Interim Evaluation of the European Research Infrastructures including e- Infrastructures in Horizon Report of the Expert Group

Transcription

1 Interim Evaluation of the European Research Infrastructures including e- Infrastructures in Horizon 2020 Report of the Expert Group September 2017

2 Interim Evaluation of the European Research Infrastructures including e-infrastructures in Horizon 2020 European Commission Directorate-General for Research and Innovation Directorate B Open innovation and Open Science Unit B4 Research infrastructures Contact Laura Esposito laura.esposito@ec.europa.eu RTD-PUBLICATIONS@ec.europa.eu European Commission B-1049 Brussels Manuscript completed in September This document has been prepared for the European Commission however it reflects the views only of the authors, and the Commission cannot be held responsible for any use which may be made of the information contained therein. More information on the European Union is available on the internet ( Luxembourg: Publications Office of the European Union, 2017 PDF ISBN doi: /63168 KI EN-N European Union, Reuse is authorised provided the source is acknowledged. The reuse policy of European Commission documents is regulated by Decision 2011/833/EU (OJ L 330, , p. 39). For any use or reproduction of photos or other material that is not under the EU copyright, permission must be sought directly from the copyright holders.

3 EUROPEAN COMMISSION Interim Evaluation of the European Research Infrastructures including e- Infrastructures in Horizon 2020 Edited by the Expert group members: Hans CHANG Alfred GAME (Rapporteur) Francesco PROFUMO (Chair) Marie-Noëlle SEMERIA Milena ŽIC-FUCHS Directorate-General for Research and Innovation 2017 European Research Infrastructures including e-infrastructures

4 Table of Contents 1. EXECUTIVE SUMMARY How relevant has the H2020 RI thematic programme been so far? How effective has the H2020 RI thematic programme been so far? How efficient has the H2020 RI thematic programme been so far? How coherent has the H2020 RI thematic programme been internally and with other (EU) actions? What is the EU added value of the H2020 RI thematic programme? lessons learned for the future INTRODUCTION Context Objectives and intervention logic Developing the European research infrastructures for 2020 and beyond Developing new world-class research infrastructures Integrating and opening existing national research infrastructures of pan-european interest Development, deployment and operation of ICT-based e-infrastructures Fostering the innovation potential of research infrastructures and their human capital Exploiting the innovation potential of research infrastructures Strengthening the human capital of research infrastructures Reinforcing European research infrastructure policy and international cooperation Reinforcing European policy for research infrastructures Facilitate strategic international cooperation IMPLEMENTATION STATE OF PLAY Overview of programme inputs and activities Participation Patterns Cross-cutting issues RELEVANCE: Is H2020 RI tackling the right issues? Flexibility to adapt to new scientific and socio-economic developments EFFECTIVENESS Short-term outputs from the programme The effectiveness with which European research infrastructures have been implemented The importance of ESFRI The effectiveness of INFRAIA: the role of speculative science and the Open Science agenda EINFRA and the future of the Digital Open Market PROGRESS TOWARDS THE OVERALL HORIZON 2020 OBJECTIVES Contribution to the achievement and functioning of the ERA EFFICIENCY Budgetary Resources Programme's Attractiveness Continuity and prioritisation Other Issues Related to Efficiency Sustainability

5 Fragmentation and Convergence Sustainability of Data-led, Distributed and Virtual Infrastructures Suitability of the financial instruments COHERENCE Internal coherence Internal Coherence of the actions implemented External Coherence EU ADDED VALUE Design and development of large European and global infrastructures Widening access and improved integration Policy development and international engagement The contribution to mobility The importance of the data pilot and EOSC Examples of added value from H2020 RI Square Kilometre Array: Infrastructure Detailed Design for SKA Phase CREMLIN: Connecting Russian and European Measures for Large-scale Research Infrastructures EOSC - The European Open Science Cloud Pilot action SUCCESS STORIES FROM FP BBMRI QUALITYNANO OpenAIREplus - making reality of the Open Access vision for science LESSONS LEARNT/CONCLUSIONS Relevance Effectiveness Efficiency Coherence EU added value RECOMMENDATIONS ON THE RI PROGRAMME Annex A Annex B

6 1. EXECUTIVE SUMMARY Research infrastructures are facilities, resources and services that are used by the research communities to conduct research and foster innovation in their fields. They may be used beyond research, e.g. for education or public services. The H2020 RI programme sets out to use the accumulated talent and resources of science capability across the EU to develop new advanced research infrastructures, widen the access to existing ones, network and interoperate national ones of wider interest and work together with third countries on global infrastructures and to address global problems. To do this H2020 RI also develops policy, provides training, and fosters workshops, studies and meetings to hone ideas and find the best ways of delivering and running infrastructures effectively. The programme title is H2020 European research infrastructures (including e- infrastructures) because the programme includes infrastructure to provide the scientific community with access to advanced and networked computing and many other services. These in turn support the rapidly-growing number of data-driven research infrastructures which are enabling the sharing and comparative study of information and research results across Europe and between disciplines. The computing and data infrastructures are together forming the basis of the European Digital Open Market, a key element in plans to build prosperity, jobs and well-being across Europe. Most of H2020 RI is managed by DG RTD. The e-infrastructures component is managed by DG CNECT, where it sits alongside officials responsible for the wider aspects of the digital economy, and the two teams work closely together. H2020 RI is a direct successor to previous Framework programmes and represents the latest component of a long-term effort to realise an effective and developing research infrastructure provision for Europe. The Commission is helped in doing this by two groups: ESFRI (European Strategy Forum for Research Infrastructures) and e-irg (e- Infrastructures Reflection Group). These groups play a major role in connecting the programme to the scientific communities and, in the case of ESFRI, to Research Ministers in member states How relevant has the H2020 RI thematic programme been so far? The five sub-programmes of H2020 RI present a cohesive set of interventions that address the objectives of the programme comprehensively, although the part to promote innovation is relatively new and needs further development. They are the result of a continuous refinement of successive programmes under previous Frameworks and work together well. This continuity and long-term vision must be retained. The programme has been delivered in a period of unprecedentedly rapid change in science and technology, particularly as it relates to IT and data. The infrastructures and resource being developed are critical to the success of the Juncker Commission s New Vision for Europe 1 and the Open Innovation/Open Science/Open to the World Moedas vision 2, because they are delivering platforms, tools and skills that will enable it to happen political priorities of the Juncker Commission 2 Open innovation, open science, open to the world: A vision for Europe 4

7 The EU has developed global leadership in Open Data policy because of policy work developed in FP7 and H2020 RI since The continuity and consistency of vision in the programme, its strong advisory structures and its ability to exploit budgetary flexibility has allowed agile response to emerging opportunity e.g. in the launch of the European Open Science Cloud (EOSC) pilot in How effective has the H2020 RI thematic programme been so far? The ESFRI Roadmapping process, coupled with the use of the INFRADEV part of the programme to catalyse the design, development and prioritisation of RI proposals, and the partnership this represents between the scientific community, member state ministers and funders, and the Commission, is one of the major achievements of international science policy in recent years and must be maintained. The pioneering development of distributed European infrastructures and networked infrastructures based around the shared distribution and access to data, materials and tools has been transformative, stimulated scientific communities across Europe into cooperation and created a solid basis for EU-level research. The two roles of the INFRAIA element of the programme to develop starter communities and to enable advanced ones to address major issues are both valuable, but the share between the two could be re-assessed so as to optimise the catalytic value of H2020 RI on new opportunities. There should be something similar to attract and explore new ideas from the community in EINFRA. The EINFRA programme has had to operate in an extraordinarily challenging and changing environment and has done well to address the objectives it was set, to maintain a vision and coherence across its programme and to absorb substantial changes such as EOSC. It sits in the right place within DG CNECT with close connections to DG RTD How efficient has the H2020 RI thematic programme been so far? The H2020 RI programme is well managed by Commission officials and the two teams in DG RTD and DG CNECT achieve a good level of overall coherence despite the considerable challenges involved. The Programme Committee performs an effective and appropriate oversight of the programme. It is key to the effectiveness of the programme going forward that all those involved in its management retain a sense of the need to prioritise and, accepting that to do this within a constrained budget involves a conscious decision to create opportunity for the new by discontinuing support for the longer-established. The EC should be very wary of accepting long-term responsibility for the operational funding of mature infrastructures, and should continue efforts to find means to obtain commitment from member states to ensure the sustainability of operational infrastructures. It is imperative that the unique role of the H2020 RI programme in catalysing the development of infrastructures of the future is not lost in order for it to become another funder of running costs. There is particular concern that a sustainable settlement must be agreed between member states to support PRACE (the HPC infrastructure) and GÉANT (networking), because of the increasing sums being contributed, seemingly haphazardly, from H2020 RI and the possibility that this will seriously restrict funds for other objectives The consideration of routes to sustainability for virtual and distributed infrastructures is very important because there are no established models. Some of the communities have 5

8 no familiarity of working with, or tradition of resourcing, such infrastructures and such concepts as recharging are little understood for intangibles such as data access. Relatively little use has been made of the potential that the financial instruments offer for loans, etc How coherent has the H2020 RI thematic programme been internally and with other (EU) actions? The H2020 RI programme provides a major role in furnishing the excellent science base that underpins the bulk of H2020. There are good examples of infrastructures that directly resonate with the aims of other H2020 programme areas, such as food, agriculture, health, materials science etc. For competitiveness and security reasons there is a major need to develop a more comprehensive approach to underpinning the success of the European Digital Open Market, with a clearer focus on the key areas of the Cloud and HPC. This is an undertaking which involves not just the current EINFRA programme and DG CNECT but many aspects of the wider H2020 programme. There is an opportunity to work with DGREGIO and the Structural Funds to stimulate regional investment in infrastructure relevant to European aims but this requires extensive planning and communication and should be started as soon as possible. The sharing of data, tools and materials is now becoming a substantial feature of scientific infrastructure, and can entail very serious issues of security, personal privacy and ethics. The effectiveness of measures to ensure that ethical and social issues relating to infrastructure development should be reviewed to ensure that they are properly addressed in proposals and considered in evaluation What is the EU added value of the H2020 RI thematic programme? The European added value of the H2020 RI programme is evident in several ways. The European Union as a whole is able to conceive and deliver large infrastructure projects at the European scale and develop and lead those at a global scale, of a type, size and number that would not otherwise be possible. The use of these infrastructures is a major vehicle to promote scientific mobility across the European Union. All of this represents greater value for money because of the minimisation of duplication and improved means of access. Reciprocal arrangements with third countries give the EU access to expensive facilities elsewhere. The networked provision of computing infrastructure and the development of major datadriven research infrastructures is realising the reality of a laboratory without walls in which not just scientists but policy-makers, businesses and society generally can access and integrate data and access research knowledge across the EU, wherever they are Lessons learned for the future The H2020 RI Thematic Programme is one the most important assets for European science. The catalytic role which H2020 RI plays in the early phases of the lifecycle of an international research infrastructure in Europe is unique and indispensable. 6

9 The responsibility for the implementation, construction and exploitation of research infrastructures (including e-infrastructures) is and should remain the responsibility of groups of member states and associated countries and not of the EC. In support of the catalytic role for new and emerging infrastructures the funding for Preparatory Phase projects for new infrastructures should be maintained and support for these should only be given to applications which include sound financial commitments from at least three member states. The element of the programme supporting starter communities to develop new ideas should be increased. The focus of the e-infrastructures part of the programme needs to be widened to include e-infrastructures in microelectronics and 5G technologies alongside Cloud infrastructure and High Performance Computing, taking into account interaction with other relevant parts of H2020. A clearer statement of purpose, scope and strategy for innovation-stimulation activity in the programme needs developing. The continued support for ESFRI is essential, particularly in its role as an incubator for implementation of infrastructures. The potential of eirg to play a similar role for the e- Infrastructures area would be increased by developing channels to MS ministers analogous to those that work for ESFRI. The Structural Funds offer considerable potential for the development of European-level infrastructure capability, particularly in Central and Eastern MS with less RI activity currently. Effort needs to be made to ensure that this potential is understood and actually realised. Special attention should be given to the development of national roadmaps. The pan-european and global infrastructure developed by H2020 RI and its predecessors is providing the foundation of the European Digital Open Market through the realisation of the Open Science and Open Data agendas. The complementary realisation of the European Open Innovation vision is for H2020 and its successor(s) as a whole, but the RI and e-infrastructures programme will be key to it. 7

10 2. INTRODUCTION 2.1. Context Horizon (H2020) is the European Union's research and innovation framework programme for the period H2020 is expected to contribute to the top priority of the Juncker Commission 4 to strengthen Europe s competitiveness and to stimulate investment for the purpose of job creation and growth. The goal is to ensure Europe produces world-class science, removes barriers to innovation and makes it easier for the public and private sectors to work together in delivering innovation. By coupling research and innovation, H2020 is helping to achieve this with its emphasis on three priorities: excellent science, industrial leadership and tackling societal challenges. The implementation of Horizon 2020 is also expected to take a strategic approach to programming of research and innovation, using joint actions and modes of governance aligning closely with policy development yet cutting across the boundaries of traditional sectoral policies. While H2020 sets out the general objective of that framework programme, the priorities and the broad lines, the specific programme 5 define the objectives and the activities which are specific to each of the different Parts. There is a critical need to reinforce and extend the excellence of the Union s science base and ensure a supply of world class research and talent to secure Europe's long term competitiveness and well-being. Part I Excellent science should support the activities of the European Research Council on frontier research, future and emerging technologies, Marie Curie Actions and European research infrastructures, including e-infrastructures. These activities should aim at building competence in the long term, focusing strongly on the next-generation of science, systems and researchers, and providing support for emerging talent from across the Union and from associated countries. Union activities to support excellent science should help consolidate the European Research Area and make the Union s science system more competitive and attractive on a global scale Objectives and intervention logic The 2015 Expert Advisory Group Implementation Report 6 gave a very good summary for the rationale of H2020 RI, as follows: Research infrastructures are facilities, resources and services that are used by the research communities to conduct research and foster innovation in their fields. Where relevant, they may be used beyond research, e.g. for education or public services. By offering high quality research services to users from different countries, by attracting young people to science and by facilitating networking, Research Infrastructures help structuring the scientific community and play a key role in the construction of an efficient research and innovation environment. Because of their ability to assemble a critical mass of people, knowledge and investment, they contribute to national, regional and H2020 RI EAG Implementation report

