Content uploaded by Leonardo Candela
Author content
All content in this area was uploaded by Leonardo Candela
Content may be subject to copyright.
VIRTUAL RESEARCH ENVIRONMENTS: AN OVERVIEW AND A
RESEARCH AGENDA
Leonardo Candela*, Donatella Castelli, Pasquale Pagano
Istituto di Scienza e Tecnologie dell’Informazione (ISTI) “Alessandro Faedo”, Italian National Research
Council (CNR), Via G. Moruzzi 1, 56124 Pisa, Italy
Email: leonardo.candela@isti.cnr.it, {donatella.castelli, pasquale.pagano}@isti.cnr.it
ABSTRACT
Virtual Research Environments are innovative, web-based, community-oriented, comprehensive, flexible, and
secure working environments conceived to serve the needs of modern science. We overview the existing
initiatives developing these environments by highlighting the major distinguishing features. We envisage a future
where regardless of geographical location, scientists will be able to use their Web browsers to seamlessly access
data, software, and processing resources that are managed by diverse systems in separate administration
domains via Virtual Research Environments. We identify and discuss the major challenges that should be
resolved to fully achieve the proposed vision, i.e., large-scale integration and interoperability, sustainability,
and adoption.
Keywords: Virtual Research Environment, Scientific Gateway, Digital Library, Data Infrastructure
1 STATE OF THE ART
A recent study promoted by The Royal Society in cooperation with Elsevier reviewed the changing patterns of
science and scientific collaborations and confirmed that science is increasingly global, multipolar, and
networked (Llewellyn Smith, Borysiewicz, Casselton, Conway, Hassan, Leach, et al., 2011). This trend calls for
innovative, dynamic, and ubiquitous research supporting environments where scattered scientists can seamlessly
access data, software, and processing resources managed by diverse systems in separate administration domains
through their web browser.
Dependent on context, these environments are commonly referred to as either Virtual Research Environments
(Carusi & Reimer, 2010), Science Gateways (Wilkins-Diehr, 2007), Collaboratories (Wulf, 1993), Digital
Libraries (Candela, Castelli, & Pagano, 2011) or Inhabited Information Spaces (Snowdon, Churchill, & Frécon,
2004). These environments are among the goals that e-Infrastructures (e-Infrastructure Reflection Group, 2010)
and cyberinfrastructures (Cyberinfrastructure Council, 2007) are going to realise. A variety of systems and
services fall within the scope of these definitions, from ad-hoc portals with minimal access services to content
resources held in external repositories (lightweight integration to promote resource discovery) to
general-purpose management systems with advanced services defined over a wide range of resources (strong
integration to promote resource exploitation). In some cases, motivations and design (sharing, on-demand
resource provision, economies of scale) align with the principles of grid computing and its ecology of virtual
organizations (Foster & Kesselman, 1998) as well as with cloud computing (Foster, Zhao, Raicu, & Lu, 2008).
For the purposes of this paper, the term Virtual Research Environment (VRE) is used with a comprehensive
scope, i.e., it represents a concept overarching all the environments cited above and identifies a system with the
following distinguishing features: (i) it is a web-based working environment; (ii) it is tailored to serve the needs
of a community of practice (Lave & Wenger, 1991); (iii) it is expected to provide a community of practice with
the whole array of commodities needed to accomplish the community’s goal(s); (iv) it is open and flexible with
respect to the overall service offering and lifetime; and (v) it promotes fine-grained controlled sharing of both
intermediate and final research results by guaranteeing ownership, provenance and attribution.
The VREs’ characteristic of being a web based working environment is the most common one and usually that
which contributes to misuse of the term “VRE” itself. In many cases, ad-hoc portals implementing simple
catalogue facilities and completely missing the collaborative, dynamicity, and openness features discussed
above have been tagged with the “VRE” term (Allan, 2009). Allan explains how Web-based services should be
loosely combined into portals to provide a comprehensive infrastructure for the support of research across all
Data Science Journal, Volume 12, 10 August 2013
GRDI75
academic disciplines. He feels that “VRE” portals should not only provide an environment for housing, indexing,
and retrieving large data sets but also leverage Web 2.0 technologies (O'Reilly, 2005) and social networking
solutions (Wang, Carley, Zeng & Mao, 2007) to give researchers a comprehensive environment for collaboration
and resource discovery.
