SEARCH

SEARCH BY CITATION

Keywords:

  • e-infrastructure;
  • governance;
  • e-science;
  • e-social science;
  • virtual communities

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies

The paper studies the transition to ICT-based support systems for scientific research. These systems currently attempt the transition from the project stage to the more permanent stage of an infrastructure. The transition leads to several challenges, including in the area of establishing adequate governance regimes, which not all projects master successfully. Studying a set of cases from Europe and America, we look at patterns in the size and scope of the undertakings, embeddedness in user communities, aims and responsibilities, mechanisms of coordination, forms of governance, and time horizon and funding. We find that, though configurations and landscapes are somewhat diverse, successful projects typically follow distinctive paths, either large-scale or small-scale, and become what we term ‘stable metaorganizations’ or ‘established communities.’


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies

E-infrastructure – or cyberinfrastructure in the US – is a new type of infrastructure consisting of distributed ICT-based resources which support collaborative research. It has obtained significant funding at national and international levels over the past 10 years. For instance, annual funding for e-infrastructure and e-(social-)science programmes in the UK alone was recently estimated at £170 m (Research Councils UK, 2010). The European Commission and the National Science Foundation have provided funds at the level of several million per year to e-science projects such as GÉANT (40 m €) and EGEE (24 m €) in Europe, and TeraGrid (30 m US-$) in the United States. In addition to these flagship projects, there has been a myriad of medium-sized and small-scale ventures created in a multiplicity of research domains from computer science, engineering, high-energy physics, and astronomy to biomedicine, climatology, linguistics, finance, the social sciences, and humanities.

The common characteristic of these efforts is that they draw on geographically distributed digital resources – such as data, computing power, visualization technology, and storage – in order to provide services enabling the resource sharing and collaborative tools essential to research in distributed settings. Many funded ventures fitting this definition of distributed ICT-based support systems are currently attempting the transition from project-style provision to the larger-scale production quality and sustainable financing required of an (e-) ‘infrastructure.’ For these research support activities, the transition in question is in effect a massive change in their governance1, or equivalently the emergence of appropriate governance for the sustainable provision of their services to research.

The challenges associated with such governance transitions are often underestimated, with a consequent threat of failure to establish sustainable structures and a loss of public investment. Of the little relevant social science analysis, one example is the early 2008 NSF Report from workshops on building effective virtual organizations (Cummings, Finholt, Foster, Kesselman, & Lawrence, 2008). The authors define a virtual organization (VO) as “a group of individuals whose members and resources may be dispersed geographically and institutionally, yet who function as a coherent unit through the use of cyberinfrastructure (CI)” (p. 3) encompassing systems known by other names, such as collaboratories, e-science or e-research, distributed workgroups or virtual teams, virtual environments, and online communities. Useful though this approach is overall, such an all-embracing use of the term virtual organization is too broad for our purposes. Based on the more precise definition introduced below, few emerging e-infrastructures currently qualify to be virtual organizations, and we are more often dealing with networks facilitated or enhanced by e-infrastructure. A more elaborate conceptualization is given by Bos et al. (2007), who distinguish between seven types of collaboratories and highlight typical organizational properties.2 With a focus primarily on the main goals pursued by the projects studied, they propose making distinctions along two main axes: the main resources involved (tools, information, or knowledge) and the main activity (aggregation of distributed resources versus collaborative creation across distance).

In our exploration of emerging patterns of governance of e-infrastructures, we can expect to identify some patterns at the level of individual projects which reflect governance structures at a higher level. However, it should be emphasised that we have not set out to compare regimes of e-infrastructure governance at the research domain or country level, which, of course, may either drive or restrain the microlevel. Others have taken a normative stance, suggested that a legal framework for electronically supported scientific collaboration should be adopted and “argued for a more ‘bottom-up’ approach to constructing appropriate institutional infrastructures for e-science” (David, 2006, p. 447; see also David & Spence, 2003). In contrast, our contribution lies in generating empirical information on initiatives to build ICT-based support systems for research and providing a structured view on their progress towards becoming sustainable e-infrastructure (or not). In support of this, we address the following research questions in this article:

  1. How does the transition from a project to more permanent governance of an e-infrastructure proceed and how do participants secure the capacity for collective action?
  2. Do we see different types of transition to stable e-infrastructure governance?
  3. Assuming a single model of governance is not suitable for all circumstances, what kinds of governance structures are emerging?
  4. Can we see any tensions or signs of failure in transitions to stable e-infrastructure governance?

Theory

Governance in and of science

Given that the purpose of e-infrastructure provision is to support scientists in their work, the governance of e-infrastructure must be seen in the context of scientific activity as a whole. Science itself has a complex and changing governance structure, and it appears that the governance transition processes we are studying are strongly related to these, and in particular to the intersection of endogenous and exogenous governance of scientific activity, i.e. related to where “governance in science” meets “governance of science.”

As described by Gläser (2006) and Gläser and Lange (2007), governance of scientific activity has recently undergone a transformation. Traditionally, the main actors inside science have been individual scientists, who formulated research questions and communicated research results to their peers, and scientific communities, which regulated what questions were asked and what was considered as an acceptable answer, applying a broad set of governance mechanisms. Recognition and advancement being through publication, the role of gatekeepers in these formal communication channels (reviewers and editors) was a powerful mechanism for steering research activities: Scientists who made frequent or more important contributions to the knowledge stock acquired greater influence, more students and followers, and multiple gatekeeper positions.

The publication and archiving of scientific knowledge required the support of learned societies and publishers, effectively locating this major component of the governance of scientific activity outside science itself. Universities and research institutes employed scientists and provided the infrastructure for knowledge production, exercising gate-keeping power in both respects. The role of funding provider gave foundations, ministries, research councils, or even private corporations a place in the governance of science and the ability to guide the activity of scientists towards their (policy) goals. Variations in the relative power of these diverse influences gave rise to public science systems of different types, such as those distinguished by Whitley (2010): state-shared (e.g. continental Europe, Japan), state-delegated (e.g. the UK up to the 80s) and employer-competitive systems (e.g. the US).

Though much of the traditional governance structure of science has remained in place, in some respects structures have begun to change, notably through the success and growth of the scientific undertaking itself. The growth of scientific knowledge production has led to an increased specialization of skills, and the role of individual scientists in formulating and pursuing research questions has increasingly been replaced by teams of scientists or even networks of such teams (see also Jones, Wuchty, & Uzzi, 2008; Wuchty, Jones, & Uzzi, 2007). As resources have not grown at the same rate as knowledge producers, other innovations in research governance structures were implemented - such as performance-related mandates and funding, monitored target agreements, ongoing research assessments and the like, helping to steer the distribution of funds (Whitley, 2010).

The increase in the availability and power of services based on modern information and communication technology (ICT), not least the Internet, has also had an impact on the governance in and of science. In some cases, for example, the Internet has weakened traditional gatekeeping rights, by providing new routes for the exchange of the results of scientific activity (“open access”). Or again, some new facilities and services such as the pooling of distributed computing power, the sharing of primary data, remote access to research equipment, or the activation of external resources for scientific work have shown a potential to change the process of knowledge production itself, and to some extent they have already done so (Bos et al., 2007; Cummings et al., 2008; Foster & Kesselman, 2006; Gläser, 2006).

More broadly still, Benkler (2006) has described how electronic networks enable new forms of peer-production in communities (his main example is Wikipedia). Applying this notion to research activities and to the e-infrastructure initiatives we are studying, it may be important to distinguish between communities of e-infrastructure developers and (other) research users - one dimension explored below. Further, the wider deployment of ICT segments and integrates research activity in new ways; new actors are brought more closely into the research domain, with search engines such as scholar.google (Rieger, 2009) influencing the ranking and online visibility of scholars and digital libraries (Meyer, Madsen & Fry, 2010) presenting new opportunities to make data available (Borgman, 2007). As Heimeriks and Vasileiadou (2008) argue, these and other ICT-enabled changes to research do not represent a radical discontinuity, but complement and add to existing modes of pursuing and communicating research. As we shall see, this applies to the infrastructures examined here, especially as they are still in transition towards becoming established means of doing research.