11 European economic development. They are also crucial in helping Europe move towards open, interconnected, data-driven and computer-intensive research. Reflecting this, the H2020 RI general objectives are: Developing the European research infrastructures for 2020 and beyond Fostering the innovation potential of research infrastructures and their human capital Reinforcing European research infrastructure policy and international cooperation These objectives are discussed further in the following sections Developing the European research infrastructures for 2020 and beyond Developing new world-class research infrastructures The aim is to facilitate and support the preparation, implementation, long-term sustainability and efficient operation of the research infrastructures identified by the European Strategy Forum on Research Infrastructures (ESFRI) and of other world-class research infrastructures, which will help Europe to respond to grand challenges in science, industry and society. This objective will address specifically those infrastructures that are planning to set up, are setting up or have set up their governance, e.g. on the basis of the European Research Infrastructure Consortium (ERIC) or any equivalent structure at European or international level. The Union funding will contribute to, as appropriate: (a) the preparatory phase of future infrastructures (e.g. detailed construction plans, legal arrangements, multiannual planning); (b) the implementation phase (e.g. R&D and engineering work jointly with industry and users, development of regional partner facilities aiming at a more balanced development of the European Research Area); and/or (c) the operation phase (e.g. access, data handling, outreach, training and international cooperation activities). This activity is also support design studies for new research infrastructures through a bottom-up approach Integrating and opening existing national research infrastructures of pan- European interest The aim is to open up, where appropriate, key national and regional research infrastructures to all European researchers, from both academia and industry, and to ensure the optimal use and joint development of these infrastructures. The Union will support networks and clusters that bring together and integrate, on European scale, key national research infrastructures. Funding will be provided to support, in particular, the transnational and virtual access of researchers and the harmonisation and improvement of the services that the infrastructures provide Development, deployment and operation of ICT-based e-infrastructures The aim is to achieve by 2020 a world-leading capability in networking, computing and scientific data in a single and open European space for online research where researchers enjoy leading-edge, ubiquitous and reliable services for networking and computing, and seamless and open access to e-science environments and global data resources. To achieve this goal, support will be given to: global research and education networks providing advanced, standardised and scalable inter-domain services on-demand; grid and cloud infrastructures providing virtually unlimited computational and data processing capacity; an ecosystem of supercomputing facilities, advancing towards exascale; a 9

12 software and service infrastructure, e.g. for simulation and visualisation; real-time collaborative tools; and an interoperable, open and trusted scientific data infrastructure Fostering the innovation potential of research infrastructures and their human capital Exploiting the innovation potential of research infrastructures The goal is to stimulate innovation both in the infrastructures themselves and in their supplier and user industry. To this end, support will be provided to (a) R&D partnerships with industry to develop Union capacities and industrial supply in high-tech areas such as scientific instrumentation or ICT; (b) Pre-commercial procurement by research infrastructure actors to drive forward innovation and act as early adopters or developers of cutting-edge technologies; (c) Stimulate the use of research infrastructures by industry, e.g. as experimental test facilities or knowledge-based centres; and (d) Encourage the integration of research infrastructures into local, regional and global innovation ecosystems The Union actions will also leverage the use of research infrastructures, in particular e- infrastructures, for public services, social innovation, culture and education Strengthening the human capital of research infrastructures The complexity of research infrastructures and the exploitation of their full potential require adequate skills for their managers, engineers and technicians, as well as users. The Union funding will support the training of staff managing and operating research infrastructures of pan-european interest, the exchange of staff and best practices between facilities, and the adequate supply of human resources in key disciplines, including the emergence of specific education curricula. Synergies with the Marie Skłodowska-Curie actions will be encouraged Reinforcing European research infrastructure policy and international cooperation Reinforcing European policy for research infrastructures The aims are to exploit synergies between national and Union initiatives by setting up partnerships between relevant policy makers, funding bodies or advisory groups (e.g. ESFRI, e-infrastructure Reflection Group (e-irg), EIROforum organisations, and national public authorities), to develop complementarities and cooperation between research infrastructures and activities implementing other Union policies (such as regional, cohesion, industrial, health, environment, employment, or development policy), and to ensure coordination between different Union funding sources. Union actions will also support survey, monitoring and assessment of research infrastructures at Union level, as well as relevant policy studies and communication tasks Facilitate strategic international cooperation The aim is to facilitate the development of global research infrastructures i.e. research infrastructures that require funding and agreements on a global scale. The aim is also to facilitate the cooperation of European research infrastructures with their non-european counterparts, ensuring their global interoperability and reach, and to pursue international agreements on the reciprocal use, openness or co-financing of infrastructures. In this respect due account will be taken of the recommendations of the Carnegie Group of Senior Officials on Global Research Infrastructures. Attention will also be given to ensure 10

13 adequate Union participation in coordination with international bodies such as the UN or the OECD. The specific objectives contribute to the Horizon 2020 general objectives on endowing Europe with world-class research infrastructures that are accessible to all researchers in Europe and fully exploiting their potential for scientific advancement and innovation. The lines of activities are aiming at developing the European research infrastructures for 2020 and beyond, fostering their innovation potential and human capital and reinforcing European policy. Coordination with the cohesion funding sources is pursued to ensure synergies and a coherent approach for the development of the research infrastructures. The objectives and the European approach to research infrastructures have evolved over time and progress was made in recent years with the implementation of the European Strategy Forum on Research Infrastructures (ESFRI) roadmap, integrating and opening national research facilities and developing e-infrastructures underpinning a digital European Research Area. The networks of research infrastructures across Europe strengthen its human capital base by providing world-class training for a new generation of researchers and engineers and promoting interdisciplinary collaboration. While the role of Member States remains central in developing and financing research infrastructures, the Union is playing an important part supporting infrastructure, fostering the emergence of new facilities, opening up broad access to national and European infrastructures, and making sure that regional, national, European and international policies are consistent and effective. It is not only necessary to avoid duplication of efforts and to coordinate and rationalise the use of the facilities, but also to pool resources so that the Union can also acquire and operate research infrastructures at world level. The comparison with the objectives of the Specific Programme Research Infrastructures in the Seventh Framework Programme for Research and Technological Development (FP7) shows a high degree of continuity forward to H2020 RIs. The objective of the FP7 programme was: Optimising the use and development of the best research infrastructures existing in Europe, and helping to create in all fields of science and technology new research infrastructures of pan-european interest needed by the European scientific community to remain at the forefront of the advancement of research, and able to help industry to strengthen its base of knowledge and its technological know-how. The action areas were: - Integrating activities (with emphasis on the role of ESFRI) - e-infrastructures - Support for new infrastructures/upgrades - Support for policy development and implementation (which included innovation) These broad action areas all continue forward into Horizon 2020 and there are no conspicuous areas of discontinuity. The changes between FP7 and H2020 RI are reflections of the wider structure of the framework and of the developing policy and scientific context. In FP7 Research Infrastructures formed part of a wider area on Capacities: This part of the Seventh Framework Programme will enhance research and innovation capacities throughout Europe and ensure their optimal use. This aim will be achieved through: 11

14 - Optimising the use and development of research infrastructures, - Strengthening innovative capacities of SMEs and their ability to benefit from research, - Supporting the development of regional research-driven clusters, - Unlocking the research potential in the EU's convergence and outermost regions, - Bringing science and society closer together for the harmonious integration of science and technology in European society, - Support for the coherent development of research policies, - Horizontal actions and measures in support of international cooperation. In H2020 the Research Infrastructures (including e-infrastructures) specific programme is included within the theme of Excellent Science, which includes: - European Research Council (ERC) - Future and Emerging Technologies - Marie Sklodowska Curie actions (MSCA) - Research Infrastructures (including e-infrastructures) There two significant changes here from FP7. The first is that e-infrastructures have more specific emphasis in H2020 RI. This was already emerging clearly during FP7 as the emphasis on the digital economy, as it is now called, began to increase in successive calls and is a reflection of increasing scientific and technological capability and societal and economic policy drivers. The second change is the reorientation of Research Infrastructures from the Capacities area to the Excellence Science one. Most of the areas within FP7 Capacities, inasmuch as they relate to research infrastructures, are still addressed within H2020 RI. However, action to stimulate the establishment and development of research infrastructures or their increased access in convergence and outermost regions is not a specific purpose in H2020 RI, and the coupling of RIs with ERC and MSC in a grouping called Excellent Science is a powerful statement of intent, because it reinforces - in terminology familiar to the scientific community - that this specific programme is science led rather that strategically or end user driven. The envisaged activity is seen as science-driven with a medium- to long-term impact horizon. In the long term these outputs are expected to underpin a wide variety of European and wider scientific, economic, social and other results leading to highly beneficial impacts. Figure 1 represents the intervention logic for H2020 and Figure 2 is an elaborated version more clearly indicating the aspects relevant to H2020 RIs. Apparently notwithstanding the above, an expected output of Research Infrastructures is Creation of new (or upgrading of existing) Centres of Excellence in low R&I performing MS and AC (see section 7.4). 12

15 Figure 1 - Intervention Logic of H

16 Figure 2 - Intervention Logic of H2020 (more detailed) 14

17 3. IMPLEMENTATION STATE OF PLAY 3.1. Overview of programme inputs and activities The data analysed here addresses the situation as of 1 st October The EC contribution allocated to the implementation of the calls included in Work Programmes and which have been closed at the date of 1 st October 2016 has been EUR billion, about 49% of total expected budget allocated to H2020 RI in Horizon 2020, which is EUR billion for the period This leaves approximately EUR billion for any remaining parts of the work programme and whatever is expected in a call. The allocations to each line of activity of H2020 RI and the numbers of awards made over the period are in Table 1. The figures include one award (GEANT Connectivity to Latin America) which was funded jointly from the 2014/15 and 2016/17 e-infra allocations. At the time of the interim evaluation, no projects are completed (the earliest having started mid-2015, there are no scheduled ends before 1 st October 2016). There are 102 projects ongoing from 2014/15 calls. A further 93 awards have been made from the 2016/17 calls, a negligible number of which have started. Table 1 - Activities and allocated share of budget dedicated to European research infrastructures, including e-infrastructures for the programming period Source: Corda, calls in 2014 and 2015, Signed Grants cut-off date by 1/10/2016 (including ad hoc calls) 15

18 3.2. Participation Patterns In this section is described the main participation patterns, per WP, per country and organisation type. Table 2 summarises the EU contribution, participation, implementation and KPI for RI. Table 2 - Summary table on EU contribution, participation, implementation and KPI for RI Summary Total EU contribution EU contribution to signed grants in calls (EUR million) EU-28 (EUR million) Associated Countries (EUR million) Third Countries (EUR million) Participation in signed grants Number of signed grants Number of participations Private sector participation (private/overall) 7.2% 10.0% 8.1% SMEs participation (SME/overall) 5.1% 6.8% 5.6% Implementation Time-to-Grant (% of projects within TTG benchmark) 59.0% 81.6% 67.7% Success Rate (projects/proposals) 23.9% 24.8% 25.2% KPI Number of researchers who have access to research infrastructures through Union support Source: Corda, calls in 2014 and 2015, Signed Grants cut-off date by 1/10/2016 (including ad hoc calls) Significant figures in Table 2 are a) The success rate of 25% of proposals. This is rather higher than the H2020 average (which is about 14%), but a number of components of the programme have, in effect, been targeted at one or a very narrow pool of potential applicants, suggesting that those areas which are intended to be broadly competitive have lower success rates b) The level of private sector participation and, within that, of SMEs (8.1% and 5.6%), is relatively low, but reflects that this is basic research infrastructure development. c) The KPI of numbers of researchers having access through EU support is more than This is a complex area to measure and the data largely relate to FP7 investments. Although of itself impressive, it does not properly capture the levels or diversity of research access being provided eg by distributed and data-led infrastructures - there are, for example, data services within infrastructures supported 16

19 by the H2020 RI programme with annual hit rates counted in millions and registered industrial users in the tens of thousands. No does the figure reflect in any way the scale and breadth of usage enabled by the EU investments in networked and high performance computing. d) The time-to-grant figure has risen between 2014 and 2015 but still shows one third of grants not being announced within the agreed target. Table 3 - Summary table of Budget, Participations, Implementation and KPI under European Research Infrastructures actions Summary Total Budget Estimated total budget in WP (EUR million) 279, EU funding to signed grants in calls (EUR million) Average EU funding per signed grant (EUR million) Participation signed grants Number of signed grants Total number of participations Newcomer participations (EU-13/overall) 6.6% 6.7% 6.6% EU-13 participation (EU-13/overall) 10.1% 11.7% 10.6% Associated countries participation (associated countries/overall) 7.5% 10.7% 8.4% Third countries participation (third countries/overall) 4.7% 1.6% 3.7% Private sector participation (private/overall) 7.2% 10.0% 8.1% SMEs participation (SME/overall) 5.1% 6.8% 5.6% Implementation Time-to-Grant (% of projects within TTG benchmark) 59.0% 81.6% 67.7% Success Rate (projects/proposals) 23.9% 24.8% 25.2% Success Rate ( allocated/requested) 29.0% 25.1% 28.3% KPIs Number of researchers who have access to research infrastructures through Union support Source: Corda, calls in 2014 and 2015, Signed Grants cut-off date by 1/09/2016 (including ad hoc calls) The participation figures (Table 3) are mainly notable for the relatively low level of participation by the 13 newer EU MS, which is exceeded by the involvement of ACs. This is not particularly surprising; given the concentration of existing infrastructure capability 17

20 in NW Europe, but it is a concern. The participation of third countries - 1.5% - is low, but probably underestimates the level of developing global reach of some EU-funded research infrastructures. Under the Work Programme , 4 calls with 6 sub-calls were launched in 2014 with EUR 277 million of estimated budget, as shown in Table 4. Table 4 Summary of the 2014 Calls Developing new world class research infrastructures H2020-INFRADEV Budget: EUR 70 million Integrating and opening research infrastructures of European interest H2020-INFRAIA Budget: EUR 90 million This call aimed to facilitate and support the implementation, longterm sustainability and efficient operation of the research infrastructures identified by the European Strategy Forum on Research Infrastructures (ESFRI) as well as other world-class research infrastructures, which will help Europe respond to grand challenges in science, industry and society. In addition, the next generation of new research infrastructures have been identified through design studies. This call focused on opening up key national and regional research infrastructures to all European researchers from both academia and industry and ensuring their optimal use and joint development. Through a targeted approach, specific types of research infrastructures or research communities were invited to submit proposals, ranging across all fields of science and technology. e-infrastructure for Open Access H2020-EINFRA Budget: EUR 13 million e-infrastructures H2020-EINFRA Budget: EUR 82 million This sub-call targeted an infrastructure supporting reliable and permanent access to digital scientific records. A key element is capacity building to link literature and data in order to enable a more transparent evaluation of research and reproducibility of results. This action includes an analysis of alternative means of public support to Gold Open. A key objective is the provision of service driven infrastructures to enable wide participation in the Open Data Pilot. This sub- call, by enabling more cooperation as well as computationintensive and data-intensive research across disciplines, contributed in making European research academic or industrial - more innovative and efficient. E-Infrastructures contributed to make every European researcher digital, increasing creativity and efficiency of research and bridging the divide between developed and less developed regions. Support to policy H2020-INFRASUPP Budget: EUR 2 million Support to innovation, human resources, policy and international cooperation H2020-INFRASUPP Budget: EUR 20 million This call focused on facilitating trans-national co-operation between NCPs for research infrastructures with a view to identifying and sharing good practices and raising the general standard of support to programme applicants. This call focused on fostering the innovation potential and developing the human resources of research infrastructures especially in areas that suffer from shortages in supply or where new skills and professions need to emerge, e.g. in 'data science'. It also aimed at reinforcing European research infrastructures policy and international cooperation. The topics for e-infrastructures aimed to optimise related investments in Europe by coordinating European, national and/or regional policies and programmes for e- infrastructures, it also facilitated the cooperation of European e- infrastructures with their non-european counterparts, by promoting, among others, connectivity and global e-infrastructure services. In complementarity with the HPC activities led under the FET part, a study on High Performance Computing (budget: EUR 0.2 million) with the aim of following the progress 18