The VREs’ characteristic of being the framework expected to support communities of practice is what makes
VREs definition very heterogeneous and VREs implementation a challenging activity. “Community of practice”
is a term coined to capture an “activity system” that includes individuals who are united in action and in the
meaning that “action” has for them and for the larger collective. The communities of practice are “virtual”, i.e.,
they are not formal structures, such as departments or project teams. Instead, these communities exist in the
minds of their members and are glued together by the connections they have with each other as well as by their
specific shared problems or areas of interest. The generation of knowledge in communities of practice occurs
when people participate in problem solving and share the knowledge necessary to solve the problems (Wenger,
1998). Creating and supporting communities of practice as a strong alternative to building teams was an early
observation (Nirenberg, 1994). This is particularly true in science and scientific collaborations as confirmed by
the Royal Society study previously cited. It is evident that realising working environments for communities with
features for communities of practice is a challenging task: the service has to be guaranteed at a level of quality
of service although the requirements and needs are highly evolving and the membership is volatile.
The VREs’ characteristic of being the system that offers the whole array of needed commodities is another
aspect concurring to difficulties in defining VREs’ scope boundaries and enlarging realization challenges. The
larger the pool of expected commodities (both quantitatively and qualitatively) the bigger the effort needed to
implement the related VRE. It is quite common to describe a VRE’s commodities by decoupling the resources
managed through the VRE from the VRE’s services facilitating resources management. Resources range from
data sets, collections, storage facilities, and computing power to services realising specific utilities and research
objects. Research objects themselves evolve from traditional research outputs, such as papers and experimental
data, to living reports (Candela, Castelli, Pagano, & Simi, 2005; Candela, Akal, Avancini, Castelli, Fusco,
Guidetti, et al., 2007), executable research papers (Van Gorp & Mazanek, 2011; Nowakowski, Ciepiela,
Harężlak, Kocot, Kasztelnik, Bartyoski, et al., 2011), scientific workflows (De Roure, Goble, & Stevens, 2009),
and enhanced publications (Hoogerwerf, Lösch, Schirrwagen, Callaghan, Manghi, Iatropoulou, et al., 2013). In
addition to that, a VRE is required to offer a unified and sometimes virtualised view on a pool of resources that
might come from different “providers”.
The VREs’ characteristic of being open and flexible with respect to the overall service offering brings
development approaches into question. Traditional approaches, mainly based on from scratch development of
ad-hoc portals, are not sustainable in community-of-practice-oriented scenarios. There is a need for innovative
approaches aimed at promoting and maximising sharing and reuse of existing commodities to build and operate
a number of VREs. In the context of the DILIGENT, D4Science, D4Science-II triplet of EU projects an
approach that has been developed and deployed based on: (i) an infrastructure making available a rich pool of
resources including datasets, computing power, and hosting machines; (ii) a software framework offering
resource management facilities for a rich array of resources including software packages; (iii) a wizard-based
mechanism allowing users to characterize the VRE they are interested in; and (iv) automatic VRE deployment
facilities that acquire the constituents needed to satisfy the VRE specification by relying on the infrastructure
and software framework offering (Assante, Candela, Castelli, Frosini, Lelii, Manghi, et al., 2008). In the three
projects, the intended communities of practice ranged from humanities research to biodiversity. Moreover, this
initiative is among the first to rely on cloud technologies to implement VREs (Candela, Castelli, & Pagano,
2010).
Finally, the VREs’ characteristic of supporting fine-grained controlled sharing of both intermediate and final
research results while guaranteeing ownership, provenance, and attribution is somehow a consequence of the
scenarios VREs are going to serve. Many science users will not be willing to contribute unless mechanisms
guaranteeing their work are in place (De Roure, Goble, & Stevens, 2009). These mechanisms can be either
explicit, e.g., the visibility of a resource is defined by its creator/owner through a set of policies, or implicit, e.g.,
it is the framework implementing the VRE that injects provenance metadata in the research outputs.