The broader view of how ICT is affecting the organization of research is beyond the scope of this essay (but see Jankowski, 2009 and Borgman, 2007, for recent overviews); we focus on trends with the relevance to the issue of governance in e-infrastructure. A number of perspectives have been offered on related topics. For example, Fry and Schroeder (2009) have argued that how social sciences address the question of ICT transformations of research can itself be categorized into several different modes, including taking an active (‘advocacy’) or critical part in shaping how this transformation takes place. In this paper, we attempt to identify different modes of governance, which may then play a role in supporting ‘advocacy’ or policymaking3.

Transition to stable infrastructure provision

Star and Ruhleder (1996), in discussing general characteristics of infrastructure emergence, point to the fact that infrastructure provision emerges out of a certain setting with a certain history. They characterise the provision of infrastructure as embedded (in other structures, social arrangements and technologies), transparent (i.e. pre-existing, standardised, self-explaining, and invisible except in breakdowns), of more than local and short-term scope, learned as part of membership in a community, and shaped by conventions of practice in these communities (and shaping them in turn). They highlight evolutionary, systemic, and relational aspects of infrastructures. Others have argued that the transition of heterogeneous localised systems into more stable infrastructure provision requires adaptation and mutual adjustment in several respects (technological, social, organizational, cultural, legal, institutional, etc. properties) and the development of gateways (e.g. standards, protocols) which permit the linking of the local systems (Jackson, Edwards, Bowker, & Knobel, 2007). Along the same lines, Iacono and Freeman (2006) stress the bridging of boundaries between organizations, disciplines, countries, and the scaling up of the ventures as major challenges in e-infrastructure development. In relation to discipline boundaries, Fry and Schroeder (2009) have discussed the relationship between structures in e-infrastructures or e-research and questions of resource concentration and pluralism versus task certainty within disciplines, though the role of research technologies also on occasion cuts across these boundaries. Bany Mohammed and Altmann (2010) understand infrastructure as a public good, ubiquitous and ‘sustainable’ in the sense that it is “independent of specific funding streams” but sustained by “business models that guarantee the provision of the necessary funding” (Voss, Procter, Hewitt et al., 2007).

These points raise the larger question of what an ‘infrastructure’ consists of in the cases discussed here. No summary answer can suffice for all our cases, though it is worth noting, first that unlike the infrastructures that are normally considered as such (transport, energy, communications – which support whole populations), the infrastructures discussed here are infrastructures that support a particular part of the population, namely, researchers (Bany Mohammed & Altmann, 2010); and secondly that many (as mentioned) are transitioning from being ‘projects’ to becoming established as longer-term (infra)structures. In terms of governance, this transition and emergence of an infrastructure is in our view best described by the following six constructs.

1) Size and scope. A process of scaling up, i.e. growing activities in size and scope, is one of the major characteristics of the establishment of an infrastructure (Avery, 2007; Foster & Kesselman, 2006; Iacono & Freeman, 2006; Jackson et al., 2007; Star & Ruhleder, 1996). The providers of locally constructed and delimited systems are linked or assembled into larger networks. In the process, the number of providers and users grows and extends beyond the original areas of use. This growth, of course, has wide-ranging consequences for the governance of the system. For instance, local control systems need to be replaced by distributed but coordinated controls, and mechanisms to deal with reverse salients, i.e. the challenges, limits, and sticking points of system development need to be established (Jackson et al., 2007). An additional perspective on the coordination of the larger scale sociotechnical systems for research is that if there are long-established mechanisms within a particular discipline community, such as in the physics collaborations studied by Shrum, Genuth and Chompalov (2007), then bureaucratic organizations can overcome some of the issues of trust and governance that beset the emerging collaborative infrastructures that we are concerned with here, where socio-technical barriers against cross-disciplinary, cross-organizational or cross-national collaboration have to be overcome (Iacono & Freeman, 2006).

2) Embeddedness in user communities. Strong involvement of science communities in e-infrastructure development and deployment has been described as beneficial (Avery, 2007). Several of the properties of an infrastructure listed by Star and Ruhleder (1996) can be reduced to the question of embeddedness in the user communities. When a system has achieved this status, it will be transparent to its users and connected to other systems, structures and social practices in the communities. In addition, it will be part of the socialisation into these communities and new members will be taught its use and significance. The new support systems and applications discussed in this paper did not appear out of nowhere; they were typically conceived by scientists, adopting important elements and input from existing services (computing centres, instruments, data archives, etc.) and utilising existing contacts to funding and other stakeholders. These scientists may themselves belong to the communities of users – in some cases with support from developers and computing engineers – or they may belong to a developer community of computer scientists and engineers searching for possible uses of their tools and applications among domain scientists. In the second case, we would expect that at least at the start, the embeddedness in and interaction with the user communities is low, leading to problems of interdisciplinary collaboration between computer and domain scientists that have been described frequently in the literature (Barjak et al., 2009; Berman & Brady, 2005; Iacono & Freeman, 2006; Voss et al., 2007a). Others have stressed that the analysis of information and communication technologies in scientific practice benefits from the perspective of sociotechnical interaction networks (STINs) in which humans/organizations as well as material and electronic ‘nodes’ interact and become tightly and sustainably interwoven with each other (Kling, McKim & King, 2003).

3) Purpose and responsibility. As Star and Ruhleder (1996) stress, an infrastructure is not reinvented each time it is used. In order to undergo a successful transition, goals of and responsibility for an ICT-based facility or system of supporting science must change at the juncture of institutionalization (Mackie, 2007). Developers and providers must adapt to the change in the purpose of their activity - from the pursuit of research goals, which include the production of new knowledge, in the form of publications, software codes or artefacts – to the ongoing maintenance of a service to other scientists and potentially to customers outside of science. Within this transition and where applicable, the degree and form of involvement of the private sector also has to be decided. Secondly, and related to this, responsibility for the infrastructure must pass at least in part out of the hands of scientists and scholars into the hands of (newly constituted or existing) management bodies, which include administrators and technicians. This transition also entails achieving the best mix of centralised and decentralised decision-making and safeguarding interoperability in decentralised institutions and ensuring the transparency and accountability of resource provision and involvement of users. The necessity of providing and administrating a robust service will also result in a larger degree of automation and exclusion of human interference in day-to-day operations, a requirement noted by Iacono and Freeman (2006).

4) Elementary mechanisms of coordination. The notion of governance of interdependent activities is clearly linked to that of coordination. The discussion of the relative advantages of organizational hierarchies on the one hand and markets on the other to achieve economic coordination of interdependent activities was put forward by Williamson (e.g. 1971, 1975). Markets support the coordination of actors whenever they wish to exchange clearly described and priced goods. In an organization or hierarchy, coordination is achieved by routines, procedures and orders which are supervised closely and enforced. Others added to these mechanisms of coordination: Streeck and Schmitter (1985) distinguished spontaneous solidarity (community), dispersed competition (market), hierarchical control (bureaucracy), and organizational concertation (association). Powell (1990) and Thompson (2003) stressed the significance of networks as a form of coordination. Accordingly, in cases when the value of a commodity is not easily measured and expressed, the relationship between the organizations involved is as important as the commodity itself. It needs to be trust-based, long-term and reciprocal to facilitate successful exchange. In networks, actors are not as independent as in markets, but less interdependent than in hierarchies. The starting points of e-infrastructures are those of communities and networks (interorganizational, sociotechnical) coordinated by means of solidarity and trust. In order to make them sustainable, stronger forms of coordination (routines, procedures, and orders) need to be established, typically linked to (virtual) organizations (Iacono & Freeman, 2006).