21 on the implementation on the HPC strategy and presenting the evolution of the European HPC ecosystem was launched in Under the Work Programme , 3 calls with 5 sub-calls were launched in 2015 with EUR 200 million of estimated budget. In addition 107 M were allocated to fund further proposals in the ranked lists of 2014 sub-calls. Table 5 Summary of the 2015 Calls e-infrastructures (H2020-EINFRA ) Budget: EUR 80,5 million The call covered two topics, Centres of Excellence (CoE) for computing applications and e-infrastructures for virtual research environments (VREs). CoEs ensure EU competitiveness in the application of HPC for addressing scientific, industrial or societal challenges. They are userfocused, develop a culture of excellence, both scientific and industrial and are placing computational science and the harnessing of 'big data' at the centre of scientific discovery and industrial competitiveness. VREs support capacity building in interdisciplinary research communities to empower researchers through development and deployment of service-driven digital research environments. They integrate resources across all layers of the e-infrastructure (networking, computing, data, software, user interfaces), foster crossdisciplinary data interoperability and provide functions allowing data citation and promoting data sharing and trust. Developing new world class research infrastructures (H2020-INFRADEV ) This sub-call targeted the implementation and initial operation of new research infrastructures which were identified by ESFRI, in the context of the prioritisation exercise, as essential to extend the frontiers of knowledge in the fields concerned and mature enough to be under implementation by Budget: EUR 97 million Developing new world class research infrastructures (H2020-INFRADEV ) Budget: EUR 6 million This sub-call supported the preparatory Phase of ESFRI projects targeting in particular those projects, which were already supported by EU through a first preparatory phase grant and had not yet entered in implementation phase. These grants will allow these projects to finalise their preparatory phase, bringing them to the level of legal, financial, and, where applicable, technical maturity required for implementing it. This sub-call provided support to the development of new professions and skills for e-infrastructures. The changing methods of (digital) science and research require that researchers, professors and students receive adequate support in computing and networking, as well as in handling, analysing and storing large amounts of digital content. Professional recognition of professions of infrastructure operators such as research technologists, data scientists or "data librarians" and the development of appropriate curricula, training and skills are crucial to ensure effective services to institution staff and students. Support to innovation Support to human resources (H2020- INFRASUPP ) Budget: EUR 2,5 million (H2020-INFRASUPP ) This sub-call targeted Innovative procurement pilot actions in the field of scientific instrumentation exploiting the innovation potential of research infrastructures using Pre-Commercial Procurement (PCP) and/or Public Procurement of Innovation (PPI) schemes. Budget: EUR 14 million Other actions launched in 2015 were: 19

22 The 1 st specific grant agreement (SGA) for GÉANT has been awarded within the GÉANT Framework Partnership Agreement (FPA). GÉANT is the European communications commons that supports the rise of compute- and data-intensive collaborative research and education through innovative services, operational excellence and global reach. The budget for the implementation of the first year of the action plan is EUR 25 million. A specific action aiming at re-enforcing the connectivity with Latin America as a strategic and long term investment for research and education was launched. The budget for this action is EUR 5 million implemented as a specific grant agreement under the GÉANT2020 FPA. A complementary funding of EUR 5 million is foreseen under WP16-17 A grant to identified beneficiaries (H2020-Adhoc ) to support the organisation of The International Conference on Research Infrastructures (ICRI-2016). The budget for the implementation of this Coordination and support action was EUR 0.3 million. ICRI 2016 will be hosted in South Africa in October In the recent years, ICRI has become the platform for stakeholders to engage and interact on many of the critical research infrastructure (RI) related questions. The event bringing together experts from across the globe and also seeking to make recommendations going forward for the RI community Cross-cutting issues Table 6 shows a rough analysis by broad subject area of the DG-RTD awards made in H2020 RI as of February Table 6 - Scientific classification of H2020 RI projects funded by DG-RTD up to February 2016 i.e. calls in 14/15. (Approx. figures) Science area Number of awards Total value M % by value Environment & energy Life Sciences inc. health and food Physical sciences & materials Social sciences & humanities Policy Sustainability, biodiversity, climate change Table 6 shows that about one third, by value, of the resource awarded from the DG-RTD component of the programme, up to Feb 2016, was in the general area of environment and (sustainable) energy. This included investments such as: - the GLOBIS-B global cooperation of infrastructures focused on services to support research predicting the biosphere and measuring the indicators of biodiversity change, - the European Carbon Dioxide Capture and Storage Laboratory Infrastructure - Support for the implementation phase of EPOS, the very large pan-european infrastructure for geo-physical monitoring which will be an essential resource for, inter alia, developing more effective earthquake prediction and management for the vulnerable regions of Europe ( 19M contribution to a 31M programme) 20

23 Social sciences and humanities In addition to the significant number and value of projects funded in these specific areas that is shown in Table 4, including infrastructures on the ESFRI roadmap, there is a very major contribution, both existing and anticipated, catalysed by access to networked computing and HPC and by the delivery of open data agenda. What is less evident is the level of involvement of SSH researchers in the delivery of H2020 RI projects more generally and what any claims of consideration of SSH issues within funded projects actually involve. This is a significant issue because of both the very important wider issues for society involved in both the increasing digital economy and the Three O s agenda 7, and also the opportunities these developments also offer to inform policy in novel ways. See also section 7.3. Gender issues The cross disciplinary nature of the programme, and the fact that gender balances are known to vary very widely across different subject disciplines means that conclusions that specifically relate to the H2020 RI programme will be fairly meaningless. SMEs and Innovation The SME participation rate in the programme (Table 2) to 1 October 2016 is 5.8% - this refers to the proportion of partners in grants awarded that were from SMEs. There have been a number of innovation-related call topics, clustered under INFRAINNOV within the DG RTD component, and within EINFRA, which will have engaged SMEs. There is also SME engagement in the large partnerships around HPC and networking (PRACE and GEANT). There is a significant network of HPC competence centres for SMEs that has been developed as a result of components of several EINFRA calls. The need for a generally more coherent approach to innovation is discussed in section RELEVANCE: 4.1. Is H2020 RI tackling the right issues? The background to H2020 is set very much in the context of general budgetary constraint following the world financial crisis and subsequent economic downturn Although negotiations reflected this, H2020 still represents the largest - and longest - EC Research and Innovation programme thus far, and the extent of commitment from member states reflects the view that investment into the research base is seen as a major means to underpin future economic growth and societal development. To this extent the expectations of impact and benefit from the whole H2020 programme, including H2020 RI, are considerable. However, as stated before, H2020 RI is seen as science-driven with a medium- to longterm impact horizon. Consequently in examining its relevance there are three main issues: - Have the objectives set out for H2020 RI been appropriately and well addressed? 7 Open innovation, open science, open to the world: A vision for Europe 21

24 - Have they proved to be appropriate in the light of changing circumstances and needs since they were elaborated? - How are they contributing to the broader aims of the EU? The four (now five) components of H2020 RI, as it has been delivered so far, closely map to the main objectives of the Specific Programme and therefore, in the broadest sense, are completely relevant. INFRADEV addresses the development of new pan-european infrastructures, and is significantly (but not entirely) concerned with supporting the implementation of the strategic recommendations of ESFRI, the European Strategy Forum on Research Infrastructures. INFRAIA is about increasing and improving European-wide access to infrastructures at the national or regional level or of wide likely EU interest. It also explores the identification of new ideas for European Infrastructures ( starting communities ). EINFRA addresses e-infrastructures, and is in various ways distinctive. The infrastructure it addresses - for computing and data services - is not always solely, or in some cases even primarily, intended for the scientific research community in the narrow sense. In some ways it has more in common with the common utilities - water, electricity, the telephone etc. Research directly underpins the development and delivery of effective services in the field, but the services themselves are much closer to, and address a much wider potential user base than, those being developed in most other parts of the RIs programme. This broad impact landscape is described as the digital single market. This gives this part of H2020 RI a considerably wider stakeholder community and a very different innovation perceptive and expectation. INFRASUPP addresses support to innovation, human resources, policy and international cooperation and comprises several different types of activity. INFRAINNOV: the 2016/17 work programme the innovation area of INFRASUPP was elevated to a new area called INFRAINNOV, with the innovation-related parts of INFRASUPP in 2014/15 being reclassified accordingly. In their general structure these areas and the objectives they address are in most ways a furtherance of those in FP7, the ex-post evaluation of which was broadly very positive. The growing emphasis on Innovation is a significant development, as is the greater pervasiveness of the virtual/distributed infrastructures and the e-science aspects, and is very consistent with the Junker vision for EU priorities set out in the document A New Start for Europe 8 in This stressed the role of research and innovation in underpinning a boost for jobs, growth and investment, and also the critical importance of the realisation of the digital single market. There is a very through and well-argued Expert Advisory Group report on the proposed H2020 RI work programme for produced in June There is very little to disagree with in this document and, in consequence, there is not the need to repeatedly referenced it throughout this report. INFRADEV INFRADEV is the largest component of H2020 RI in financial terms. Its investments so far have addressed four broad areas: Design Studies 8 A New Start for Europe 9 EAG RI H

25 Preparatory Phase of ESFRI projects Individual implementation and operation of ESFRI projects Implementation and operation of cross-cutting services and solutions for clusters of ESFRI and other relevant research infrastructure initiatives Three of these four areas directly reference ESFRI and its Roadmap. ESFRI is a body established in 2002, primarily comprised of representatives of all MS and AC and the EC, and appointed by and reporting to MS research ministers. Its principal role is to act as an incubator for new and developing European (and, increasingly, global) research infrastructure concepts and inform their prioritisation and development to implementation. In doing this ESFRI maintains and updates a strategic roadmap of new research infrastructure requirements. It does this by consultation with the scientific and user communities via call and evaluation exercises. There has been a close relationship between ESFRI, the evolving roadmap and the research infrastructure investment by the EC since ESFRI began its work, and the birth of ESFRI itself was very considerably assisted by the EC. The most recent ESFRI Strategy Report/Roadmap was published in early The portfolio of RIs supported within INFRADEV in H2020 RI thus far is impressive in its quality but also in its diversity - the use by ESFRI of discipline-based working groups has ensured that infrastructure needs in life sciences, environment, energy, social sciences, humanities etc are addressed as well as physics and materials, and that there is strong support for the distributed and data-led infrastructures that are an increasing feature of the digital and open data era. The introduction of cluster awards, enabling infrastructures with aspects in common that would benefit from common development and shared approaches is particularly clever because it also resonates with the opportunities of the open data agenda and the need for more interoperability between diverse datasets. Not all of INFRADEV is ESFRI-related and the portfolio so far includes many other substantial programmes to develop infrastructures across the disciplines and not developed via the ESFRI route, Overall, INFRADEV appears to be delivering the expected objectives set in H2020 RI exceptional well. INFRAIA The INFRAIA area is about opening up existing European infrastructures to the wider European community. The first call of INFRAIA 14/15 supported a very diverse portfolio of 19 projects, all of which met the general aim of the call in their different ways. It is difficult to see much commonality between them beyond a commendable focus on excellence and wide scientific and beneficiary engagement. Such a list is always going to be vulnerable to criticism of its apparent randomness. It is, however, important to recognise that the costs allocated reflect the large sums needed to make a difference at the European scale, that the overall impact of the portfolio in terms of engagement with a wide diversity of scientific users across so many member states is very considerable, and that the effect of these programmes as beacons of good practice and demonstrators of pan-european scientific integration and outreach is very considerable. 10 ESFRI Roadmap

26 There is relatively little specific emphasis in the INRAIA work plans on the potential of this intervention to promote access to research infrastructures to accelerate research to underpin policy or innovation. This is presumably reflecting the excellent science lead. However, there are e.g. specific areas mentioned which are of clear applicable relevance in the list of topics in INFRAIA 01 e.g. vaccine research (a major global need) and plant phenotyping (a technology of direct relevance to plant breeding and global food security) to name two. It should be considered whether placing greater emphasis on relevance to policy and innovation goals as an objective of the INFRAIA area, rather than seeing it as almost incidental, would be of benefit. EINFRA Amongst the many distinctions EINFRA has from the rest of H2020 RI is that it is broken into numerous areas in the work plans, many of which seek to solicit a single proposal or address a very specific requirement. Nevertheless, most of the expectations of the EINFRA programme at the outset have thus far been addressed. The following is of note: - Networking: world-class infrastructure and governance mechanisms have and are being developed. - Distributed computing: a proven grid computing infrastructure is in place. This is now being transformed in cloud services - High Performance Computing (HPC): an efficient mechanism for allocation of cycles has been devised. - Data: The Research Data Alliance was launched as a bottom-up organisation tackling research data issues such as interoperability and standardisation - Open access: A robust infrastructure supporting open access policies is in place, regarding both publications and data. - Centres of Excellence: critical mass of HPC is being built - Virtual research environments (VREs): user experimentation with e-infrastructure platforms is taking place. Also, despite being already rather broad, EINFRA has shown sufficient agility to respond the rapid policy developments in the Three O's/Open Data area by e.g. assimilating the European Open Science Cloud pilot into its vision in a very effective way. Thus far, during the course of the EINFRA programme, there have been a series of specific, and sometimes very large, investments into aspects of HPC provision (PRACE) and networking (GEANT). In the case of GEANT this involves, inter alia, the funds to join the various national components into a cross-european whole. Each of these funding investments appears sensible in its specific purpose, and there is no doubt that these infrastructures are absolutely vital to the future of European science. However this piecemeal injection of H2020 funding (which has now spread beyond EINFRA and INFRASUPP into INFRADEV) is neither effective nor efficient because it is not addressing the main issue here, which is that there is no long-term agreement for the European-level funding and development of these essential resources by member states. It also seems likely that by continuing to support the glue which holds the parts together, H2020 RI may be removing the incentive for such an agreement to develop. The sums being invested from H2020, although very substantial in the context of the RIs programme, are relatively small in relation to even the current investment in HPC and networking by MS. It is clear that the provision of HPC, networking and distributed computing and data infrastructure across Europe is ultimately a pervasive economic and social necessity and 24