Data Science Journal, Volume 12, 10 August 2013
GRDI76
2 TEN-YEAR VISION
In ten years, it is expected that the trend characterizing science and scientific collaborations discussed above
continues, thus becoming the “default” approach for scientific investigations as well as for any societal
collaboration-based activity. Virtual Research Environments will be integrated into standard practices and tools
used by communities of practice, thus becoming the “enabler” working environments for implementing
investigation and collaboration activities efficiently and effectively.
The creation and management of Virtual Research Environments will be a very straightforward process that
relies on specific services – VRE Management Services – built atop a “global virtual infrastructure” resulting
from the aggregation and interoperation of a number of existing infrastructures and systems. The VRE
Management Services will support the phases of VRE definition, deployment, and monitoring / maintenance.
The VRE definition phase will guide an authorised actor of an application domain in characterizing the expected
VRE service in very abstract terms, e.g., defining the policies and procedures governing the VRE community
building, defining the policies governing the VRE operation, identifying the datasets the VRE community is
willing to play with, describing the data types the VRE community is going to manage, and identifying the
facilities the VRE is requested to support. Which characterizations are allowed depends on the current offering
of the “global virtual infrastructure”, i.e., the “global virtual infrastructure” is actually playing the role of
“resources provider” and a VRE will be an application built by dynamically acquiring the needed constituents
from the overall offering. The quality of service of the resulting VRE is declared in its specification; thus it is
known a-priori and depends on the amount of resources spent to acquire the resources needed to realize the
VRE.
The VRE deployment phase will be almost automatic. The Management Services will crunch the specification
of the expected VRE service including (i) the directives on the quality of service and (ii) the available budget to
identify the “optimal” set of resources to be acquired from the “global virtual infrastructure”. The Management
Services will take care of creating the application context changing this set of resources from a complex whole
into an integrated system.
The VRE monitoring / maintenance phase will require little direct human control. The Management Services
will take care of checking the state of the set of resources allocated to implement the VRE service. When needed,
they will perform corrective actions aiming at guaranteeing that the VRE service specification is satisfied, e.g.,
by dynamically acquiring new resources from the “global virtual infrastructure”.
The resulting Virtual Research Environment will be very flexible and customizable. Every single user can
simply define its own workflows – workflows realizing a scientific investigation – by combining existing
facilities without taking care of implementation details and computational resources acquisition. The
computational resources as well as the workflow constituents will be dynamically acquired and combined by the
Management Service, in accordance with the VRE specification.
Thus Virtual Research Environments creation and management will become a societal and organisational
process rather than a technological one.
3 CURRENT CHALLENGES
There are three major issues to be resolved to realise the above vision as well as to implement sustainable
Virtual Research Environments: large scale integration and interoperability, sustainability, and adoption.
Because of their intrinsic nature, any Virtual Research Environment is built as a “collection” of existing systems
and resources; thus their developers have to deal with the entire stack of issues that go under the interoperability
umbrella. Interoperability is actually a multi-layered and context-specific concept, which encompasses different
levels along a multi-dimensional spectrum ranging from organisational to semantic and technological aspects.
From the VRE developers’ point of view it is fundamental to rely on a rich array of systems and resources –
both in terms of variety and size – that can be seamlessly accessed and combined in innovative ways to satisfy
the evolving needs of the community of practice. Part of the resources can be acquired and put in place from
scratch for specific purposes while other resources have necessarily to be acquired from existing systems either
because they are produced by those systems or for opportunistic reasons, e.g., economic ones. However, the
Data Science Journal, Volume 12, 10 August 2013
GRDI77
challenges affecting Virtual Research Environments are actually very broad and include those characterising
every aspect of a data infrastructure. In fact, Virtual Research Environments are at the higher level in a
conceptually layered architecture of a virtual and scattered system as they represent the application layer that is
built on top of one or more layers offering at least (i) raw resources (e.g., computing, storage, network, and
software resources), (ii) communication and authentication protocols, (iii) protocols for publication, discovery,
negotiation, monitoring, accounting, and payment of resources usage, and (iv) protocols allowing the definition
and management of groups of resources. In the context of a (global) research data infrastructure, the majority of
these challenges are expected to be assigned to the infrastructure itself, i.e., the infrastructure should take care of
putting in place a rich array of mechanisms enabling interoperability with existing systems – conceptually acting
as resource providers – to build a unified space of resources – ranging from data sets, collections, storage
facilities, and computing power to services realising specific utilities and research objects. The richer the array
of interoperability mechanisms the infrastructure is equipped with, the larger the resources space and,
consequently, the domain of “VREs” that can be built.