5) Formality of governance relates to the extent to which agenda setting and decision making are structured by pregiven sets of rules or institutions and how compliance with these decisions can be enforced. Wittek (2007) describes how formal and informal forms of governance are linked to differences in the incentives motivating compliance, differences in the relationship between the actors, and differences in the source of legitimacy. Formal governance rests on material incentives (e.g. revenues or other resources, lower costs, opportunities for future development) and vertical relationships in which control is exercised by superior positions in the hierarchy or by either clients or vendors. Legitimacy is based on legally enforceable rules and routines laid down in contracts or binding agreements. Informal governance uses only social incentives (e.g. reputation, esteem), is characterised by horizontal relationships in which peers state and monitor the goals of coordination, and does not have strong means of enforcing compliance. Instead, compliance is achieved through reciprocity and reputational sanctioning.

6) Sustainability of an infrastructure is a complex issue covering several of the concepts described above (Voss et al., 2007b). At its core is, however, the certainty and permanency of the funding arrangement. In the history of infrastructure development we often see public investment at the start that is over time replaced by private funding sources (Bany Mohammed & Altmann, 2010). ICT-based support services to science are generally – at least in Europe and North-America – set up and funded by means of short-term public grants (typically not exceeding 3 years) issued by national research councils, science foundations, ministries, and international organizations such as the European Commission, plus some cofunding from the organizations in receipt of the funding (Barjak et al., 2010). Though second and even third funding rounds are not uncommon, projects need to reach beyond these short-term grants and find other funding schemes to establish themselves on a more permanent basis. It has been argued that oftentimes funding will need to come from institutional sources and projects need to be aware of this early in their lifetime to ensure smooth transition when grants run out (Mackie, 2007). Alternatively, market-based funding models can be designed and established (Bany Mohammed & Altmann, 2010).

Table 1 summarises the various aspects of the transition from project-based, potential infrastructures, small-scale support systems in a project environment, to sustainable large-scale infrastructure in a production environment.

Table 1. Two types of support system
 Support system in the project environmentInfrastructure
Size and scopeFew providers, few users, local scope of serviceDistributed providers, many users, non-local (in spatial, thematic, or other respects)
Embeddedness in user communitiesNot widely embedded, mainly pilot users and early adopters with a particular interestEmbedded, qualification to use the infrastructure is part of the socialisation into the community
Purpose and responsibilityAcademic organizations focussing on technical development and scientific discoveryNonacademic organizations focussing on service provision
Mechanisms of coordinationSolidarity, trustRoutines, procedures and orders
Formality of governanceInformal governanceFormal governance
Sustainability of fundingShort-term, grant-based, eventually renewable (upon application)Long-term, renewable grants, user fees or budgetary contributions

Sample and data collection

This paper is based on case studies for 16 e-infrastructures/projects which were assembled during the eResearch 2020 study for the European Commission, DG Information Society & Media (see Table 2). The cases cover different types of service: services based on Grid computing (OSG, EGEE), supercomputing services (TeraGrid, DEISA2), providers focussed on raw connectivity (GEANT), and providers of access to specialist data (C3-Grid, CLARIN). The study included systems specialised for their field of science, including the life sciences (MediGrid, Swiss BioGrid), the physical sciences (US-NVO), social sciences and humanities (CLARIN), as well as services for communities outside traditional academic disciplines (Driver for libraries or CineGrid for cinematic production). The cases were sampled to cover the discipline categories of the ESFRI roadmap (ESFRI, 2008). There is also considerable variation in the geographical range of the cases, which corresponds to today's e-infrastructure landscape: While there is a European bias in the cases, some span multiple continents; some cater to regional populations of scientists; while yet others concentrate their activities on a specific country. It is worth drawing attention to the fact that developing countries have not yet received much attention in the literature on e-Research/e-infrastructures (but see Ynalvez, Duque & Shrum, 2010; Luo & Olson, 2008), though several of the projects described here have links to, even if they are not based in, the developing world. Asian cases could not be included (but see for more recent descriptions of e-infrastructure developments in Asia Park, 2010 and Soon & Park, 2009).

Table 2. Sample of cases
Case acronymShort DescriptionDisciplines targeted with the service
C3-GridImproves access to distributed datasets in collaborative environments for Earth Science in GermanyClimatology, geophysics, biogeography, hydrology, oceanography, other earth system sciences
CineGridResearch, development, and demonstration of networked collaborative tools to enable the production, use and exchange of very-high-quality digital media over photonic networksComputer networking, scientific visualization, media science
CLARINImproves access to language resources across Europe and develops tools and technologies to allow computer-aided language processing(Computational) linguistics, languages, literature
D4ScienceEstablishes networked, Grid-based, and data-centric Virtual Research Environments (VREs) for Environmental Monitoring (EM) and Fisheries and Aquaculture Management (FARM) communitiesEnvironmental monitoring, fisheries and aquaculture resources management
DEISAIntegrates European national supercomputing platforms through Grid technologiesNuclear fusion, climate/earth system research, astrophysics/cosmology, computational neurosciences, plasma physics, computational bio sciences, materials sciences
DRIVERLinks distributed data from repositories across Europe and makes it accessible to a wider audience, develops tools and standards for digital repositoriesN/A (any)
EELA-2Extends the use of EGEE Grid middleware and infrastructure to Latin America and contributes to the establishment of a Latin American grid communityHigh-energy physics (HEP), biomedicine and bioinformatics, earth sciences, AI and optimization, chemistry, civil protection, engineering, environmental science
EGEE-IIIBuilds a Grid infrastructure to provide a production service to scientific researchers for sharing computing resources to analyse data from the Large Hadron Collider and other collaborative projectsHEP and several other disciplines across the domains (natural sciences, engineering, social sciences & humanities)
ETSFSupports experimental physicists in Europe with theoretical services and computer codesCondensed matter physics, chemistry, biology, material science, nanotechnology
GÉANTProvides and further develops a multi-gigabit pan-European backbone research network interconnecting Europe's national research and education networks (NRENs)N/A (any)
MediGridDevelops a Grid infrastructure for medical and biomedical research in Germany(Clinical) medicine, biomedicine, biomedical informatics
US-NVODevelops and provides technologies and standards for the integration of various astronomical digital data sets from diverse instruments in the USAstronomy
Open Grid Forum (OGF)Global forum for setting standards in Grid computingGrid computing
Open Science Grid (OSG)Builds a Grid infrastructure to provide a production service to scientific researchers for sharing computing resources to analyse data from the Large Hadron Collider and other collaborative projects in the USHEP (∼90%), others (10%) across the domains
Swiss BioGridProvides a Grid infrastructure for computing intensive applications and data sharing in the life sciences in SwitzerlandBiological sciences, pharmaceutical research
TeraGridIntegrates U.S. supercomputing platforms through Grid technologiesMolecular biosciences, physics, chemistry, astronomical sciences, materials research, earth sciences, advanced scientific computing, chemical thermal systems, atmospheric sciences (+ 19 other fields)

The cases also differ in regard to their “maturity,” which reflects the current situation where infrastructures can be found at many stages of development. Some cases have already developed most of their services and have likely reached their peak (OGF), while other projects have offered tools and engaged users but expect to considerably expand their service repertoire and extend their user community (US-NVO, DEISA2, OSG, TeraGrid), and others are still at a much more formative stage (CLARIN).

Data on these 16 cases was collected through a range of different methods. The approach, though strict in respect of the results that had to be produced, was open in regard to the methods employed in the field phase as most of the data were expected to be insensitive to method. Common templates were used to compile detailed quantitative and qualitative information on the infrastructure, and organizational documents and published materials were a chief source to examine operational logic and structure in each case. The documents included designated web sites for the selected infrastructures, publications, and presentations.