27 not solely or even primarily an infrastructure for research. It has not been found no persuasive argument that extending or increasing EC support further in this area in the way it is currently happening will do anything to accelerate the uptake of this responsibility by MS. the need for a much more comprehensive approach to underpinning the European Digital Open Market is discussed further in section The Competitiveness Council meeting in May invited ESFRI to explore mechanisms for better coordination of Member States' investment strategies in e- infrastructures, covering also HPC, distributed computing, scientific data and networks. ESFRI has set-up a working group for this purpose but it has not yet delivered its report. An effective high-level European e-infrastructure roadmapping group connected to MS science ministers and with some engagement with Finance Ministries, able to advise the catalytic role of the EC, would be extremely helpful in progressing this issue. (See section 6.3 and 7.2). Until it is progressed there will be continuing and growing concern that the demands of MS for more investment from H2020 RI into computing development is placing unreasonable pressure on the programme and in particular EINFRA, and is a displacement exercise, diverting attention away from the real issue rather than contribution to its solution. Consequently, suggestions that the EC should fund the provision of exascale computing for European research (see section 6.2) is a somewhat alarming prospect unless there are very clear terms and timescale for the operating costs to be taken over my MS. Despite the above concerns, the EINFRA programme has made outstanding progress within the aims set for it. In particular, by the end of H2020 it should be able to have generated a coherent and complete catalogue of services, supported by public e- Infrastructures, that serves the scientific community with a view to serving wider communities beyond Such a catalogue of services will be the foundation of the European Open Science Cloud (EOSC). INFRASUPP - Coordination and support actions The INFRASUPP programme addresses support to innovation, human resources, policy and international cooperation and comprises several different types of action. In terms of effectiveness, the majority of what has been funded in 14/15 and what is proposed in 2016/17 is clearly focussed on adding value to the major investments in the rest of the programme - whether by exploring innovation potential for or from European RIs, developing skill sets in needed to operate and use them effectively, undertaking studies informing European RI policy or exploring international links. INFRAINNOV - Support for innovation The INFRAINNOV badge was invented sometime after the H2020 RI call was issued, and comprises two topics rebadged from INFRASUPP in that call, plus two further calls under the INFRAINNOV head in the call, one of which was a large (EUR 20million) focus on advanced on imaging technology development (an exceedingly timely and appropriate subject, but not uniquely so). The INFRAINNOV area only covers the aspect of the H2020 RI programme managed by DG RTD - there are several significant innovation related investments being made through EINFRA. The overall objectives of H2020 RI in relation to innovation are poorly articulated, notwithstanding the intelligent analysis in the Expert Group report on priorities. There are three broad ways in which activities in the programme create innovation potential: EAG RI H

28 Science and technology has to be developed to enable the development, construction and delivery of access to advanced research infrastructures. Some of the capability to do that is in the private sector or public-sector technology transfer entities. This may have innovation potential well beyond the specific purpose for which it is being developed. Research undertaken using research infrastructures frequently has innovation and benefit potential Platforms which enable access to data and tools by problem owners (in industry, the public sector and society more generally) are creating an innovation ecosystem in which the mutual dependency of the infrastructure, the research base and the user community is greatly increasing. Elements of these areas of innovation potential are being addressed piecemeal, but there needs to be more comprehensive and intelligible aims and strategy Flexibility to adapt to new scientific and socio-economic developments The period since H2020 was conceived has seen some exceptionally rapid developments in science with major consequences for infrastructure needs and development. In particular, developments in distributed computing, overall computer power and highvolume data transmission have combined to produce an explosion of data-driven science, giving scientists in many disciplines inter-operable access to research data of a hithertounimagined scale and diversity. From this has come a great number of platform needs around the data sharing/data access area and the associated requirements for standardisation, tool development etc. In life sciences this has coincided with the development and availability of highthroughput and automated omics and imaging technologies giving extensive insights into the function and diversity of genetic and molecular processes and of phenotypes. In the social and environmental sciences, healthcare, the humanities and many other fields, digitisation techniques of all kinds have enabled broad access to previously inaccessible information, a wealth of time-series data, and to all manner documents and images, in some cases dating back millennia, that were hitherto accessible only to very few. Moreover, these developments do not solely concern scientific research but address the key technological platforms relating to computing and data. The European vision which this underpins is set out in two key policy documents. The Junker Commission vision of a New Start for Europe (2014) 13 stresses the central importance of the digital single market and the growing need for Europe be a global actor. The Open Innovation/Open Science/Open to the World (known as The Three O s) 14 vision document produced by Commissioner Moedas articulates how the enabling power of ICT, open data and open science enables global approaches the great challenges and opportunities on the 21st century. The Research Infrastructures H2020 and FP7 specific programmes (and earlier) have substantially underpinned the ability of the EU to face up to these issues, working closely with commission officials over many years. The first significant international vision on the open data agenda was an EC High Level Expert Group Report, Riding the Wave (2010) 15 Subsequent support for the Research Data Alliance (RDA) to work globally has placed the EU at the forefront of transnational data access policy. The data sharing policy - first political priorities of the Juncker Commission 14 Open innovation, open science, open to the world: A vision for Europe

29 released as a pilot 16 and now to be fully implemented across H was the first such policy in a transnational research funder. This EU leadership at a global level would not have been possible without the combined efforts of the European scientific community, the EC and the expert advisors - particularly in ESFRI and EIRG - working together over many years. They identified future scientific and technological potential and framed, funded and delivered a series of European interventions which have established the basis for open data and open science in Europe, supported by a growing research infrastructure capability and more people to enable it. This is a great achievement of H2020 RI and its predecessors and its transformative potential is economically and socially almost limitless. All these developments bring with them issues of ownership, commercial and personal privacy and freedom, or ethical concerns of many kinds. The H2020 RI programme has had to recognise and respond to these issues as they have arisen. One aspect that has been important in realising this vision has been the way the budgetary framework has enabled flexibility and agility. The 2016/7 work plan for INFRADEV included the European Science Open Cloud pilot. This is an excellent example of the flexibility inherent in the model of a longish funding framework with successive work programmes - this major policy opportunity underpinning the three O's was included rapidly and effectively and has delivered a further leg to the European structure to deliver the open data and research agenda and underpin the digital single market. It was also notable how ESOC integrated effectively into the forward vision for EINFRA reflecting the effective level of strategic planning that is maintained within the programme. 5. EFFECTIVENESS The H2020 RIs programme is about doing and enabling excellent science and its effectiveness is about how well that is achieved. There are several key areas: The effectiveness with which European research infrastructures have been implemented The importance of ESFRI The effectiveness of INFRAIA: the role of speculative science and the Open Science agenda EINFRA and the future of the Digital Open Market 5.1. Short-term outputs from the programme The effectiveness with which European research infrastructures have been implemented The ex-post review of FP7 17 lamented that a target set by the Innovation Union of launching or beginning the construction of 60% of the 2010 ESFRI roadmap by 2015 was proving slow to realise. The ESFRI Roadmap demonstrates that this target was met. However, the issue highlighted by the FP7 review was of considerable significance, drawing attention to a need to gain more engagement from MS and to show more selectivity and structure in the process of identifying and taking forward European Infrastructure proposals from initial concepts to launch. These issues have been reviewed by ESFRI and are reflected in the structure of the INFRADEV part of the programme. In particular, there is a structure for each of the design, preparatory and implementation phases, with clear and different criteria appropriate to these levels, no expectation of an automatic ascent up a ladder of successive support and an overriding rule that an

30 infrastructure can only remain on the Roadmap for 10 years. ESFRI has thus moved from a phase of growing the roadmap to maintaining a more steady state situation in which a realistic number of the best cases can be brought through to maturity The importance of ESFRI The ESFRI process, and the relationship between it and the EU funding of infrastructures has to be seen as one of the major achievements in the development and delivery of international science in recent years, bringing together as it does the European scientific community, the collective action of MS science ministers, the pan-european resource of the EU and the strategic leadership of the Commission into a coherent strategic direction. In doing this it has overcome the otherwise insurmountable difficulty of negotiating the genesis and development of numerous major infrastructures between multiple partners. The catalytic role of the EC in achieving this has been vital and will continue to be so. In addition to this the process has led to a widening of the concept of Pan European infrastructures beyond very large centralised facilities to include distributed and integrated infrastructures, many based around needs for acquisition, distribution and standardised access to data. Amongst other things this has helped propel Europe into the global forefront of the open data/open science agenda. However, the wider value has been in leading a recognition that the integrative working of diverse resources across member states is possible, and that by doing so a larger, more coherent and effective scientific effort can be enabled in many areas which makes more effective use of resources to do better science. By supporting the development of outstanding exemplars of this approach INFRADEV has raised awareness of the burgeoning potential and stimulated scientific communities across the EU. Recently ESFRI has developed closer working with EIRG and has also begun to acquire greater representation of computational and data-led science expertise in its membership. It will be interesting to see what role it can play in advising the development of EOSC The effectiveness of INFRAIA: the role of speculative science and the Open Science agenda Member state delegations to the Programme Committee that have made input to this review have regarded INFRAIA together with INFRADEV as the most appreciated parts of the overall programme. INFRAIA is particularly well regarded because the nature of its calls is such that delegations and the scientific community can have input to the process and prioritisation. It is also the part of H2020 RI that most frequently elicited comments from project participants (and unsuccessful applicants). In the 2016/17 work programme 19 INFRAIA was split into two parts, for advanced communities (INFRAIA-01) and starter communities (INFRAIA-02). There seems a sensible separation, given the inevitable advantage that established areas would have over new ones in a competition with one another. It also seems a good idea to try to focus the topics for established areas, and to allow a more open call for the starter ones and allow more time for those proposals to be evaluated. The list of call topics in INFRAIA was very specific, some based around opportunities to integrate related infrastructures, often with a view to underpinning the realisation of particular policy goals etc (eg vaccine research, earthquake research, poverty and living conditions, widening access to HPC) or improving integration of activities around existing

31 and developing large infrastructures (eg neutron beams, synchrotrons, lasers etc). The requirements for each are specified, and most (but not all) ask those seeking funding to actually demonstrate that they are attracting significant users from other countries. There is strong emphasis on good data management practice. All of this makes the call very relevant. The 4:1 balance in budget in favour of INFRAIA-01 is of some concern. The number of important scientific and policy areas potentially addressed by the networked infrastructures in INFRAIA-01 is large. Some of the areas and consortia in INFRAIA-01 are not particularly new - they represent a repackaging of consortia or activities previously funded in other ways. It is appreciated that this is sometimes desirable, and also that preventing it is very difficult, but it must be recognised that if it is done too much then the essentially bottom-up and science-driven that is securing the future vitality and purpose of the programme is weakened. The way that topics in INFRAIA-01 had been chosen was not clear in the call - it was in fact the result, in part, of a consultative exercise that - inter alia - engaged with other DGs in the EC and the MS reps on the Programme Committee, alongside other inputs. This is a positive way to consider relevance and could have been explained, although it does also raise the issue of why the more policy-relevant infrastructures could not be funded, or at least co-funded, with other parts of H2020 responsible for research in the relevant areas. There is a strong argument for investing a greater proportion in the new ideas sought by INFRAIA-02. Given the speed at which research technology is developing there needs to be regular opportunity for new ideas to be explored based on scientific merit. It is therefore recommend that the share of funding between the two components be reassessed in future calls. See also section 8.2. There is no broad equivalent to INFRAIA-02 in the EINFRA programme, a point raised by members of EIRG and others in the relevant research community. It would be helpful if DG-CNECT could consider how to ensure that there is adequate opportunity for bottomup idea input from the wider scientific community. (See also section 5.6) One of the reasons given for the reluctance to allow more open invitation of speculative proposals from the scientific community has been the concern that some of them will fail to deliver the anticipated results, or that these will not prove useful. Several points need to be considered here: - investment in scientific ideas involves risk and although a review process can assess these relative to cost and likely potential benefit, seeking to eliminate risk in the conduct of science is not sensible - almost all well-conducted and internationally competitive science delivers beneficial outputs tools, methods, research materials, training and skilled people etc, and those that do not deliver all their original objectives usually contribute to the base of knowledge by eg disproving significant hypotheses, demonstrating the limits of technologies, suggesting different routes of future exploration - and, sometimes, achieving something unanticipated. - the open science and open data agendas mean that far less duplication and repetition in science need take place in future, if results, methods and tools are shared. Consequently, the EC will maximize the scientific value of all its research investments if it vigorously enforces policies to release data and tools into infrastructures able to present these in usable form to the widest community. This does mean that the terms by which public open science sits alongside public/private partnerships, and how relevant infrastructures can provide platforms within which SMEs and private sector organizations to use public data securely, need to be better 29

32 articulated. It also means stopping grant holders and national funders withholding publicly-funded data and tools etc. on vague and generic grounds of future commercial interest or maintaining competitive advantage. Funding should be provided within grants to support the preparation and delivery of data, tools and materials into appropriate repositories, in order to ensure that this completed as part of the deliverables of the award. This will help to accelerate the development of data-led infrastructures and their key role in delivering the Open Science agenda EINFRA and the future of the Digital Open Market The e-infrastructures field is probably the fastest-developing and most economically and socially significant part of the global technology infrastructures landscape at the present time. It is the focus of intense political and commercial interest, with which the EC is deeply involved. As an example, in just two months April-May 2016 the EC produced five significant policy outputs relating to research in the area: - The European Science Cloud Initiative 21 - Digitising European Industry Communication 22 - ICT Standardisation Prioritisation for the Digital Single Market 23 - Communication on EU e-government Action Plan Competitiveness Council Conclusions on the transition towards an Open Science system 25 In addition, the Three O's which describes a global innovation landscape underpinned by an open science agenda fundamentally predicated on open sharing of research and other data. This is entirely dependent on the types of information and computing infrastructure which are primarily the focus of EINFRA. The component of H2020 RI that is delivered through DG CNECT has, at its heart, a confusion between whether it is about developing e-infrastructure for research or doing research to support the development of e-infrastructure. In practice the two are intimately related and the potential scope of the landscape being addressed is very large indeed. There has been an effort to rationalise this using a concept of Technology Relevance Levels (TRL), with EINFRA concentrating funding on TRL6 or above, and more basic work at TRL5 of below being considered mainly elsewhere in H2020. The draft Work Programme is structured on two themes: Theme 1 - Integration and consolidation of e-infrastructure platforms supporting European policies and research and education communities. The call recognises the two technology-push and user-pull approaches to this issue Theme 2 - Prototyping innovative e-infrastructure platforms and services for research and education communities, industry and the citizens at large. This structuring does present a more explicit recognition of the wider impact of the programme than was evident in the previous call, and also - in the user-pull approach - the recognition that in some areas, particularly data infrastructures, there are very substantial developments arising from the user communities