Sustainability is definitely one of the major challenges affecting Virtual Research Environments development.
VREs require effort and money to be built and maintained according to the communities of practice needs. It is
a waste of effort and money building them without having a long term support although costs can be mitigated
by devising innovative development approaches eventually based on “global virtual infrastructures”. As
proposed in (Carusi & Reimer, 2010), there are three key strategies for sustainability that might be put in place
either singly or in combinations: (i) acquire further funding from diverse research bodies; (ii) develop business
models aiming at self-sustainability; and (iii) rely on community support. However, given the volatile nature of
communities of practice the sustainability issue remains a challenging problem.
Although several Virtual Research Environments have been developed in various application domains and a
plethora of communities of practice are in action, the majority of these systems are not yet fully integrated into
standard practices, tools, and research protocols used by real life communities of practice. This reluctance to
migrate from traditional and consolidated research practices and facilities to the innovative ones promoted by
VREs is among the most difficult barriers affecting the entire VRE domain. As recognised by Carusi and Reimer
(2010), among the factors causing this issue are: (i) the lack of support of both technical (e.g., bug fixing and
further development of the VRE service) and instructional (e.g., training – especially in early stages) nature; (ii)
the gap between the community of practice needs and the actual service implemented by the VRE; (iii) the
reliability of the technology (very often VREs are based on cutting edge and evolving technologies); (iv) legal,
ethical, and cultural issues (the willingness to “share” research outputs and participate in web based research
investigations might be nullified by fear for ownership and attribution); and (v) interdisciplinarity (differences in
“languages” and working practices are a need, a potentiality and an issue as well). The lack of community
uptake has cascading effects on the entire VRE research domain, in particular its impacts on sustainability.
4 RESEARCH DIRECTIONS PROPOSED
Virtual Research Environments represent innovative working environments that aim at enhancing the
cooperation and collaboration among researchers in all modern research scenarios. They promote novel
approaches and facilitate global and timely sharing of research findings, expertise, and any research supporting
“asset” across organizational and operational boundaries and barriers. Because of these potentialities, their
development should be guided by a number of principles and best practices aiming at promoting efficiency and
effectiveness of the resulting services.
A rich array of resources and systems has been developed, and a lot of effort is currently spent in building
infrastructures all over the world including: (a) Internet infrastructures, e.g., GÉANT (www.geant2.net), the
high-bandwidth Internet serving Europe research and education community, and Internet2 (www.internet2.edu),
the network designed to serve the US research and education community; (b) grid infrastructures, e.g.,
European Grid Infrastructure (www.egi.eu), the European Grid Infrastructure built by federating a number of
mainly European providers, and Open Science Grid (www.opensciencegrid.org), a grid infrastructure built by
bringing together computing and storage resources from computers and research communities in the US; (c)
data infrastructures, e.g., DataONE (www.dataone.org), an infrastructure for supporting Earth observational
data mainly in US, Data Conservancy (dataconservancy.org), an infrastructure promoting scientific data
curation, OpenAIRE (www.openaire.eu), an infrastructure promoting the dissemination and sharing on open
Data Science Journal, Volume 12, 10 August 2013
GRDI78
access artifacts including data, and D4Science (www.d4science.org), an Hybrid Data Infrastructure stemming
from a series of EU Funded projects and promoting the realisation of Virtual Research Environments (Candela,
Castelli, Pagano, 2012). Moreover, a lot of momentum has been gained by cloud technologies (Foster, Zhao,
Raicu, & Lu, 2008). All these efforts should be considered as building blocks for realising Virtual Research
Environments. However, to make this possible, services and resources that are aggregated and offered by such
infrastructures should, as much as possible, be independent of a specific application domain and “designed for
reuse”. From scratch and ‘self-sustained’ approaches, e.g., approaches aimed at building the entire spectrum of
the needed resources without ‘outside’ assistance, should be discouraged and prevented because of their
intrinsic development costs and difficulties to deal with evolving scenarios. Actually, Virtual Research
Environments should be linked to existing infrastructures with both roles of consumer, i.e., VREs should benefit
from the services offered by these infrastructures, and provider, i.e., the resources produced in the context of the
VRE operation should contribute to the infrastructures offering.