Using semistructured interviews4 based on a dedicated interview guide enabled considerable insight to be gained from a diverse set of providers, while maintaining analytic consistency across the cases. The sample was purposive, collected via snowball methods: The documentary information gathered was used to identify interviewees, and also structured the interview content by guiding the researchers towards potentially important aspects of the case at hand. Interviewees were selected from the senior management and/or technical manager levels of e-infrastructures. Up to 7 persons per case were interviewed depending on the functional diversity within the organization and the amount of published information resulting in a total of more than 49 interviews and 2,415 minutes of interview time (see annex table 1 on details). Interviews were partially transcribed.

Table Annex Table 1. Case interviews
Case acronymNo. of interviewsTotal duration in min.
  1. a

    In addition to the listed interviews several other interviews were done previously within other projects and were reused for this analysis.

C3-Grid390
CineGrid6380
CLARIN2150
D4Science3170
DEISA3120
DRIVER160
EELA-27340
EGEE-III3110
ETSF170
GÉANT270
MediGrid190
US-NVO3170
Open Grid Forum (OGF)2105
Open Science Grid (OSG)3150
Swiss BioGrid4200
TeraGrid3 (+2)*140
Total492415

Interviews and documents were used to produce extended case descriptions which then were fed back to the informants with a request for validation and comments. The validated descriptions were used to evaluate and classify the cases on the core variables, frequently consulting the original interview transcripts and documents to verify classification and forfurther detail.

For all cases, a set of variables is shown that attempts to operationalise the concepts listed above: size and scope, embeddedness in the user community, purpose and responsibility, coordination mechanisms, degree of formality, and time horizon.

  • Size and scope. Size was measured by the number of organizations participating. Assessing participation is not a trivial issue, as it takes different forms, from coordinating the efforts of the entire consortium or providing integral resources and services over the full life span, to accessing and using it once in order to obtain computational or other support. We opted to use the number of participants as stated by the e-infrastructures themselves and also their estimation of annual funding. Large cases have 20 or more participants, medium-sized between 11 and 20, and small have 10 or fewer members. In regard to funding, we distinguish megaprojects (more than 20 m of national currency p.a.), large projects (more than 5 m), regular projects (1–3 m) and small projects (probably below 1 m but in most cases no data). Scope was measured as geographic scope (national – international), disciplinary scope (less or more than five disciplines involved)5 and sectoral scope (academic only versus Public-Private Partnership PPP, i.e. involvement of commercial partners in the projects).
  • Embeddedness in the user community. This concept was measured by looking at the drivers of the development and the maturity of the user community. We defined as drivers the main organizations carrying and promoting the e-infrastructure. These could either be from a community of domain scientists (e.g. astronomers, high-energy physicists, biomedical researchers etc.) wishing to enhance their research capacities through an infrastructure; or they were from a community of computer scientists who developed applications for research and offered these and supporting services to scientific communities. In regard to the size of the user communities, we collected information on different aspects of their extension (geographically, organizationally, across disciplines) and synthesized this.
  • Purpose and responsibility: Purpose was evaluated by categorizing the main goals as specified by the interlocutors from the cases. Four types of goals were distinguished in the process: a) Technical development covers the new research, development and testing of tools, applications and systems; b) sociocultural progress refers for instance to education and training activities, community-building and advances regarding the institutionalisation and political representation of e-infrastructure ventures; c) scientific discovery summarises the desire to contribute to scientific progress in specific domains; d) service provision describes the aim of providing round-the-clock ICT-based support services to scientific undertakings. Responsibility was evaluated by assessing and categorizing the coordinating organizations of each case, distinguishing between academic (universities, nonuniversity and governmental R&D facilities) and nonacademic (corporations, nonprofit organizations) organizations.
  • Coordination mechanism. We evaluated the approaches to achieving coordination and assessed the importance of solidarity (community), prices and competition (market), routines, procedures and orders (hierarchy), trust and reciprocity (network).
  • Formality of governance was assessed by examining the incentives motivating compliance, the relationship between the actors, and sources of legitimacy of action in a case.
  • Sustainability of funding. The main indicators for evaluating this are the time horizon of the funding scheme, as well as the duration and stability of a support system.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies
Size and scope

An overview of the size and scope of the cases is given in Table 3. Half of them are large, a third is medium-sized, and only few are small. Commercial partners are involved in only five out of 16. Regarding geographic scope, our cases include six national and nine international e-infrastructures/projects. Roughly the same number of cases were multidisciplinary as against cases with fewer (up to five) disciplines involved. Looking at size and scope only, we see that only a few cases could clearly be categorized as either e-infrastructures (EGEE, Géant) or local support systems (MediGrid, US-NVO, Swiss BioGrid), whereas the majority lie in between.

Table 3. Size and scope of the cases
CaseNo. of participantsFunding p.a. (nat. currency)aGeographic scopeDisciplinary scopePPP (sectoral scope)
  1. a

    ) Funding as shown in project documents and public sources. Cross-financing and contributions in kind and labour from participating organizations or third parties could not be assessed consistently and are not included.

C3-Grid11–201–3 mNationalMultidisciplinaryYes
CineGrid>20<1 mInternationalFive or fewer disciplinesYes
CLARIN>201–3 mInternationalFive or fewer disciplinesNo
D4Science11–201–3 mInternationalFive or fewer disciplinesNo
DEISA11–205–10 mInternationalMultidisciplinaryNo
DRIVER11–201–3 mInternationalMultidisciplinaryNo
EELA-211–201–3 mInternationalMultidisciplinaryNo
EGEE-III>20>20 mInternationalMultidisciplinaryYes
ETSF11–201–3 mInternationalFive or fewer disciplinesNo
GEANT>20>20 mInternationalMultidisciplinaryNo
MediGrid<111–3 mNationalFive or fewer disciplinesNo
US-NVO11–201–3 mNationalFive or fewer disciplinesNo
OGF>20<1 mInternationalFive or fewer disciplinesYes
OSG>205–10 mNationalMultidisciplinaryNo
Swiss Biogrid<11<1 mNationalFive or fewer disciplinesYes
TeraGrid11–20>20 mNationalMultidisciplinaryNo
Embeddedness in the user community

In 8 out of the 16 cases, the main impulses for set up and development were given by members of the user communities, and in the other 8 cases the drivers were from the developer or provider side (see Table 4). The relationship between drivers and size of the user communities is not clear and we see cases with many users both in community- and developer-driven projects.

Table 4. Drivers and users of the casesa
CaseDriverType of usersSizea
  1. a

    ) Large: established large user population; medium: established small user population; small: no established user population.