33 A number of the components of the call were in effect single-application and mostly continuations of funding to large infrastructure developments in computing or data management. Compared to the other parts of the H2020 RI programme, or H2020 as a whole, EINFRA funds relatively few research projects in the conventional sense. As mentioned before, it has been received some input suggesting that this may be restricting creative and speculative scientific input and that a call similar to the relatively open one in INFRAIA soliciting new ideas might be helpful. However, the implication of the TRL concept mentioned above could suggest that such speculative work is unlikely to be appropriate to EINFRA. However, the main problem perceived with EINFRA is that whilst it is delivering activity that is very relevant to the agenda that has been set for it in H2020, and delivering it extremely well, the agenda itself omits a key aspect of the requirements to deliver an effective vision for the European Digital Open Market, which is the development of European research infrastructure and development capability in ICT hardware and technology. This is discussed more fully under Coherence, below. (See section 7.2) The EINFRA programme does, to some extent, sit uncomfortably in the Excellent Science pillar because although very much delivered by an excellent science base, and underpinning capability for excellent research, the programme has considerably more extensive, pervasive and immediate relevance to the wider economy and society than much of the rest of the H2020 RI programme. Nevertheless, the logic of locating it within the H2020 RI programme is clear when the level of interaction and interdependency across the programme as a whole is considered. 6. PROGRESS TOWARDS THE OVERALL HORIZON 2020 OBJECTIVES 6.1. Contribution to the achievement and functioning of the ERA The key factors demonstrating the role of H2020 RI to the realisation of the ERA are summarised in Section 8. Box 1 Contribution to the achievement and functioning of the ERA Horizon 2020 provides support to Member States and the main stakeholders in implementing the ERA reform agenda across the following key priorities: Optimal transnational co-operation and competition on common research agendas, grand challenges and infrastructures (P2P's, ESFRI and ERIC ) In order to measure the contribution of RI Horizon 2020 to the realisation of the ERA, the following indicator has been identified: Number of national research infrastructures networked (in the sense of being made accessible to all researchers in Europe and beyond through Union support) 7. EFFICIENCY No significant criticism of the administration of the H2020 RI programme by the Commission services has been received, and the officers involved are held in high regard for their knowledge and helpfulness. Some members of the Programme Committee find the actions of the senior officials somewhat peremptory on occasion, but many praise the leadership and vision which has been shown. Planning science at an international level is not easy, and the success of the programme in setting successful strategic direction for the European Union in the development of European-level research infrastructures is undoubted. 31

34 7.1. Budgetary Resources The scientific community is particularly sensitive to success rates. Excessive oversubscription wastes a great deal of effort, whilst under-subscription may indicate a lack of the competitive pressure needed to sustain excellence. These issues are modulated in H2020 RI because much of the investment is to develop things already in train and consequently some calls may be very focussed and not expected to produce multiple applications. There has not been much comment received about success rates generally. However, the Virtual Research Environments (VRE) call EINFRA-9 14/15 was very oversubscribed, creating some disappointment in the scientific community and concern from some members of the Programme Committee. It was decided not to react to this e.g. by finding more funds or holding another call. The issue perhaps arose because the call drifted away from seeking to add European value to existing and developing member state infrastructure and associated competencies towards something that looked to many researchers like a call for research projects, because of its emphasis on real-use cases etc. Also, it fell somewhat awkwardly between clustering activities and EINFRA. Some of the issues were addressed in the Work Programme and it is advised that there will be further relevant opportunity in the Work Programme in the context of EOSC. The incident is worth mentioning because of the way it was dealt with: concerns were raised, the matter was taken to the Programme Committee, the concerns discussed, the outcome reviewed and a decision made - that is good process. The leadership of EINFRA and its associated part of INFRASUPP from DG-CNECT (the Directorate General for Communications Networks, Content and Technology) places the management of the activity within the context of those considering the wider impact landscape. This has consequent disadvantages in term both of some separation from the rest of the RTD programme, and also a reinforcement of a sense of difference between the ICT research base and much of the rest of the research infrastructure community. The integration of H2020 RI between DG CNECT and DG RTD consumes considerable time and effort from both parties, but this is primary due to the difficulty and complexity of the task, and the personnel involved work hard to ensure good management at the interface. Significant policy actions are in collaboration between the DGs. For example: shaping global data sharing and data management policies, and working to spread this type of policy in the Member States. There will be the need for greater interaction still as the issues around computing and data infrastructures continue to become more pervasive, and even closer working will be required, including joint calls in some areas. So far in H2020 RI an effective programme which meets the objectives is being delivered broadly within the budget. There are clear signs, discussed in more detail elsewhere in the report that without major increase the budget will be unable to sustain an ongoing programme to address future infrastructure needs unless there is greater uptake of responsibility for the operating costs of mature new infrastructures by member states and member-state funded international bodies. There is a particular urgent case wrt HPC and computer networking. The activities funded under H2020 RI are finding no difficulty in attracting good people to participate in and deliver the funded grants. There is, however, an ongoing issue with participation from the new member states: the great majority of current European infrastructures are based in or led from North West Europe. This is discussed further under cohesion issues in section

35 7.2. Programme's Attractiveness Continuity and prioritisation Large international infrastructures can take a very long time to develop - there are infrastructures within the H2020 RI portfolio that will not be operational before 2030 and infrastructures reaching landmark status on the ESFRI Roadmap that were first initiated in the 1990 s. Every stage from design to operation can be difficult, particularly if there is no previous model to follow. They are also very expensive and, in the case of centralised facilities, their location can be a political and economic issue as well as a scientific one. Consequently, that processes have in general been quite slow is inevitable and not itself a cause for criticism. It does however stress the need for continuity on scales longer than individual Frameworks and, in this regard, the level of continuity from FP7 through to H2020 RI has been very good. As large European infrastructure proposals develop there is coalescence of scientific objectives between regional and national communities and of strategic priorities between MS. It is clear from the INFRADEV call, and from the latest ESFRI Roadmap that it has now been recognised that decisions and priorities have to be made, and perhaps earlier. The incentive to do this came from Research Ministers and it is telling - and of great credit to ESFRI members, Commission officers and the scientific communities - that difficult decisions have been made without extensive dissent. This robust approach must continue, with a clear focus on identifying and incubating a forward-looking core of the most strategically important investments. This does not mean that resources spent developing concepts and building communities to develop infrastructure ideas that may not proceed as originally envisaged are in any sense wasted. From this point of view the aspects of INFRADEV /15 aimed at clustering and cross-working between related infrastructures are very important. These recognise, inter alia, the importance that the Open Data agenda has as a unifying approach to the provision of data services across scientific research. Notwithstanding the long-haul nature of much infrastructure development, as mentioned elsewhere, the EC has shown in the proposed targeting of INFRADEV 4 to EOSC that it is capable of rapid response to urgent strategic opportunity Other Issues Related to Efficiency Sustainability Some of the pressure for the EC to provide operational support arises from the oftenused but not well-defined notion of sustainability - essentially the idea that an infrastructure, when working effectively, should be able to perform the relevant science, satisfy user demand for its services, and make reasonable provision for maintenance and development, within an income that meets or exceeds its expenditure and is sufficiently stable to allow for staffing continuity and forward planning. There is an underlying assumption that long-term sustainability is a role for member states and that the role of the EC is as a catalyst to enable the necessary route ways to this. Much effort has been invested in FP7 and H2020 RI in trying to move developing infrastructures towards sustainability, both by developing policy tools (e.g. the ERIC - European Research Infrastructure Consortium 27 - model) to help frame the necessary cooperation and funding agreements, and through individual projects addressing the issue as part of their objectives

36 There is no one-size-fits-all recipe for achieving sustainability, given the hugely varying nature of infrastructures, the differing financial practices in host organisations and member states and the fact that the economic situation - global or local - appears to change at a faster rate than the time taken to design and implement the average large infrastructure. Large centralised experimental facilities have huge capital construction budgets, often found from completely different public sources to the recurrent costs. Virtual and distributed infrastructures have no final size, scope or lifetime, developing and changing within a general framework of purpose. E-science and data-based infrastructures draw on diverse technologies being developed by others at sometimes alarming rates, and for these robustness and flexibility is as important as stable budgets (see section 6.4) Many of these issues have been articulated very clearly in the Science Europe report on Strategic Priorities, Funding and Pan European Cooperation for Research Infrastructures in Europe (Jan 2016) 28 which gives a framework of principles for infrastructure development and funding that very much reflects approaches advocated by e.g. ESFRI and CoPoRi. DG RTD currently has a consultative discussion underway on this issue, with a survey completed 29 and a meeting to be held in November The implication of these analyses is that the current practice, e.g. in INFRAIA and some EINFRA calls, of asking proposers to briefly describe measures to address sustainability is unlikely to be achieving very much. Particularly because there is little means of ensuring that whatever tentative suggestions they include are actually delivered. DG CNECT has actually dropped it from 2016/17 calls. The development of infrastructures from starter proposals onwards needs to involve a parallel and integral consideration of the eventual operational phase and its context and how to move towards it. Consideration could be given to providing some form of workshop, training or supported guidance to successful consortia to incentivise this process. As mentioned before, the sustainability problem is particularly pressing in the key area of High Performance Computing (PRACE) and networking (GEANT). The sums contributed to running the component parts of these by member states vastly outstrip the sums contributed by the EC, but the EC money is the glue that holds the infrastructure together. During H2020 RIs there has been a succession of large awards from EINFRA, and latterly from INFRADEV, to these infrastructures, but no evident sign that they will move over to the landmark stage where their long term stability is primarily ensured by member states Moving forward to the programme in 2018/20, there are serious conflicts of priority beginning to develop in EINFRA. The work programme 14/15 is very much focussed on consolidation of generic e-infrastructure services, 16/17 is additionally picking up on the integration of e-infrastructure which is more community-specific (with some similar investments in e.g. INFRADEV) and on enabling infrastructure such as the EOSC, and 18/20 seems likely to be expected to invest further in this latter general area. It is unclear how room for these additional developments can be found within a constrained budget unless some means is found to ensure the absorption of developed infrastructure into the ongoing provision by MS. This issue will become more pressing as requirements are now getting so large in some areas that the assumption that pan-european aims can be met through the federation and development of existing national and regional resources are not necessarily valid. Exascale computing is a possible example of something so large that few or no individual member states will be able to develop sufficient capacity even for themselves. It has been suggested that this be funded directly by the EU - if so, it is important that it does

37 not signal a change of the role of the EC from leading and catalysing the development of new European infrastructures to being another funder/operator. This is not a desirable outcome: the catalytic role should be maintained and means to secure sustainability commitments from member states found, although this could possibly be incentivised by strategic investments in construction and commissioning etc with suitable agreements in place for the future operation. One possibility for exascale computing is to seed provision within an existing or new European laboratory with longer term operating costs funded by national subscriptions Fragmentation and Convergence Input from the EINFRA programme and from e-irg highlighted fragmentation (technological, geographic, thematic) and overlap and duplication (the same issues being addressed, differently or in the same way, in different centres); weak collaboration among e-infrastructure projects; multiple governance structures; competing regional / national / European interests etc. These are in fact are all features of the general research landscape and certainly not unique to e-infrastructures development. In most research contexts they are to some extent a necessary consequence of a generally beneficial system in which competition, a plurality of opportunity and a diversity of approach creates a vibrant and successful science community. The Open Science agenda offers the prospect of greater effectiveness within this system by promoting e.g. greater sharing and re-use of data. This system does to some extent work in the development of large research infrastructures - astronomers and nuclear physicists have retained effective competitive research whilst maintaining substantial dependence on successions of large centralised facilities; materials and life scientists in academe and industry successfully influence policy decisions on the provision of synchrotrons etc. But these relationships have developed over a considerable time: the e-infrastructure landscape is changing much faster and it is clear that the progress to transnational coherence is going too slowly in Europe. That pace is not going to be altered solely by offering research money at the EU level. Apart from the size of the overall investment needed being too large relative to the sums available in EINFRA, the considerations of industry, commerce and society are insufficiently integrated in this conventional research-led programme context. Fixed-term project led-funding is not a good as the sole tool for dealing with this area. In some areas, e.g. data services, there is still very little coherence and it could be argued that the most significant EC investment has been to drive policy through its sponsorship of the Research Data Alliance, and to lead community behaviour through the Data Management Plans Pilot and through training, although this would underestimate the role that H2020 RI funding has played in building e-science communities and integrating these within the in the data generators and users. In the light of the above, a body analogous to ESFRI, able to speak effectively to both MS ministers and the EC on the issues in e-infrastructure and to produce a properly-owned roadmap to guide a mutually-supportive framework for investment is urgently needed. This is not going to be achieved by the efforts of EIRG, as currently structured, despite the quality of its output, because it lacks any significant standing wrt ministries and funding agencies in MS, and because it lacks the wider participation from stakeholders beyond the public-sector research community who are vital to developing a European wide strategy for computing infrastructure. The complete merger of EIRG and ESFRI would not be desirable because of the very different stakeholder bases Sustainability of Data-led, Distributed and Virtual Infrastructures Computer-based research data management infrastructure enables distributed or virtual infrastructures - which depend on the ability to remotely access and integrate facilities or data at different locations - to exist and be accessed. Such infrastructures feature 35

38 significantly on the ESFRI Roadmap, and in FP7 and H2020 RI funding, with an increasing harmonisation of tools and platforms enabling clustering of infrastructures in related or complementary areas. The Open Data agenda and the availability of these e-science based infrastructures is a seriously disruptive development across large areas of science. It impacts on almost every aspect of how research is done, by whom, how achievement is recognised and rewarded, how outputs are owned and rights protected etc etc. Amongst this miasma of change is the issue of how the infrastructure which is going to be sustaining all this new science is to be paid for. A Science Europe report: Funding research data management and related infrastructures (June 2016) 30 more or less demonstrates that, as yet, funding agencies and infrastructure providers have little idea how to effectively fund the development and operation of such infrastructure. This is not surprising, given the very rapid rate at which the field has developed in response to emergent enabling technologies. International and national models for funding operational costs of research infrastructures that currently exist mainly address centralised facilities (large machines) and are of limited relevance. Although user-pay models exist for some databases and possibly for compute access they are relatively rare and usually relate to a single relatively well-defined specific offering. A fully developed data-led distributed infrastructure is likely to be offering e.g. thousands of heterogeneous databases and analytical and management tools from a very large number of sources to tens of thousands of users on a continuous basis. This renders any form of resource-specific recharge model to individual users very difficult. However, there needs to be some resource or reward link back to the providers. This is made more difficult by the way that the service offering will be continuously changing as new tools and data are added and the use of others declines. The user communities are likely to behave like shoals of herring, rapidly changing in large numbers from one data source or set of tools to another as better data and new analytical approaches become available. Any funding model has to cope with these rapid changes in need whilst giving some stability and continuity to the infrastructure as a whole. The advantage of distributed infrastructures - that they can be bigger/better/faster/more versatile/adaptable/efficient by processes of gradual change rather than by successive construction of new ones - carries large problems with it. The extreme level of dependency that they will eventually create within an international user base poses considerable governance and associated social and ethical challenges (c.f. the current issues around pervasive social media platforms) and the need to maintain continuity of access imposes potentially severe burdens on funders. Further difficulty is caused by the tendency of funders to operate within disciplinary silos. Put simply, e.g. astronomers and earth scientists and the organisations that fund them know they need big money to build telescopes, ships and spacecraft and have established relationships between themselves and parts of government to develop and consider cases for this. Disciplines historically with low cost requirements (eg parts of the humanities), with a tradition of individual researcher independence (eg life sciences) or which operate with a close relationship to industry and innovation (eg engineering), do not have the relationships in place to easily address how to deliver future data-led public research infrastructure and view the prospect with some trepidation. The EC has been at the forefront of discussion and innovation in many areas impacting on this general issue through being an early funder of developing distributed infrastructures and leading international consideration of the Open Data agenda. It has