Virtual Research Environments should be designed, since the beginning, to promote uptake, ensure usability,
and guarantee sustainability. These three aspects form a virtuous circle that, if properly managed, ensure the
success of a specific VRE. In reference to uptake, it is fundamental that the community served by the specific
VRE, although virtual and aggregated by the VRE itself, is provided with tools and facilities for managing and
maintaining the VRE services that have limited requirements with respect to community expertise. Moreover,
the conceivers of the VRE should plan how to engage the broader community of practice that can be served by
the VRE, e.g., it might be possible to build a core team that sustains the VRE itself in the medium and long term
by awareness raising, targeted training, and other engagement events tailored to attract and convince key
representatives of the community of practice. As regards usability, Virtual Research Environments building
should be mainly a community building process rather than a technology development process. This implies that
the focus should be primarily on using technology to identify and rationalise workflows, procedures, and
processes characterising a certain research scenario rather than having technology invading the research scenario
and distracting effort from its real needs. As far as sustainability is concerned, it is fundamental that the
resulting VRE service is conceived as a vital tool in the community of practice it is dedicated to. Moreover,
sustainability is further enhanced whenever the VRE is perceived as a useful tool in the context of larger
research initiatives and communities so to benefit from economies of scale, i.e., savings gained by an
incremental level of production, and economies of scope, i.e., savings gained by producing two or more distinct
goods when the costs of doing so is less than that of producing each of them separately.
5 ACKNOWLEDGEMENTS
The work reported has been partially supported by the GRDI2020 project (FP7 of the European Commission,
INFRA-2009.3, Contract No., 246682).
6 REFERENCES
Allan, R. (2009) Virtual Research Environments: From Portals to Science Gateways. Oxford, UK: Chandos
Publishing.
Assante, M., Candela, L., Castelli, D., Frosini, L., Lelii, L., Manghi, P., et al. (2008) An Extensible Virtual
Digital Libraries Generator. Christensen-Dalsgaard, B., Castelli, D., Jurik, B. A., & Lippincott,J.(Eds.), 12th
European Conference on Research and Advanced Technology for Digital Libraries, ECDL 2008, Aarhus,
Denmark, September 14-19, volume 5173 of Lecture Notes in Computer Science, pp 122-134.
Blanke, T., Candela, L., Hedges, M., Priddy, M., & Simeoni, F. (2010) Deploying general-purpose virtual
research environments for humanities research. Phil. Trans. R. Soc. A 368, pp 3813-3828.
Candela, L., Akal, F., Avancini, H., Castelli, D., Fusco, L., Guidetti, V., et al. (2007) DILIGENT: integrating
Digital Library and Grid Technologies for a new Earth Observation Research Infrastructure. International
Journal on Digital Libraries 7 (1-2), pp 59-80.
Candela, L., Castelli, D., & Pagano, P. (2012) Managing Big Data through Hybrid Data Infrastructures. ERCIM
News (89), pp 37-38.
Data Science Journal, Volume 12, 10 August 2013
GRDI79
Candela, L., Castelli, D., & Pagano, P. (2011) History, Evolution and Impact of Digital Libraries. In Iglezakis, I.,
Synodinou, T.-E. , & Kapidakis, S. (Eds.), E-Publishing and Digital Libraries: Legal and Organizational Issues,
Hershey, PA, USA: INFORMATION SCIENCE REFERENCE.