C3-GridCommunityPilot users, mainly from the project membersSmall
CineGridCommunityCommunity members, technology developers and innovatorsSmall
CLARINCommunityPilot users from the project membersSmall
D4ScienceDevelopersPilot users from the project members, external users from two communities (EM, FARM)Small
DEISADevelopersScientists and other users of supercomputing servicesLarge
DRIVERDevelopersDifferent types: Repository managers provide content, organizations build their repository systems with the software and end users use the portalsLarge
EELA-2DevelopersPilot users from the project members, external users of Grid computing servicesMedium
EGEE-IIICommunityScientists and other users of Grid computing servicesLarge
ETSFDevelopersScientists and other usersMedium
GEANTDevelopersAnybody in Europe transmitting data through an NREN internationallyLarge
MediGridDevelopersUsers from project membersSmall
US-NVOCommunityCommunity members, research astronomersMedium
OGFCommunityCommunity members, organizations, and individuals involved in Grid computingLarge
OSGCommunityScientists from many different organizations and fieldsLarge
Swiss BiogridCommunityScientists from a subfield, pharmaceutical researchersSmall
TeraGridDevelopersGraduate students (36%), faculty (22.3%) and postdoctorates (12.7%) from the user fieldsLarge

Many cases involve a complex web of academic disciplines, developer and IT communities, and professional or nonacademic stakeholders (see Figure 1). In order to ensure that activities are properly coordinated, it is necessary to have formal mechanisms of governance such as clearly defined rules and legally enforceable contracts, material incentives, and hierarchical subordination. In these settings, participants can not be expected to share mutual understanding, trust and identity resulting from membership of the same scientific community. Two examples can illustrate this point:

  • The Open Grid Forum (OGF) is a global forum for setting standards in Grid computing. Though its members work at various academic, commercial and government/nonprofit organizations with distinct missions, they all share a common interest in promoting applied distributed computing and personally learn and benefit from being confronted with and contributing to cutting edge developments in this field. In this sense OGF fits with Benkler's ideas of “commons-based peer production” (2006, p. 60) discussed earlier. Information exchange, coordination and decision-making are largely informal and take place at biannual conferences, working group meetings, telephone conferences and online fora supported by a small staff.
  • TeraGrid is a joint initiative of several U.S. centres for supercomputing integrating their platforms through Grid technologies. Its participants, all associated with supercomputing, report a number of organizational idiosyncrasies as obstacles to effective collaboration. Working with them requires large communication effort and close management by a small group, the Grid Infrastructure Group (GIG), with the backing of the main stakeholders or resource providers. To ensure smooth collaboration and communication with the diverse user communities, TeraGrid has had to devise and implement novel mechanisms: Science Gateways are tailored portals hiding the complexity of the supercomputing environment from the users; Campus Champions advocate TeraGrid to local constituents, recruit TeraGrid users and serve as technical representatives, and are a point of first contact to assist users with issues relating to the project.
image

Figure 1. Stakeholders in e-infrastructure projects

Download figure to PowerPoint

Where governance is applied to a group within an organization of developers, a key element of governance structure is how to address the main structural problem of how to maintain a focus on user requirements. Indeed, we found some examples in which a group of users is set up to ensure that their interests as stakeholders are properly taken into account by developers. It could be imagined that effective interactions could be organized on behalf of users – the domain scientists. In the projects we examined, however, we did not find this. In fact, and going against a key success factor that has been identified in the literature on the topic discussed earlier, governance was more often organized to fit around various developer stakeholders. Steering committees were recruited from among the developer communities, and representation from the various member organizations or states was involved - but not from user communities.

Goals and responsibility

Looking at the main goals and purposes of the cases and the allocation of responsibilities for coordination and management, we see that most cases are strongly anchored in the academic domain (see Table 5). Only few cases pursue the core purpose that we would expect for an e-infrastructure; namely to provide a permanent and robust service – e.g. computing cycles, standards – to scientists engaged in the advance of knowledge. These cases are DEISA, EGEE/EGI, GEANT, OGF, OSG, and TeraGrid. All cases are at least to some extent involved in research and technical development of ICT-based support systems to science. Several of the cases also engage in complementing their technical goals with sociocultural aims, in particular informing and preparing scientific communities, science funders, and policy makers for e-infrastructure provision and use. Scientific discovery in one or several domains was explicitly mentioned by only few cases, but we may assume that this is a driving force of more cases.

Table 5. Goals and types of coordinating organizations of the cases
CaseGoalsaCoordinator(s)
TDSCSDSPAcademicNon-academicCoordinating organization(s)
  1. a

    ) TD Technical development, SC sociocultural progress, SD scientific discovery, SP service provision, see page 17 on examples for the goals.

C3-GridX X X Nonuniversity research institute
CineGridXX   XConsultancy firm
CLARINXX  X University institute
D4ScienceXX  X Umbrella association of research institutes in informatics and mathematics, nonuniversity research institute
DEISAXXXXX Computing centre of a nongovernmental and non-profit association of research institutes
DRIVERXX  X University department, university library
EELA-2XX XX Governmental R&D institute, nonuniversity research institute
EGEE-III/EGIXX XX (C)X (F)European intergovernmental scientific research organization (C), foundation (F)
ETSFXXX X University department
GEANTXX X XLimited liability company and a “Not for Profit” organization
MediGridX   X Umbrella organization for medical research networks, university institute
US-NVOX X X University departments
OGFXX X XNonprofit corporation
OSGXXXXX Governmental R&D institute
Swiss BiogridXXX X National academic computing centre
TeraGridXXXXX Governmental R&D institute and university institute
Coordination mechanisms

None of our cases relies on the price mechanism and a market model to achieve coordination among its participants. However, some of the projects use peer reviewed ‘calls’ to identify research projects among their potential users and allocate their capacities on projects via competitions for excellence (e.g. DEISA, ETSF).

The majority of cases are currently governed by mechanisms typical of the network coordination mechanism (see Table 6): Participants align their interests via an exchange of information and negotiations in reciprocal constellations; they pool and share competences and resources such as computing power, data, software or hardware to realize pilot projects. These then achieve a higher level of service – at least for some period – or even provide a production infrastructure to their members and eventually external customers and users.

Table 6. Coordination mechanisms, formality and time horizon of the casesa
CaseCoordination mechanismsFormalitybStart yearTime horizon of funding
  1. a

    In some cases we found a change of governance in progress at the time of the case study. These cases appear with a (C) to denote the current governance scheme and (F) to denote the future scheme in preparation.

  2. b

    The formal and informal distinction is in most cases a question of a scale, rather than two absolutely distinct poles.

  3. c

    Not answerable with the collected data.

  4. d

    Start of predecessors under different names/project start.

C3-GridNetworkFormal2005Short-term
CineGridNetworkInformal2002Long-term
CLARINNetworkInformal2008Short-term
D4ScienceNetworkFormal2004/2008dShort-term
DEISANetworkFormal2002Short-term
DRIVERNetwork (C), Organization (F)n/ac2006Short-term
EELA-2Network (C), Organization (F)Formal2004Short-term (C), long-term (F)
EGEE-III (C), EGI (F)Network (C), Organization (F)Formal2001/2004dLong-term
ETSFNetworkInformal2004Short-term
GEANTNetworkFormal2000Long-term
MediGridNetworkInformal2005Short-term
US-NVONetworkInformal1999Long-term
OGFNetworkInformal1999Long-term
OSGOrganizationFormal1999/2006dLong-term
Swiss BiogridNetworkInformal2004Short-term
TeraGridOrganizationFormal2001Long-term

Two out of the 16 cases are coordinated by organizational routines and procedures. TeraGrid and OSG are the major providers of distributed computing power in the US. OSG is governed by an Executive Board and council of central stakeholders which is responsible for resource provision and planning. TeraGrid is run by the GIG which coordinates and oversees the work of the resource providers. These have independent control over their local resources, but they have to meet intensive communication and reporting requirements to secure the coordination of service provision and other activities. Three further cases DRIVER, EELA-2, and EGEE-III/EGI were in the process of increasing the role of organization-like coordination mechanisms. EGEE was probably the most advanced in this regard and its successor organization, the European Grid Infrastructure (EGI) has been set up as a new organization (foundation) under Dutch law in February 2010.

GÉANT is unique in having secured very high levels of funding – it was the largest project in our set with a total budget of more than 40 m € per year – in consecutive funding rounds, probably due to its position as indispensable and unrivalled provider of cross-national connectivity. GÉANT has a particularly complex governance structure. Though it is not alone in being governed in a consortium of independent organizations, this ‘network of networks’ links closely into the activities of the national research and educations networks (NRENs) and governance involves NREN Policy and Executive Committees. Project governance structures continue to be used to fund GÉANT partly for expediency – other governance structures in which the necessary funding can be provided at European or global level had not been identified or not agreed – and partly because the services demanded by users of NRENs joined by GÉANT push at the limits of provider ingenuity and require continuous addition of expertise, best planned in individual projects.