39 the opportunity - and a responsibility - to catalyse effective discussion to try and clarify the routes to ensuring early solutions to the funding issue. It also needs to do this without appearing to relieve MS organisations of the prime responsibility for addressing it. The forthcoming DG RTD RI Long term sustainability Stakeholders Workshop 31 is a welcome and very important move and should be the start of a major effort to foster strategic thinking in this area Suitability of the financial instruments The existing financial instruments appear to work adequately for most purposes that the H2020 RI specific programme seeks to fund. There is little use of the provisions that exist for e.g. loans, and it is possible that this might prove a more useful tool in addressing the sustainable funding of operating costs by MS than it currently is. For infrastructure development, and the stimulation of innovation from it, more could be done to facilitate the role of TTOs which sit at the interface between university and national labs and the private sector and the available instruments may not be optimal for that. Complaints that the standard funding model is difficult for small and unusual proposals (studies, workshop programmes etc) have been noted. No evidence has been seen of good things not happening because of this, and INFRASUPP appears to be supporting a good programme of policy related activities and international linkage development. 31 LONG TERM SUSTAINABILITY of Research Infrastructures Exploring RI's full potential - Stakeholders Workshop =none 37

40 8. COHERENCE 8.1. Internal coherence As explained previously, the component parts of H2020 RI together comprise a wellfigured and complementary set of intervention areas which have been developed over a long time and work together very well: large infrastructure conceived and developed through INFRADEV, new ideas and networking of existing national and regional infrastructures via INFRAIA, the core underpinning e-infrastructure through INFRAIA and measures to support training, international outreach policy development and other supplementary measures, the whole forming a comprehensive and interlocking package. A more coherent approach to innovation needs to be added to this but otherwise it is a serviceable and effective approach and has shown itself to work very well indeed. This has not, however, led to a comprehensive approach in all areas. Specifically, the EINFRA programme is not addressing significant aspects of the infrastructure requirement for the European Digital Open Market. In the majority of other contexts the H2020 RI programme resonates well with other H2020 programmes, in that it provides large numbers of scientists in a whole range of disciplines with tools and resources to enable their research. Figure 3 (based on the Likert scale assessment of an internal EC questionnaire) shows that there are no overlaps with the different actions under H2020. Figure 3 Internal coherence of RI with other EU policy programmes Internal Coherence of the actions implemented A COMPREHENSIVE APPROACH TO EUROPEAN DIGITAL INFRASTRUCTURE EINFRA has so far shown a broad footprint with many agile initiatives but is not comprehensive, having some strengths but lacking all the components of a sustainable backbone. There is a need to design a global and comprehensive vision focused on the main initiatives needed to realise and sustain the European Digital Open Market. This 38

41 needs much stronger coordination with other parts of H2020 and particularly the ICT specific programme. A key driver for this is security: put simply, at the moment the majority of cloud computing requires reliance on facilities located outside Europe, and only small elements of the microelectronics capability is in Europe, disadvantaging the EU in terms of future leadership in innovation and creating very serious long-term vulnerabilities in the security of data and the access to key elements of technology. Four main European Research Infrastructures are needed for building the European Data House: Microelectronics, 5G technologies, Cloud infrastructure and High Performance Computing (HPC) are the four building blocks of the foundation. Two Important Projects of European Common Interest (IPCEI) initiatives under construction, IPCEI Microelectronics and IPCEI HPC, can underpin the main e-infrastructures and start to design a global vision leading to big data enabled applications. Eight needs can be identified: a) Build a foundation based on microelectronics: Chips are the enabling infrastructure for the Cloud, big data and HPC. These chips are complex and are the essential building block of fully controlled systems devoted to data storage and data processing. The first role of the foundation e-infrastructure is to ensure European industrial sovereignty which requires a European approach leveraging, in a coherent way and along the full value chain, the different national initiatives of MS with an appropriate manufacturing and industrial capability and appropriate anticipation with RTOs. Europe can only be successful with leadership in innovation. This means first in R&D and first in the full chain, starting with semi-conductor infrastructures establishing manufacturing and research capabilities, going through HPC, Cloud and Big Data infrastructures enabling services and data driven application. The eri and EII (European Industrial Infrastructure) involved in the foundation infrastructure in microelectronics should include the major competitive technologies of Europe: FDSOI, Microcontrollers, Power Semiconductor, Smart Sensor, Compound Semiconductor, and support for the constellation of innovative technologies in SMEs and start-ups. b) Capital expenditure (Capex) is critical to maintain leading edge eri to prepare the future. Research Technological Organizations are committed to innovate and to transfer innovation in the European industry to ensure their competitiveness through innovation; they are part of the pillar 2 of the HLG vision, filling the gap between academia and industry. Successful existing eri operated by RTO or new complementary eri need public support to be able to invest ahead of industrial needs. Their positioning require a steady level of public investment in cutting edge equipment and prototypes to make proves of concept at the right scale before setting up strategic partnerships with industry and before leading technology transfers. The subset of principles to consider before investing in equipment in existing or new eri to prepare the future should be: - No duplication with existing eri: invest the right money - Coherence with the local ecosystem : invest in the right place - Sustainable plan : invest in the long term c) Build the data pipeline for public and private exploitation based on convergence of HPC, Big data and Cloud computing: HPC infrastructures and associated facilities are today mainly operated by governmental and academic institutions. In the private sector, only a few large industrial companies can afford operating their own in-house HPC systems. Furthermore, HPC applications are generally designed for modelling and simulation of complex physical and quasi-physical systems. For example, HPC is used for weather and climate prediction, in the military and nuclear 39

42 energy sector, the aerospace, automotive and pharma industry. HPC Centres are operated, for obvious reasons, by high qualified personnel, limited in number and often inaccessible to the majority of private companies. However, the world of High Performance Computing is changing. Driven by the Internet of Things and Big Data, the scope of HPC is expanding in a way that previous definitions with a limited focus on modelling and simulation are breaking down. Big Data applications are driving HPC in the Cloud, and in parallel, a democratization process to make HPC more accessible to private companies has started. The convergence of HPC, Big Data and Cloud Computing is enabling complete new application fields for new user groups wanting to transform the possibilities offered by these new technologies into real business opportunities. Online services and user generated data are driving the emerging digital economy. However, transforming raw data streams into information with high economic value requires easy access to computing and storage facilities combined with new types of Big Bata enabled software applications. Handling and manipulating huge amounts of data volumes, in an off-line or in a real-time mode, generate high IT workloads in a way that standard enterprise data management systems can t deliver the required computing power anymore. Data driven applications are changing the HPC landscape. It is observed a transition from HPC as a monolithic technology towards an essential tool required for supporting digital business processes and data management challenges. The crucial role of HPC in a data driven economy has been well understood by global companies and nations. USA, China, Japan and India have declared HPC and Big Data a strategic priority and fund large programs to develop next generation HPC-BD ecosystems. d) Build a public European Open Cloud: Next generation HPC and Big Data systems are designed to run in the Cloud to guarantee access to a wider range of users. In addition, cloud computing enables the convergence of HPC and Big Data. By using stateof-the-art data management technologies, it will become possible to run parallel algorithms on cloud platforms and when applied to Big Data applications, the analytics can be far more complex, taking advantage of multiple parallel nodes and accelerators interconnected via ultrafast local and international telecommunications facilities and networks. In a well-designed HPC-BD ecosystem, it will be possible to bundle distributed resources, share capacities, balance workloads and share applications. The convergence of HPC, Big Data and Cloud technologies will lead the way to High Performance Data Analysis (HPDA). Bundling HPC technologies together with Big Data applications to form HPDA will significantly increase the overall performance of these powerful tools. Today, the only option for a nation wanting to use HPC technologies is to invest in its own local infrastructures. Larger nations have the necessary means to realize and maintain this sophisticated kind of installations, but smaller countries have difficulties for operating such systems. A federated European environment has the potential to deliver HPC-BD resources in a more cost effective way. Open access under fair and equal conditions to a European cloud based HPC-BD ecosystem has the potential to create a level playing field and to put all European countries on a same level regarding the access to essential IT resources required for the digital transformation of their national economies and to be prepared for the Digital Single Market. On-demand access to virtually unlimited resources opens the way for private companies, administrations, researchers, scientists and entrepreneurs to new, fantastic business opportunities because they no longer have to restrict their ambitions and activities to the actual size of in house available IT infrastructures. In the future, the importance of an HPC installation will no longer only be measured in double precision LINPACK performance benchmarks, which only measures the capacity to perform a huge amount of calculations. Instead, achieved results in important socioeconomic domains will be used to measure how much progress has been made, for example how much HPC systems have contributed to the development of new treatments 40

43 to cure terrible diseases, how quality of life has improved or how many new jobs have been created in a specific timeframe. e) European coordination and leverage of ERI funding with critical mass. As the second phase of e-infrastructures should be more focus, directed to 4 main infrastructures enabling data driven services, the funding framework need also to be more strategic and more bundled. Combined funding between MS and Europe, as well as cross border funding have to be promoted. f) Combined funding between MS and Europe to sustain eri. Conditions to succeed in transforming research results into economic value are to support innovation with eri, RTOs and to leverage anchoring initiative with major industry players. It can be seen that investing in dynamic ecosystem connecting research, education, industry and public authorities are the key factors of success in the global worldwide map. European ambition will consolidate European ecosystem driven by sustainable innovation roadmap involving industry leaders, experienced RTO with track record in transferring innovation into industry, with clear sharing of the risk and of the value. g) Cross border funding leveraging European support to encourage seamless innovation and critical mass, in the respect of local research players and in agreement with local and national policies of the host MS. When an ERI brings its expertise to a company located in another MS, the contribution of the ERI to the growth of the industry of the third MS should be eligible to the European funding. The subset of principles to consider for leveraging European funding in case of cross border cooperation between an eri and a private company from a third MS should be : - Agreement with the local eri : no competition but complementarity - Coherence with the local and national policy : reinforcement at the EU level - Request from local industry : added value for economic growth h) Create a new European Body with a dedicated funding line to coordinate and monitor combined funding and cross border European funding directed to eri. This new body will label existing relevant eri and new ones, align the different roadmaps of eri to build a coherent vision, and allocate funding according to the global coherence of a competitive Europe. A manifesto for this body is attached at Annex B. After wide-spread local initiatives, a selection phase based on common criteria is recommended to drive the funding of the second-phase of the eri Program towards the most active, structured and relevant eri. SOCIAL SCIENCES AND HUMANITIES The H2020 RI programme includes support for major level social science and humanities infrastructure in both INFRADEV and INFRAIA. CLARIN (the European Research Infrastructure for Language Resources and Technology) 32 and DARIAH (a pan-european infrastructure for arts and humanities scholars working with computational methods) 33 - both ESFRI Landmark projects - are both making initial steps in attaining global cooperation with more and more countries interested in joining, something that could not have been possible without FP7 and H2020 support. This is innovative in cultural and linguistic infrastructures in the sense that it showcases what Europe has to offer internationally. It also triggers off possible new research connections with countries such as China etc. thus providing the building ground for both European and international impact. Among the emerging ESFRI RIs there is E-RIHS PP - The

44 European Research Infrastructure for Heritage Science (HS) Preparatory Phase - which aims at the preservation of the world's heritage by enabling cutting edge research in HS, but by combining RIs in HS with those in the domain of natural heritage and research within certain areas of the natural and biomedical sciences. This example again shows a high level of potential for social and even economic impact. RIs of the kind mentioned above can be seen as incubators for interdisciplinary research and provide the foundation for learning and gaining intercultural knowledge. They are also obviously of value inter alia to those in the Commission concerned with matters of culture and heritage. There is a growing role for ethicists as well as specialists from other disciplines, such as Anthropology, in the development and delivery of some research infrastructures and their eventual relevance to society. The report Assessing the projects on the ESFRI Roadmap 34 commissioned by DG RTG in 2013 included the examination of ethical issues relating to every infrastructure. Since then, one could claim that on a general level Ethics/Bioethics is an area to which more attention is paid at various levels. This includes issues of very serious societal importance around, for example: the access and use of patient data; ownership and rights over digitised items of cultural heritage (text, images) the issues of working and data transmission across borders in scientific areas with divergent legal and cultural research frameworks (eg human embryo research, technologies with defence applications); the issues around consent and permission when working in developing countries and with other cultures. However having examined the questionnaire relating to ethical and social issues used as part of the project application process, and some panel members having had past experience of its use in evaluation, it is not clear that significant ethical issues within proposals are being properly and systematically captured or that the grant holders are being effectively tasked with addressing them. This is important when so many proposals involve widening access to data, tools and technologies that have potential ethical and/or security risks that may not have been anticipated by the originators or be understood by the users. It is recommended that the process for examining ethical issues and assessing risks is reviewed, perhaps to provide more discipline-specific issues to be addressed in the questionnaires and their assessment External Coherence ENGAGEMENT OF ALL MEMBER STATES IN THE USE AND PROVISION OF RIS Earlier, attention was drawn to relatively low levels of participation of newer member states in the H2020 RIs specific programme. For some considerable time ESFRI devoted time and energy to thinking about ways of widening participation, including the articulation of the concept of Regional Partner Facilities - effectively outstation/access centres to major facilities located in other parts of the EU. Some of this focus was because of the identification of Supporting the development of regional research-driven clusters, and unlocking the research potential in the EU's convergence and outermost regions as objectives for the Capacities component of FP7, an emphasis that is reduced in H2020 RI. In any case, any implication of positive measures to favour proposals on grounds of regional development is so contrary to the excellent science ethos as to be impossible. In 2013 DG RTD and DG REGIO collaborated on an effort to promote the possibility of accessing the Structural Funds to provide investment into local infrastructure related to 34 Assessing the projects on the ESFRI Roadmap 23sept13.pdf 42

45 RIs on the ESFRI Roadmap, in the context of Regional Smart Strategies - i.e. on the basis that the relevant infrastructure was an investment consistent with the realisation of a specific strategy to develop particular capabilities for the region concerned. 35 The relevant strategies were due to be agreed in 2014, the process was completely unfamiliar to the majority of the scientific community (and alien in its conception and operation), and many countries had well established systems in place to manage these funds that had been addressing their submissions for some time. Consequently all aspects of this process were hurried, hard to understand and even harder to do, and no very clear picture of the outcome is available. Anecdotally, results appear to have varied greatly between countries, with some successes in the sense that funds have been used in some places to contribute to important scientific infrastructure, but not in a way that suggested any particularly systematic approach. The most common comment from scientists that have become involved is that the process lacks transparency and lacking any sense of a European dimension. In the circumstances it seems obvious that this process was too hurried and too late. Nevertheless, the Structural Funds do represent a genuine opportunity for synergistic benefit to both European research infrastructure and regional development. The next round of Structural Funding begins in 20020, and it is recommend that DG RTD, DG CNECT and DG REGIO develop a more systematic plan to inform those who can bring this about as soon as possible so that local and European viewpoints can be focussed together appropriately. Figure 4 (based on the Likert scale assessment) explains to what extent this intervention is coherent with many of other EU interventions, but is showing a specific need (gap) on the complementarities and synergies with ESIF and EFSI funds. Figure 4 External coherence of RI with other EU policy/programmes