Candela, L., Castelli, D., & Pagano, P. (2010) Making Virtual Research Environments in the Cloud a Reality:
the gCube Approach. ERCIM News (83), pp 32-33.
Candela, L., Castelli, D., Pagano, P., & Simi, M. (2005) From Heterogeneous Information Spaces to Virtual
Documents. Digital Libraries: Implementing Strategies and Sharing Experiences, 8th International Conference
on Asian Digital Libraries, ICADL 2005, Bangkok, Thailand, December 12-15, 2005, Proceedings. Springer.
Carusi, A., & Reimer, T. (2010) Virtual Research Environment Collaborative Landscape Study. JISC.
Cyberinfrastructure Council. (2007) Cyberinfrastructure Vision for the 21st Century Discovery. National
Science Foundation.
Davies, S. (2011) Still Building the Memex. Communications of the ACM 54 (2), pp 80-88.
De Roure, D., Goble, C., & Stevens, R. (2009) The design and realisation of the myExperiment Virtual Research
Environment for social sharing of workflows. Future Generation Computer Systems (25), pp 561-567.
e-Infrastructure Reflection Group (2010) Blue Paper. E-IRG.
Foster, I. & Kesselman, C. (1998) The Grid: Blueprint for a New Computing Infrastructure. Morgan Kaufmann.
Foster, I., Zhao, Y., Raicu, I., & Lu, S. (2008) Cloud Computing and Grid Computing 360-Degree Compared. In
Grid Computing Environments Workshop, 2008. GCE ’08.
Hey, T., Tansley, S., & Tolle, K. (2009) The Fourth Paradigm - Data-intensive Scientific Discovery. Microsoft
Research.
Hoogerwerf, M., Lösch, M., Schirrwagen, J., Callaghan, S., Manghi, P., Iatropoulou, K., et al. (2013) Linking
Data and Publications: Towards a Cross-Disciplinary Approach. International Journal of Digital Curation 8 (1),
pp 244-254.
Lave, J. & Wenger, E. (1991) Situated Learning: Legitimate Peripheral Participation. New York, NY:
Cambridge University Press.
Llewellyn Smith, C., Borysiewicz, L., Casselton, L., Conway, G., Hassan, M., Leach, M., et al. (2011)
Knowledge, Networks and Nations: Global Scientific Collaboration in the 21st Century. The Royal Society. The
Royal Society.
Nirenberg, J. (1994) From team building to community building. National Productivity Review 14(1), pp 51-62.
Nowakowski, P., Ciepiela, E., Harężlak, D., Kocot, J., Kasztelnik, M., Bartyoski, T., et al. (2011) The Collage
Authoring Environment . Procedia Computer Science 4, pp 608-617.
O'Reilly, T. (2005) What Is Web 2.0 - Design Patterns and Business Models for the Next Generation of Software.
O'Reilly.
Snowdon, D. N., Churchill, E. F., & Frécon, E. (2004) Inhabited Information Spaces: Living with your Data.
London, UK: Springer-Verlag London Ltd.
Van Gorp, P. & Mazanek, S. (2011) SHARE: a web portal for creating and sharing executable research papers.
Procedia Computer Science 4, pp 589-597.
Wang, F.-Y., Carley, K., Zeng, D., & Mao, W. (2007) Social computing: From social informatics to social
intelligence. Intelligent Systems, IEEE 22(2), pp 79-83.
Data Science Journal, Volume 12, 10 August 2013
GRDI80
Wenger, E. (1998) Communities of Practice: Learning, Meaning and Identity. Cambridge, UK: Cambridge
University Press.
Wilkins-Diehr, N. (2007) Special Issue: Science Gateways - Common Community Interfaces to Grid Resources.
Concurrency and Computation: Practice and Experience 19 (6), pp 743-749.
Wulf, A. (1993) The collaboratory oppurtunity. Science 261, pp 854-855.
(Article history: Available online 30 July 2013)
Data Science Journal, Volume 12, 10 August 2013
GRDI81