Formality

As can be seen, approximately half of the cases studied work primarily by applying informal governance mechanisms, and in the other half formal mechanisms seem stronger. It is important to point out that all the cases had at least some formal structures in place. For instance, in EELA-2 the consortium is bound by a consortium agreement and contract with the funders. Though in theory it would be possible to resort to contractual sanctions for ensuring compliance and resolving conflicts, in practice, science-based incentives such as maintaining reputation and esteem from the peer community and the funders appear more important. Relationships are either horizontal or vertical between the EELA-2 participants depending on their status and functions, at national level and in the project overall. In CineGrid, to give a second example, trust and mutual understanding facilitate collaboration and the sharing and joint use of resources. Use is dependent on membership, which requires an application, the setup of a minimum infrastructure and the payment of membership fees. Nevertheless, sharing of resources and participation in CineGrid projects are voluntary. Members are engaged in horizontal relationships, and leadership varies across demonstrator projects.

If we compare Tables 3 and 6, we see that there is a clear relationship apparent between the involvement and role of different scientific communities and the formality of governance: Multidisciplinary and developer-driven projects are more formal than those within one or few related communities and driven by these communities themselves. Informal governance appears to function only if the participants have a common basis in the same academic discipline (such as SwissBioGrid, for example). This mirrors the problems which collaboration with computer scientists faces in multiple settings: a negative attitude towards technology and computer-enhanced research among users (here: domain scientists), little understanding of user requirements (here: domain-specific practices), general problems of interdisciplinary cooperation – field jargon and communication across disciplines – as well as divergent motivations (cutting-edge research versus service provision). These were mentioned as some of the strongest challenges in e-infrastructure development in the case interviews.

Sustainability of funding

Most of the governance structures for the infrastructures are of short-term validity – they are research projects with a specified duration and follow-up funding still pending. U.S.-based cases (CineGrid, US-NVO, OGF, OSG, and TeraGrid) are on average 3 years older than European cases. It is interesting to see that small activities such as CineGrid, the U.S. National Virtual Observatory or the Open Grid Forum have found it possible to establish quite solid long-term funding, despite their relatively informal governance structures and management processes. Most European cases still need to put in place long-term funding arrangements, and only the ‘flagships’ Géant and EGEE/EGI have succeeded in this respect.

Clustering of cases

In a last analytical step we grouped all cases by means of a hierarchical cluster analysis. We included in this analysis all variables shown in annex table 2, standardised them to a value range between 0 and 1, calculated the squared Euclidean distances and grouped the cases by means of the Ward algorithm. The three cluster solution provided the best fit to the data:

  • Cluster 1 Stable metaorganizations (EGEE-III/EGI, GEANT, OSG, TeraGrid): In the first cluster we find four cases that have reached the level of an e-infrastructure in several respects: Serving multiple disciplines and countries – except for the two US cases OSG and TeraGrid – they are large in respect of number of service providers and size of user community. In most cases, nonacademic administrators and technicians coordinate the participating organizations by means of hierarchical procedures, formal rules and incentives. They all have a significant funding base and funding security over the long term.
  • Cluster 2 Established communities (CineGrid, OGF): The second, less populated cluster comprises cases which are strongly anchored in a few communities, notably in scientific communities as well as communities of practice. Nonacademics have a strong position in the coordination of the cases and participants are motivated to comply by means of trust and reciprocal activities. These communities have a relatively low funding requirements and meet these in a stable manner through membership fees and contributions.
  • Cluster 3 ICT-support systems in a state of flux (C3-Grid, CLARIN, D4Science, DEISA, DRIVER, EELA-2, ETSF, MediGrid, Swiss Biogrid, US-NVO): This cluster is the largest of the three but also the most heterogeneous. The cases appear either to lean towards the metaorganizational structures of the first cluster, exhibiting increasingly formalised governance features, including hierarchy components (DEISA, DRIVER, EELA-2), or towards the established communities of the second (US-NVO). In all these cases sustainability either still needs to be proven or has been tested and found wanting - some in the cluster (at the time of writing two: MediGrid, Swiss Biogrid) have already ceased activity, funding having been withdrawn. The main differences between this cluster and the established communities of the second cluster are the smaller and shorter-term funding base, fewer users, and the lesser focus on service provision. DEISA is in all regards the closest to the first cluster and certainly at least a borderline case of stable metaorganization.
Table Annex Table 2. Variables in the cluster analysis
Var.LabelMeanMinMax
nopartNumber of participants (grouped)2.2513
fundFunding per year (grouped)2.3114
geoscopeGeographical scope0.6301
disscopeDisciplinary scope0.5001
secscopeSectoral scope0.3101
driverDriving forces0.5001
sizeuseSize of the user communities (grouped)2.0613
TDGoal of technical development1.0001
SCGoal of sociocultural progress0.8101
SDGoal of scientific discovery0.4401
SPGoal of service provision0.3801
acadAcademic coordinators0.8101
nonacadNonacademic coordinators0.2501
coordCoordination mechanisms0.1901
formFormality of governance2.3813
ageAge of the case6.50110
horizonTime horizon of the funding0.4401

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies

We found considerable variety in the governance arrangements among the cases of current e-infrastructure provision we studied. This may be due to different requirements for governance mechanisms, relating to the specific services, the targeted user groups and the geographic spread of provision, but seems to derive in large measure from the varying progress made towards large-scale, production-quality, sustainable e-infrastructure provision. When viewing clusters of cases in a space formed by six constructs, we can see a group of ventures which have established themselves with a sustainable governance structure, or are very close to reaching this stage (CineGrid, EGEE/EGI, Géant, OGF, OSG, TeraGrid). Such cases are mainly to be found among those ventures in our sample which have already been active the longest. Quite distant from these are ventures which were found to be still in project-style governance (MediGrid, Swiss BioGrid). Such structures appear to be increasingly difficult to sustain over time - both investigated examples have in the meantime terminated their activities. In between the project-style and mature e-infrastructure forms of governance there are many cases at varying stages of development.

Among the six cases having achieved a sustainable governance structure for their e-infrastructure services, there is a predominance of large-scale structures providing connectivity or distributed computing services (EGEE/EGI, Géant, OSG, TeraGrid). These differ in their geographic reach and the numbers of disciplines they serve. Metaorganization structures predominate, with organizational membership and formal governance resting on material incentives, vertical relationships (between the coordinator and the distributed members), and written contracts or other binding agreements. The second type of governance of mature e-infrastructure is the established community, notably with a mix of academic and private sector participants (CineGrid and OGF). These are coordinated based on trust relationships and ongoing reciprocal action, and secure their financial stability by small-scale funding through membership fees and other contributions.

Our analysis revealed no clear third alternative to the stable metaorganization and established community as a governance structure for mature e-infrastructure provision. If there is indeed no third way for sustainable provision, it seems that the other ventures studied (C3-Grid, CLARIN, D4Science, DEISA, DRIVER, EELA-2, ETSF, US-NVO) will need to choose one of these two roads to avoid the destiny of stalling in project-style structures, the fate met by MediGrid and Swiss BioGrid. It is too early to tell if the emerging governance of e-infrastructure exposed here will be affected by the current broader transition in governance in science, characterised by a weakening of decision-making processes internal to scientific communities and a shift of power towards external governance structures. Currently, e-infrastructure provision continues to rely quite strongly on network mechanisms for coordination, and less on classical hierarchical organization or market, price-based mechanisms for balancing supply and demand. Perhaps surprisingly, the role of engagement with user communities in establishing infrastructures more securely appears to be quite weak. As today's e-infrastructure ventures try to establish themselves on firmer ground, they have to contend with an difficult economic climate, but can profit from increasing recognition of the importance of science and resulting innovation in the economy and wider society. Nevertheless it can be expected that existential tensions will remain, even among our clusters of stable metaorganizations and established communities. External governance pressures can be expected to increase, requiring that e-infrastructures continually and transparently monitor their effectiveness, and the usefulness of their services, and that such monitoring is firmly anchored in their governance arrangements. Recent work on providing a new legal instrument for European research infrastructures, the European Research Infrastructure Consortium (ERIC) is a sign of the recognised importance of governance of these new services.