46 9. EU ADDED VALUE This section is intended to address value that has arisen from the investments through the programme that are over and above those that would arise from interventions at national or regional level. The overall focus of H2020 RI and the nature of its major interventions is such that the whole programme is seeking deliver added value of this kind, and much of this is described in the preceding chapters. Here we try to capture the major added value characteristics of H2020 RI, their importance and where they might be improved Design and development of large European and global infrastructures The INFRADEV programme, in close conjunction with ESFRI, has enabled the EU to be effective in conceiving and delivering large research infrastructure projects at the European sc and global scale. These would not otherwise have been realised because of their large size, cost and complexity, which has required an EU-wide common vision and the combined efforts of several member states to initiate them. The ESFRI Strategy Report on Research Infrastructures/Roadmap 2016 lists 29 such infrastructures that have reached the landmark (implementation) phase, and another 21 in development. These include world-leading infrastructures across all the disciplines of science. All are potentially open to all EU members states, and many are attracting participative interest more globally. The efficiency gain here is not just in the provision for better and larger scale scientific activity that makes better use of European-wide resource and expertise. The ESFRI-led prioritisation process itself provides a framework that obviates the need for complex international discussions to begin afresh for every new concept. Models for international agreements, particularly the ERIC, developed by the EC, provide the structure to better progress those discussions towards implementation. Several of the newer pan-european RIs on the roadmap are distributed - that is to say, they operate through the integration of components at many different sites through the use of e-infrastructure, bringing access to data, tools and facilities to European scientists wherever they are. Key pan-european e-infrastructures in computer networks and highperformance computing and - in the near future - the European Open Science Cloud (EOSC) are key enablers of this, 9.2. Widening access and improved integration Substantial effort in H2020 RI is devoted (mainly in INFRAIA and EINFRA) to widening access to existing national and regional research infrastructures, improving their interoperability and integration and encouraging coordinated development of further capability to address problems at the Europe-wide level. These interventions improve effectiveness by opening resources to wider European access and application, and enable synergy by enabling the articulation of common standards, cooperation on training etc. They also represent the major contribution to the key indicator for the achievement of the ERA (number of RIs opened up to Europe) referred to in Box 1 (section 5.2.1). Some concern has been expressed in section about the relative level of investment into advanced communities (those with an already-established level of European working). One reason for this is that although these can achieve very effective collective focus, being well established, they appear to lack incentive to achieve the efficiency of resourcing, pooling and coordination that can be achieved through EC funding of starter communities" in consortium development. This is because it is, to some extent, easier to build a more economically rationalised and sharing approach into a consortium at the 44

47 outset than it is to impose it on well-developed individual national infrastructures. A similar issue, on a larger scale, is one of the challenges faced in constructing and sustainably operating the large distributed infrastructures on the ESFRI roadmap in an integrated way Policy development and international engagement Relative to the investments in infrastructure development, access and deployment the funding for policy activity and international outreach in H2020 RI, mostly in INFRASUPP, are relatively small but add considerable value. H2020 RI is the only major international collaboration addressing Research Infrastructure on a broad scale. As such it is increasingly attractive to third countries, and there have been interventions of many different kinds intended to explore and develop infrastructure links with different countries - the G7, BRIC, the developing world etc. Reciprocal arrangements with third countries have been enabled to give the EU access to expensive facilities elsewhere, and collaborative links developed to achieve global approaches to global problems such as climate change a food security. Some of this activity has moved into what is sometimes described as "scientific diplomacy" - scientific cooperation as a route to the affirmation of intent for wider friendly international relations. Large fundamental science infrastructures are particularly good for this because they provide a specific and tangible focus for negotiation, can involve significant numbers of scientists in interactions, and can offer benefits to all parties. They often involve great scientific challenges around e.g the fundamentals of the universe or life that are of interest to all people regardless of politics etc. However, this is almost never the primary reason for the collaboration being established, and they must be based on scientific excellence rather than political expediency. Policy development for internationalisation of research infrastructure has of necessity been an area where the EC has developed an international lead, and the ESFRI process for developing and delivering collective priorities is internationally admired. The H2020 RI programme has also played a major role in the development of the Open Science and Open Data agenda worldwide, largely through its sponsorship of the Research Data Alliance and the role played in taking its message to the G The contribution to mobility H2020 RI plays a major role in promoting research mobility, within the EU and more globally. This is not just a consequence of the movement of scientists to work at different sites - although this is itself very considerable. The synergistic development of common standards, research protocols, tools and platforms means that they are engendering a greater portability of skills, data and knowledge across the European scientific community, and disseminating these both through face-to-face training and e-learning The importance of the data pilot and EOSC The networked provision of computing infrastructure and the development of major datadriven research infrastructures is realising the reality of a laboratory without walls in which not just scientists but policy-makers, businesses and society generally can access and integrate data and access research knowledge across the EU, wherever they are. The single greatest effectiveness, efficiency and synergy gain from H2020 RI - and from H2020 overall - is from opening as much of the data from EC-funded research to the wider scientific community as possible, thereby potentially e.g. maximising access and usage and reducing unnecessary replication. Various things are key to realising these benefits. The Open Data Pilot - in which award holders are encouraged to include proposals to release data as part of their H2020 project - needs to become a standard requirement. EOSC needs to develop in order to 45

48 provide the best framework for EU scientists to share, access, manage and interoperate data. The development of the convergent cluster process to aggregate data-led research infrastructures into appropriate umbrella organisations needs to continue. Effective guidelines for private sector usage and public/private partnerships need to be developed Examples of added value from H2020 RI The following three projects illustrate some of the the wide diversity of added value benefits from H2020 RI interventions Square Kilometre Array: Infrastructure Detailed Design for SKA Phase 1 The Square Kilometre Array (SKA) is an international project to construct the largest and most sensitive radio telescope ever conceived. The SKA will use hundreds of thousands of radio telescopes receivers, in three unique configurations, which will enable astronomers to monitor the sky in unprecedented detail and survey the entire sky thousands of times faster than any system currently in existence. The SKA telescopes will be co-located in Africa and in Australia. The objective of the IN-SKA proposal is to support the implementation of the SKA, enabling the start of construction in 2018 and delivery of first science around In addition to driving innovation in many areas, and supporting the SKA as an inspirational vehicle for skills, industry and outreach, the ultimate deliverables of tender documentation ready for construction will place European and other national industry in leading positions to benefit directly from implementation of the SKA in CREMLIN: Connecting Russian and European Measures for Large-scale Research Infrastructures CREMLIN aims to foster scientific cooperation between the Russian Federation and the European Union in the development and scientific exploitation of large-scale research infrastructures. It has been triggered by the recent so-called megascience projects initiative launched by and in the Russian Federation which is now very actively seeking European integration. The proposed megascience facilities have an enormous potential for the international scientific communities and represent a unique opportunity for the EU to engage in a strong collaborative framework with the Russian Federation. The proposal is a first and path finding step to identify, build and enhance scientific cooperation and strong enduring networks between European research infrastructures and the corresponding megascience facilities to maximize scientific returns. The proposal follows the specific recommendations of an EC Expert Group by devising concrete coordination and support measures for each megascience facility and by developing common best practice and policies on internationalisation and opening. 46

49 EOSC - The European Open Science Cloud Pilot action Research Infrastructures such as the ones on the ESFRI roadmap and others, are characterised by the very significant data volumes they generate and handle. Effective data preservation and open access for immediate and future sharing and re-use is a fundamental component of today s research infrastructures and Horizon 2020 actions. In this context, European research stakeholders make increasing use of cloud services to effectively handle such data. This pilot action will demonstrate how wide availability of scientific data and data-analysis services for European researchers can be ensured through a cloud infrastructure. It will address the federation, networking and coordination of existing research infrastructures and scientific clouds for the purpose of increasing indability, accessibility and interoperability, improving the services provided to research communities, and facilitating re-use of data by a wider user community. Trust, easy accessibility and use by researchers should be duly taken into account. Particular attention is being paid to storage, access and re-use needs for data and knowledge from Horizon 2020 projects, as well as to the needs of the long tail of science, including orphaned scientific communities. The action will build on existing infrastructures and design a stakeholder driven governance framework, with the involvement of the research user community, the research infrastructures and the research funding bodies to ensure its sustainability. Links with related national and European initiatives will be established. 10. SUCCESS STORIES FROM FP7 Among the many success stories from FP7 finished projects, the following three have been selected as illustrations of success in developing European added value from national or individual capabilities and outputs and having potential to deliver major social, economic, policy and scientific benefits BBMRI 36 Is an infrastructure for human tissue and clinical data, managing personal data security across borders and pioneering a European RI Consortium model for a distributed infrastructure Biological samples of healthy subjects or patients alongside clinical data constitute the foundation of most biomedical studies. Such samples are central to unravelling the complexity of the human body and identifying novel therapeutic targets for challenging diseases such as cancer. The scope of the FP7-RI BBMRI (Biobanking and biomolecular resources research infrastructure) project was to organise and define the structures needed to bring existing biobanks and resources across Europe into a pan-european infrastructure. The Austriancoordinated consortium comprised 54 participants and 224 associated organisations from 33 countries. The project successfully provided comprehensive information, standard operational procedures and codes of conduct for European biobanks. An access policy on human biological samples and associated data was implemented as well as an inventory of existing major population-based and clinical biobanks in Europe. Technologies and reagents were made available to make optimal use of patient sample collections. A new database for molecular methods was also introduced. One of the major challenges was the generation of an IT-infrastructure capable of linking the existing biobank data with patient registries and clinical information. Partners

50 established a publicly accessible, common web-based portal as a centralised information site for European technology resources and platforms serving the major biobanks. The portal is functional and available at the BBMRI website. A major aspect of this was developing systems to prevent the personal information of patents associated with the samples flowing inappropriately QUALITYNANO 37 A networked approach developing and disseminating protocols to ensure the safer realisation of the benefits of nanomaterials Engineered materials of sizes of nanometers are increasingly ubiquitous in applications as diverse as energy storage, electronics, paints and cosmetics. Nano-scale objects with sizes comparable to proteins and viruses can interact with living organisms with great potential for the application of nanotechnologies in medicine, as nanoparticles may be used for the transport of drugs in the treatment of diseases. However, the properties associated with their small size, which make these innovations possible, are not completely understood nor are the precise implications for human health and the environment, a matter of growing concern. QUALITYNANO has networked researchers to develop testing procedures and a European framework to assess the safety of nanomaterials and set up a transnational access programme to experts and equipment across 15 European sites. The work focuses on nanomaterial synthesis, labelling protocols, presentation to living systems and characterisation of the properties of nanomaterials in contact with biological systems. It has included training young researchers to test nanosafety in standard and comparable ways across different labs, enabling the spread of spread of competencies and excellence and quality in nanosafety testing across Europe. This should help ensure that the benefits and innovations to be gained from the use of nanotechnologies in Europe can be realised more safely, with better attention to likely damage to human health and the environment OpenAIREplus 38 - making reality of the Open Access vision for science OpenAIREplus is the 2nd generation of Open Access Infrastructure for Research in Europe. The project capitalised on the successful efforts of the OpenAIRE project which moved rapidly from implementing the EU Open Access Pilot project into a service phase, enabling researchers to deposit their FP7 and ERA funded research publications into Open Access repositories. Work in tandem this OpenAIREplus extended the mission further to facilitate access to the entire Open Access scientific production of the European Research Area, providing cross-links from publications to data and funding schemes. This brought together 41 pan-european partners, including three cross-disciplinary research communities. The publication repository networks were expanded to attract data providers from domain specific scientific areas. Innovative underlying technical structures was deployed to support the management of and inter-linking between associated scientific data. Access to and deposit of linked publications via the OpenAIRE portal is supported by a Help Desk, and OpenAIRE's collaborative networking structure promote the concept of open enhanced publications among user communities. Liaison offices in each of the project's 31 European countries work to support the needs of researchers in Europe. The project has now moved on to leveraging its international connections to contribute to common standards, data issues and interoperability on a global level

51 For examples of successful projects from FP7 in the social sciences and the humanities see section LESSONS LEARNT/CONCLUSIONS The H2020 RI Thematic Programme is one the most important assets for European science, and in key aspects of its activity it has given the EU a global lead. The principal support components have been developed and refined, but been essentially constant, over several Research Framework Programmes, and this continuity and stability of approach has been essential in supporting the very long-term process of catalysing the development of international research infrastructures, which can take 10 years or more to the point of implementation. The catalytic role which H2020 RI plays in the early phases of the lifecycle of an international research infrastructure in Europe is unique and indispensable. It must be continued and, given the increasing scientific and policy drivers for infrastructure at the European level, the volume of this catalytic activity will need to increase. The responsibility for the implementation, construction and exploitation of research infrastructures (including e-infrastructures) is and should remain the responsibility of groups of member states and associated countries and not of the EC. Consequently, the efforts to establish a clearer shared understanding between the EC, member states and the scientific community on the pathways for sustainability of RIs (including e- infrastructures) through member state support must continue. INFRADEV The INFRADEV programme is substantially concerned with progressing infrastructures on the ESFRI roadmap. This is an extraordinarily successful partnership between the EC, ESFRI and the ministers in member states and its key components should continue. The Preparatory Phase Project funding in INFRADEV02 is crucial, and in future support should only be given to applications which include sound financial commitments from at least three member states. INFRADEV04 has supported the European Open Science Cloud pilot project (EOSC): as well as ensuring that the foundations for this are in place, attention should be given to seeding clusters. INFRAIA This element of the programme supports a variety of e.g. networked and regional infrastructures to integrate at a European level and widen access. It plays a vital role. The share of funding between advanced communities - groupings already to some extent established - and starting communities is favouring the former and involves too often repeated or supplementary support and repackaging of established partnerships. Intervention from H2020 RI is best focussed on catalysing new opportunities at the cutting edge. The balance should be shifted in favour of more support for starting communities seeking to establish European-wide or regional infrastructure access in new and emerging areas of infrastructure opportunity or need and awarded against open competitive science-led criteria Advanced community support should focus on those proposals able to show a realistic likelihood of integrated European operation being maintained beyond the period of the 49

