-Infrastructure. This is used to describe IT infrastructures that very often are key components of the RIs: computing and storage platforms, networks, software and applications. There are also e-Infrastructures that are RIs in their own right. For example PRACE – the European high-performance computing infrastructure coordinates access to European top computing resources.

Establishing and operating Research Infrastructures is a large effort that requires wide collaboration and expertise from various areas from engineering to scientific topics, and from computer science to management and law. There needs to be common access policy for the RI, for example, and common understanding how the costs of the infrastructure are shared.

Modern Research Infrastructures are essential for novel and groundbreaking research aiming to solve grand challenges of society and understanding the universe around us. It is certainly worthwhile to join the effort of building these instruments of science, working together with the research communities. There are many challenges ahead – but also great rewards.