The establishment of e-infrastructures, with sustainable governance, is part of the wide-ranging change of research processes and structures taking place as the potential of information and communication technology for knowledge advancement is increasingly realised. The infrastructures examined in this paper are seen as being at a more or less advanced stage of transition from project-style structures, through various forms of collaborative socio-technical systems, towards sustainable existence - in the shape of meta-organizations or established communities. In most ventures, governance mechanisms are still in flux: In many, the typical early bias towards developer rather than user communities has to be overcome and user needs take centre stage, if the problem of longer term, sustainable resourcing is to be solved. The ventures studied may leave a rather mixed picture of different forms, levels of maturity and of governance structures, but this appears to be a true reflection of the stage science has reached today in nurturing the emergence of sustainable governance of e-infrastructure.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies

The paper mainly draws on data collected within the eResearch 2020 study (http://www.eresearch2020.eu). The U.S.-American cases within this study were prepared by Zack Kertcher and Erica Coslor from the National Opinion Research Center at the University of Chicago, the other cases by the authors and Oliver Bendel, Gordon Wiegand (both FHNW), and Tobias Hüsing (empirica). The study itself was funded by the European Commission, DG Information Society & Media. Part of the work (at the Oxford Internet Institute) was also supported by ESRC grant RES-149-25-1022 for the Oxford e-Social Science (OeSS) project: Ethical, Legal and Institutional Dynamics of Grid-enabled e-Sciences. We would also like to thank the two reviewers for this journal for their very helpful comments.

Notes
  1. 1

    We are here applying Stoker's appropriately broad understanding of governance as “ultimately concerned with creating the conditions for ordered rule and collective action” (1998, p. 17).

  2. 2

    Their definition of a collaboratory is as follows: “A collaboratory is an organizational entity that spans distance, supports rich and recurring human interaction oriented to a common research area, and fosters contact between researchers who are both known and unknown to each other, and provides access to data sources, artifacts, and tools required to accomplish research tasks.” (Bos et al., 2007, p. 656).

  3. 3

    This second step relates to ‘steering and aligning structures’ described by Fry and Schroeder (2009).

  4. 4

    Since providers are often widely spread geographically, interviews were primarily carried out via telephone. Although this imposes some limitations on the ability to interpret the responses of interviewees without non-verbal cues, the interviewers saw little evidence that responses were biased as a result of the interview mode.