52 award without further EC support. Overall, support for what is in essence the same consortium (or a subset of it) from H2020 RI or its predecessors should be limited to a maximum of 10 years. E-INFRA The E-INFRA programme currently plays an essential role in the support of widening European access to HPC and advanced networking and the sharing of related tools. It has also developed a global lead for the EU in the Open Data agenda. However, there is an urgent need to address capability in other key research areas that underpin the digital open market and future EU competitiveness and security. In the short term, the focus of EINFRA needs to be widened to include enabling e- infrastructures in microelectronics and 5G technologies alongside Cloud infrastructure and High Performance Computing to ensure sustainable services. In addition, there is a need to make IPCEI a reality, crucially with capital funding for public infrastructures underpinning public-private partnerships and services for scientific communities, based on an efficient data pipeline. The European Open Science Cloud can be the reference framework for this. For the long term, a new approach is needed, based on from services to public infrastructures" instead of being research-project based. This will mean combining MS-EC funding and cross-border funding. In the first instance a high-level implementation group with MS and EC representatives should be considered to set up the most efficient funding channels to support the priority needs of the scientific community and to be the foundation of a new body in charge of global coordination and EU/MS joint funds INFRAINNOV The separate INFRAINNOV focus is new within H2020 RI and is still in something of an exploratory phase. The development of a clear area of focus on innovation within the programme is welcomed. There are two main areas of focus - promoting the engagement of industries an SMEs in the construction of RIs and in the exploration of the wider potential of technologies and tools developed in the process, and stimulating companies (inc SMEs) and the public sector to become users of RIs. A clearer statement of purpose, scope and strategy for INFRAINNOV needs developing. INFRASUPP and OTHER OCTIONS The true support activities in this area - workshops, studies, third-country collaborative discussion etc - all add value to H2020 RI in very effective ways. However, a large proportion of the substantial budget is committed to a miscellany of large one-off activities, for which the rationale, in terms of the overall aims of H2020 RI, is not very clear and which sometimes seem unintegrated with the rest of the programme. Some would more suitably fit elsewhere - either within H2020 RI or in another part of H2020. Consideration should be given to whether this represents an appropriate and effective way to commit funds for science. OPERATIONAL MATTERS Overall the programme is very well run and overseen. The level of cooperative action needed between DG RTD and DG CNECT is bound to increase over the foreseeable future given the pervasive nature of the open science/open 50

53 data agenda and of computing infrastructure generally and it will be good to see e.g. more joint policies and shared calls. Mostly the programme seems very well received by the research communities. Concerns about oversubscription in some calls, lack of bottom-up opportunity for new ideas, and the lack of competition in calls procuring a single application all suggest that there needs to be a continued focus in sustaining a healthy level of competition, which gives a success rate that is neither too high nor too low. The continued support for ESFRI is essential, particularly in its role as an incubator for implementation of infrastructures. This implies that the ESFRI Roadmap should be seen as a tool for realising research infrastructures of pan-european interest, and not as a goal in itself. More should be done to encourage member states to build and implement national roadmaps. The role of EIRG as an advisory link between the IT community and the EINFRA programme management is very effective in providing technical and other advice, stimulating the community response to opportunity and in providing guidance on aspects of technical policy. However, EIRG lacks the linkage to ministers in member states that is critical to ESFRI s success and this would improve its role considerably. Structural Funds Despite specific encouragement by MS Ministers, the deployment of Structural Funds to support the development of Research Infrastructures has not been widely successful. It is important to raise the awareness and better facilitate the possibility of using structural funds for RIs, whether they are at the European level or national, to support the outreach towards European and international collaboration and the development of components of RIs on the ESFRI Roadmap. This requires long-term collaboration between DG REGIO and DG RTD/DG CNECT to raise awareness of this opportunity in good time for it to be exploited, and to enable the alignment of financial regulations (as well as time frames) between Structural Funds and funding coming from H2020. A higher level of alignment for financial inputs needs to be maintained through the different phases of the RI life cycles. Part of the problem with using Structural Funds for RIs stems from the diversity found in National Roadmaps across Europe. This ranges from countries that have no Roadmap whatsoever, to those which have finely developed ones. In the less developed Roadmaps, visible strategic priorities are lacking, and hence follows the recommendation that a certain level of mentoring in setting up Roadmaps (especially in Central and Eastern Europe) would be welcomed Relevance Key findings: The five sub-programmes of H2020 RI present a cohesive set of interventions that address the objectives of the programme comprehensively, although the part to promote innovation is relatively new and needs further development. 51

54 They are the result of a continuous refinement of successive programmes under previous Frameworks and work together well. This continuity and long-term vision must be retained. The strengths are: The EU has developed global leadership in Open Data policy because of policy work developed in FP7 and H2020 RI since The infrastructures and resource being developed are critical to the success of the Juncker Commission s New Vision for Europe 39 and the Open Innovation/Open Science/Open to the World Moedas vision 40, because they are delivering platforms, tools and skills that will enable it to happen. The bottlenecks are: The programme has been delivered in a period of unprecedentedly rapid change in science and technology, particularly as it relates to IT and data. This will overcome by: The continuity and consistency of vision in the programme, its strong advisory structures and its ability to exploit budgetary flexibility has allowed agile response to emerging opportunity e.g. in the launch of the European Open Science Cloud (EOSC) pilot in Effectiveness The ESFRI Roadmapping process, coupled with the use of the INFRADEV part of the programme to catalyse the design, development and prioritisation of RI proposals, and the partnership this represents between the scientific community, member state ministers and funders, and the Commission, is one of the major achievements of international science policy in recent years and must be maintained. The strengths are: The pioneering development of distributed European infrastructures and networked infrastructures based around the shared distribution and access to data, materials and tools has been transformative, stimulated scientific communities across Europe into cooperation and created a solid basis for EU-level research. The bottlenecks are: The EINFRA programme has had to operate in an extraordinarily challenging and changing environment and has done well to address the objectives it was set, to maintain a vision and coherence across its programme and to absorb substantial changes such as EOSC. It sits in the right place within DG CNECT with close connections to DG RTD. This will overcome by: The two roles of the INFRAIA element of the programme to develop starter communities and to enable advanced ones to address major issues are both valuable, but the balance between the two is too much in favour of the second, and the catalytic value of H2020 RI political priorities of the Juncker Commission 40 Open innovation, open science, open to the world: A vision for Europe 52

55 would be better served by more emphasis on new opportunities. There should be something similar to attract and explore new ideas from the community in EINFRA Efficiency Key findings: The H2020 RI programme is well managed by Commission officials and the two teams in DG RTD and DG CNECT achieve a good level of overall coherence despite the considerable challenges involved. The Programme Committee performs an effective and appropriate oversight of the programme. The strengths are: The consideration of routes to sustainability for virtual and distributed infrastructures is very important because there are no established models. Some of the communities have no familiarity of working with, or tradition of resourcing, such infrastructures and such concepts as recharging are little understood for intangibles such as data access. The bottlenecks are: There is particular concern that a sustainable settlement must be agreed between member states to support PRACE (the HPC infrastructure) and GÉANT (networking), because of the increasing sums being contributed, seemingly haphazardly, from H2020 RI and the possibility that this will seriously restrict funds for other objectives. Relatively little use has been made of the potential that the financial instruments offer for loans, etc. This will overcome by: The EC should be very wary of accepting long-term responsibility for the operational funding of mature infrastructures, and should continue efforts to find means to obtain commitment from member states to ensure the sustainability of operational infrastructures. It is imperative that the unique role of the H2020 RI programme in catalysing the development of infrastructures of the future is not lost in order for it to become another funder of running costs. It is key to the effectiveness of the programme going forward that all those involved in its management retain a sense of the need to prioritise and, accepting that to do this within a constrained budget involves a conscious decision to create opportunity for the new by discontinuing support for the longer-established Coherence Key findings: The H2020 RI programme provides a major role in furnishing the excellent science base that underpins the bulk of H2020. There are good examples of infrastructures that directly resonate with the aims of other H2020 programme areas, such as food, agriculture, health, materials science etc. The strengths are: For competitiveness and security reasons there is a major need to develop a more comprehensive approach to underpinning the success of the European Digital Open Market, with a clearer focus on the key areas of the Cloud and HPC. This is an undertaking which involves not just the current EINFRA programme and DG CNECT but many aspects of the wider H2020 programme. 53

56 The bottlenecks are: The sharing of data, tools and materials is now becoming a substantial feature of scientific infrastructure, and can entail very serious issues of security, personal privacy and ethics. The effectiveness of measures to ensure that ethical and social issues relating to infrastructure development should be reviewed to ensure that they are properly addressed in proposals and considered in evaluation. This will overcome by: The effectiveness of measures to ensure that ethical and social issues relating to infrastructure development should be reviewed to ensure that they are properly addressed in proposals and considered in evaluation EU added value Key findings: The European added value of the H2020 RI programme is evident in several ways and throughout the programme. The strengths are: The European Union as a whole is able to conceive and deliver large infrastructure projects at the European scale and develop and lead those at a global scale, of a type, size and number that would not otherwise be possible. The use of these infrastructures is a major vehicle to promote scientific mobility across the European Union. The networked provision of computing infrastructure and the development of major datadriven research infrastructures is realising the reality of a laboratory without walls in which not just scientists but policy-makers, businesses and society generally can access and integrate data and access research knowledge across the EU, wherever they are. The processes of developing and initiating pan-european research infrastructures and particularly the e-science and data sharing aspects, has placed the EC at the global forefront of thinking and policy in the Open Science and Open Data area. All of this represents greater value for money because of the minimisation of duplication and improved means of access. Reciprocal arrangements with third countries give the EU access to expensive facilities elsewhere. Bottlenecks and weaknesses: The establishment and maintenance of comprehensive access to word class research infrastructures and enabling e-infrastructure across the EU is a large, long-term project with many challenges, particularly in securing the prioritisation of pan-european approaches when national resources are very constrained. So far the EC and research ministers in MS have shown remarkable tenacity and consistency of vision in pursuing this goal, and it is crucial that this is maintained and that short term economic and other national pressures do not undermine it. The second great challenge is to put the third O - Open Innovation - alongside the open science and open data capability and capacity which H2020 RI is helping to realise. How these bottlenecks will be overcome: The persistence of purpose in pursuing the long term vision is greatly aided by the role which ESFRI plays in advising ministers in MS as well as the EC. The changes ESFRI has 54

57 made in its processes to improve prioritisation and focus in its role catalysing and incentivising the realisation of the most important RIs for the future are essential. H2020 RI and its successor must reinforce this by ensuring that the EC remains focused on this catalytic role for future Europe and and global RIs and does not become diverted into subsidising operational costs of existing research infrastructures, which should remain the role of MS. Grasping the Open Innovation opportunity is very much a role for H2020 and its successor as a whole, with European RIs and e-infrastructures providing platforms that will enable it. Actions recommended in this report that will help include the major investment in developing European infrastructure and capability in advanced IT technologies, the realisation of EOSC, the full implementation of the Open Data policy for all Research Framework project data (and the greater consideration of issues around public/private and commercial interactions, and of privacy and ethics that this implies), and the development of a comprehensive strategy for innovation arising from RI and E- infrastructure investment. 55

58 12. RECOMMENDATIONS ON THE RI PROGRAMME The H2020 RI Thematic Programme is one the most important assets for European science. The catalytic role which H2020 RI plays in the early phases of the lifecycle of an international research infrastructure in Europe is unique and indispensable. The responsibility for the implementation, construction and exploitation of research infrastructures (including e-infrastructures) is and should remain the responsibility of groups of member states and associated countries and not of the EC. In support of the catalytic role for new and emerging infrastructures the funding for Preparatory Phase projects for new infrastructures should be maintained and support for these should only be given to applications which include sound financial commitments from at least three member states. The element of the programme supporting starter communities to develop new ideas should be increased. The focus of the e-infrastructures part of the programme needs to be widened to include e-infrastructures in microelectronics and 5G technologies alongside Cloud infrastructure and High Performance Computing, taking into account interaction with other relevant parts of H2020. A clearer statement of purpose, scope and strategy for innovation-stimulation activity in the programme needs developing. The continued support for ESFRI is essential, particularly in its role as an incubator for implementation of infrastructures. The potential of eirg to play a similar role for the e- Infrastructures area would be increased by developing channels to MS ministers analogous to those that work for ESFRI. The Structural Funds offer considerable potential for the development of European-level infrastructure capability, particularly in Central and Eastern MS with less RI activity currently. Effort needs to be made to ensure that this potential is understood and actually realised. Special attention should be given to the development of national roadmaps. 56

59 1. ANNEXES A. Panel Membership and acknowledgements B. Manifesto for the future - A new body for einfra 57

60 Annex A Panel Membership and acknowledgements - Prof Francesco Profumo (Chair) - Prof Hans Chang - Dr Marie-Noelle Semeria - Prof Milena Zich-Fuchs - Dr Alfred Game (Rapporteur) Acknowledgements During the course of this report members of the Panel consulted widely with those involved in the planning, management and delivery of the H2020 RI specific programme and those delivering its science. We are particularly grateful to the following for the assistance and advice given: - Members of the Commission services in DG RTD and DG CNECT - The Chairs of, and individual delegates to, ESFRI and EIRG - Members of the H2020 RI Programme Committee - Grant holders and applicants to the H2020 RI programme - Attendees and organisers at the International Conference on Research Infrastructure (ICRI), Cape Town, 3-5 October Authors of various of the documents referenced in the report 58

61 Annex B Manifesto for the future - A new body for e INFRA November, 2016 There is a road for a bright future in Europe. Europe is first in research, our high skilled students are ranked first in many companies, European industries are leaders in automotive, aeronautics, space, IoT with microcontrollers, sensors, power electronics. We are used to work in networks and to mutualize large infrastructures. There is still an industry of chips able to insure the sovereignty of Europe. European Members States and the European Commission are believing in their future and are investing in it: as a sign of trust in the future, numerous research infrastructure including e INFRA were born in the first phase of H2020. There are more than ever tremendous opportunities of new services enabled by the digitalization of our societies, more and more start-ups are growing. However, Europe is lagging, struggling to catch the value of its huge investments, transformation is too slow. What is wrong? Compared to other continents and nation, a coherent vision is missing: European e INFRA are still fragmented, more and more often underfunded due to the decrease of national public money, impacting their sustainability, limiting the reach of their services, difficult to be accessed by SMEs and start-ups. It is time to act, to draw the global picture of einfra based on solid foundations and turned towards services to improve the quality of live and ensure European economic growth: - Foundation einfra are prerequisite for insuring an independent and sustainable backbone for services. They address Microelectronics, 5G technologies, Cloud infrastructures and High Performance Computing. - Service einfra need to address at the right scale the major societal, economic and environmental challenges. They target Research communities, Public Administration, Precise Medicine, Space, Manufacturing, Agrifood, Smart City Applications, in close connection with Industry. People are at the core: data owners, data providers and data users. E INFRA have to be complemented with an office dealing with ethic, trust, privacy and regulation. Furthermore, Large scale Test Beds are mandatory to transform innovations and new technologies from e INFRA into real-world services to test cross-border data exchange mechanisms under real-life conditions and to foster the digitalization process of industry and cities. To enable this future, a new body is mandatory to label and to finance einfra insuring the sustainability of the critical einfra and creating new einfra to fill the gaps. This body will be in charge of the coherence of the einfra policy and of its implementation considering cross-border funding and combining Member States and European funding. We call for a global vision of a data sharing society taking care of people, bridging e INFRA and technologies to industry and services. We call for a new body in charge of the coherent implementation of the e INFRA policy and of the combined funding of e INFRA with Member States. 59

62 60