  5. 5

    Five disciplines is an arbitrary threshold, but we find that it is well suited to divide disciplinary from multidsciplinary cases. Usually disciplinary cases serve fewer and multidisciplinary cases considerably more disciplines and in the disciplinary cases the focus is narrower on one academic domain (e.g. life sciences, earth sciences, physical sciences).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies
  • Avery, P. (2007). Open science grid: Building and sustaining general cyberinfrastructure Using a collaborative approach [Electronic Version]. First Monday, 12. Retrieved 9. November 2006 from http://www.firstmonday.org/issues/issue12_6/avery/.
  • Bany Mohammed, A., & Altmann, J. (2010). A funding and governing model for achieving sustainable growth of computing e-infrastructures. Annals of Telecommunications, 65(11), 739756.
  • Barjak, F., Lane, J., Kertcher, Z., Poschen, M., Procter, R., & Robinson, S. (2009). Case studies of e-infrastructure adoption. Social Science Computer Review, 27(4), 583600.
  • Barjak, F., Lane, J., Poschen, M., Procter, R., Robinson, S., & Wiegand, G. (2010). e-infrastructure adoption in the social sciences and humanities - cross-national evidence from the AVROSS survey. Information, Communication & Society, 13(5), 635651.
  • Benkler, Y. (2006). The wealth of networks: How social production transforms markets and freedom. New Haven & London: Yale University Press.
  • Berman, F. & Brady, H. (2005). Final Report: NSF SBE-CISE Workshop on Cyberinfrastructure and the Social Sciences. Retrieved 3 October 2006, from: http://vis.sdsc.edu/sbe/reports/SBE-CISE-FINAL.pdf.
  • Borgman, C. (2007). Scholarship in the digital age: Information, infrastructure, and the Internet. Cambridge, MA: MIT Press.
  • Bos, N., Zimmerman, A., Olson, J., Yew, J., Yerkie, J., Dahl, E., et al. (2007). From shared databases to communities of practice: A taxonomy of collaboratories. Journal of Computer-Mediated Communication, 12(2), 652672.
  • Cummings, J. N., Finholt, T. A., Foster, I., Kesselman, C., & Lawrence, K. (2008). Beyond being there: A blueprint for advancing the design, development, and evaluation of virtual organizations. Retrieved 28.04.2009, from http://www.ci.uchicago.edu/events/VirtOrg2008/VO_report.pdf.
  • David, P. A. (2006). Toward a cyberinfrastructure for enhanced scientific collaboration: Providing its “soft” foundations may be the hardest part. In B. Kahin & D. Foray (Eds.), Advancing knowledge and the knowledge economy (pp. 431453). Cambridge: MIT Press.
  • David, P. A., & Spence, M. (2003). Towards institutional infrastructures for e-science. Final report of the OII project on “The Institutional Infrastructure of e-Science: The Scope of the Issues” to the Joint Information Systems Committee (JISC). Retrieved 29.09.2003, from http://www.oii.ox.ac.uk/E-ScienceRpt_Final-150903.pdf.
  • European Strategy Forum on Research Infrastructures ESFRI (2008). European roadmap for research infrastructures: Update 2008. Luxembourg: Office for Official Publications of the European Communities.
  • Foster, I., & Kesselman, C. (2006). Scaling system-level science: Scientific exploration and IT implications. Computer, 39(11), 3139.
  • Fry, J., & Schroeder, R. (2010). Disciplinary differences in e-research. In W. Dutton & P. Jeffreys (Eds.), World wide research (pp. 257271). Cambridge: MIT Press.
  • Fry, J., & Schroeder, R. (2009). Towards a sociology of e-research: Shaping practice and advancing knowledge. In N. Jankowski (Ed.), e-Research: Transformation in Scholarly Practice (pp. 3553). New York: Routledge.
  • Gläser, J. (2006). Wissenschaftliche Produktionsgemeinschaften: Die soziale Ordnung der Forschung. Frankfurt am Main: Campus.
  • Gläser, J., & Lange, S. (2007). Wissenschaft. In A. Benz, S. Lütz, U. Schimank & g. Simonis (Eds.), Handbuch Governance: Theoretische Grundlagen und empirische Anwendungsfelder (pp. 437451). Wiesbaden: VS Verlag für Sozialwissenschaften.
  • Heimeriks, G., & Vasileiadou, E. (2008). Changes or transition? Analysing the use of ICTs in the sciences. Social Science Information, 47(5), 529.
  • Iacono, C. S., & Freeman, P. A. (2006). Cyberinfrastructure-in-the-making: Can we get there from here? In B. Kahin & D. Foray (Eds.), Advancing knowledge and the knowledge economy (pp. 455477). Cambridge: MIT Press.
  • Jackson, S. J., Edwards, P. N., Bowker, G. C., & Knobel, C. P. (2007). Understanding infrastructure: History, heuristics and cyberinfrastructure policy [Electronic Version]. First Monday, 12. Retrieved 03.09.2010 from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/1904/1786.
  • Jankowski, N. (2009). The contours and challenges of e-research. In N. Jankowski (Ed.), e-Research: Transformation in scholarly practice (pp. 331). New York: Routledge.
  • Jones, B. F., Wuchty, S., & Uzzi, B. (2008). Multi-university research teams: Shifting impact, geography, and stratification in science. Science, 322(5905), 12591262.
  • Kling, R., McKim, G., & King, A. (2003). A bit more to it: Scholarly communication forums as socio-technical interaction networks. Journal of the American Society for Information Science and Technology, 54(1), 4767.
  • Luo, A. & Olson, J. (2008). How collaboratories affect scientists from developing countries. In G. Olson, A. Zimmerman & N. Bos (Eds.), Scientific collaboration on the Internet (pp. 365376). Cambridge, MA: MIT Press.
  • Mackie, C. J. (2007). Cyberinfrastructure, institutions, and sustainability [Electronic Version]. First Monday, 12. Retrieved 9. November 2006 from http://www.firstmonday.org/issues/issue12_6/mackie/.
  • Meyer, E.T., Madsen, C., & Fry, J. (2010). Digital resources and the future of libraries. In W.H. Dutton & P. Jeffreys (Eds.). World wide research: Reshaping the sciences and humanities (pp. 8397). Cambridge, MA: MIT Press.
  • Meyer, E.T., & Schroeder, R. (2009). Untangling the web of e-research: Towards a sociology of online knowledge. Journal of Informetrics 3(3), 246260.
  • Park, H. W. (2010). Mapping the e-science landscape in South Korea using the webometrics method. Journal of Computer-Mediated Communication, 15(2), 211229.
  • Powell, W. (1990). Neither market nor hierarchy: Network forms of organization. In L. Cummings & B. Staw (Eds.), Research in organizational behavior (1st ed., vol. 12, pp. 295336).
  • Research Councils UK (2010). Delivering the UK's e-infrastructure for Research and Innovation. Report commissioned by the Department for Business Innovation and Skills. Retrieved 30 June 2011 from www.rcuk.ac.uk/documents/research/esci/e-infrastructurereviewreport.pdf.
  • Rieger, O. (2009). Search engine use behaviour of students and faculty: User perceptions and implications for future research [Electronic Version]. First Monday, 14 (12). Retrieved 10 June 2011 from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/viewArticle/2716/2385.
  • Soon, C., & Park, H. W. (2009). The rise of e-science in Asia. Dreams and realities for social science research: Case studies of Singapore and South Korea. In N. Jankowski (Ed.), e-Research: Transformation in scholarly practice (pp. 109128). New York: Routledge.
  • Star, S. L., & Ruhleder, K. (1996). Steps toward an ecology of infrastructure: Design and access for large information spaces. Information Systems Research, 7(1), 111134.
  • Shrum, W., Genuth, J., & Chompalov, I. (2007). Structures of scientific collaboration. Cambridge, MA: MIT Press.
  • Stoker, G. (1998). Governance as theory: Five propositions. International Social Science Journal, 50(155), 1728.
  • Streeck, W., & Schmitter, P. C. (1985). Community, market, state - and associations? The prospective contribution of interest governance to social order. European Sociological Review, 1, 119138.
  • Thompson, G. F. (2003). Between hierarchies and markets: The logic and limits of network forms of organization. Oxford: Oxford University Press.
  • Voss, A., Mascord, M., Fraser, M., Jirotka, M., Procter, R., Halfpenny, P. et al. (2007a). e-Research Infrastructure Development and Community Engagement. Paper presented at the UK e-Science 2007 All Hands Meeting. Retrieved 31 October 2007 from http://www.allhands.org.uk/2007/proceedings/papers/866.pdf.
  • Voss, A., Procter, R., Hewitt, T., Asgari-Targhi, M., Daw, M., Baun, C., et al. (2007b). Sustainability of e-infrastructures (for the Social Sciences). Paper presented at the Third International Conference on e-Social Science. Retrieved 30 June 2011 from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.126.4642&rep=&rep1type&=pdf.
  • Whitley, R. (2010, October). Changing governance and authority relations in the public sciences. Paper presented at the “Tentative governance in emerging science and technology” conference, University of Twente, Enschede, The Netherlands.
  • Williamson, O. E. (1971, May). The vertical integration of production: Market failure considerations. The American Economic Review, 61, 112123.
  • Williamson, O. E. (1975). Markets and hierarchies: Analysis and antitrust implications. New York. Free Press.
  • Wittek, R. (2007). Governance from a sociological perspective. In D. Jansen (Ed.), New forms of governance in research organizations: Disciplinary approaches, interfaces and integration. (pp. 7397). Dordrecht, The Netherlands: Springer.
  • Wuchty, S., Jones, B. F., & Uzzi, B. (2007). The increasing dominance of teams in production of knowledge. Science, 316(5827), 10361039.
  • Ynalvez, M.A., Duque, R.B., & Shrum, W. 2010. Shaping research in developing areas. In W.H. Dutton and P. Jeffreys (Eds.), World wide research: Reshaping the sciences and humanities (pp. 325342). Cambridge, MA: MIT Press.

Biographies

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Acknowledgements
  7. References
  8. Biographies
  • Franz Barjak (franz.barjak@fhnw.ch) is Professor at the School of Business of the University of Applied Sciences Northwestern Switzerland. He earned a Dr. phil. in Geography from the Ruhr-University Bochum. His research interests include knowledge and technology transfer and university-industry cooperation, scientometrics, e-science and e-social science. Address: School of Business, University of Applied Sciences Northwestern Switzerland, Riggenbachstrasse 16, CH-4600 Olten, Switzerland.

  • Kathryn Eccles (kathryn.eccles@oii.ox.ac.uk) is a Research Fellow at the Oxford Internet Institute at the University of Oxford. Her research focuses on the impact of new technologies on scholarly behaviour and research, particularly in the Humanities. Prior to becoming interested in e-Research and the Digital Humanities, she gained a D.Phil in Modern British History from the University of Oxford. Address: Oxford Internet Institute, University of Oxford, 1 St Giles, Oxford, OX1 3JS, United Kingdom.

  • Eric T. Meyer (eric.meyer@oii.ox.ac.uk) is a Research Fellow at the Oxford Internet Institute at the University of Oxford. His main area of research has been understanding, from a social informatics perspective, how digital technologies are enabling innovation in research practices and the creation of new types of knowledge. Before joining the faculty at Oxford, he earned a Ph.D. from Indiana University in the United States. More information is available at: http://www.oii.ox.ac.uk/people/?id=120. Address: Oxford Internet Institute, University of Oxford, 1 St Giles, Oxford, OX1 3JS, United Kingdom.

  • Simon Robinson (simon.robinson@empirica.com) is director of empirica Communication and Technology Research in Bonn, Germany. He has principal responsibility for policy-oriented research on knowledge transfer, innovation and the organization of science at empirica, and in this capacity led the European eResearch 2020 study. Address: empirica Gesellschaft für Kommunikations- und Technologieforschung mbH, Oxfordstr. 2, D-53111 Bonn, Germany.

  • Ralph Schroeder (ralph.schroeder@oii.ox.ac.uk) Schroeder is Professor at the Oxford Internet Institute at the University of Oxford. His books include ‘Rethinking Science, Technology and Social Change’ (Stanford University Press 2007) and ‘Being there Together: Social Interaction in Virtual Environments’ (Oxford University Press 2010). Before coming to Oxford, he was Professor at Chalmers University in Gothenburg, Sweden. His current research is focused on the digital transformations of research. Address: Oxford Internet Institute, University of Oxford, 1 St Giles, Oxford, OX1 3JS, United Kingdom.