e-IRGSP Governance Report

\chapter{e-Infrastructure Governance and IT Governance}

\label{chapter:4}

In general, e-Infrastructure governance may not be identical with IT

governance. In this Chapter we investigate the main points of

e-Infrastructure and where they may differ from corresponding IT

governance. Although sibling fields, it is not obvious that findings

in the IT governance literature apply directly to e-Infrastructure

governance. 

\section{e-Infrastructure Governance}

e-Infrastructure governance has been discussed in the 7th

e-Infrastructure Concertation Meeting and the e-IRG Workshop in

Brussels, in 2009 and 2010 respectively. It is framed as pertinent to

funding sustainability and promotion of shared interests but at the

same time it is not clearly defined. The discussion around governance

has been fragmented around legal frameworks, financial investment,

management, data ownership and decision-making \cite{Potts}. It is

further associated to transparency and accountability in resource

provision, policy definition and implementation,

centralised/decentralised structures, balance between providers and

users, as well as trust between all stakeholders \cite{Sawley}.

e-Infrastructure governance is also called to resolve intellectual

property and liability issues \cite{PARADE}, to promote open access

principles, common standards while supporting the distributed nature

of information \cite{ConcertationReport}. Although these elements can

only be interrelated, a stronger definition and demarcation of the

different domains of e-Infrastructure governance is required to

provide a strong basis for discussion. At the same time, different

levels of national or organisational governance, as well as the

diverse requirements emerging from the rich landscape of

e-Infrastructure types add to the complexity.

\section{Actors and Proposed e-Infrastructure Governance Models}

There seems to be general consensus regarding the actors involved in

e-Infrastructure governance. The \ac{parade} suggests three

stakeholder groups involved in high-level, strategic decision-making

(see Figure below): (1) user communities providing input on required

activities, (2) service providers responsible for the technological

issues, (3) funding bodies and research agencies ensuring

cost-effective and sustainable operation. In their proposed governance

structure, the Executive Board consists of members from all three

groups and retains the responsibility for daily operations and

implementation, supported by a project team and an International

Advisory Board \cite{PARADE}.

(PARADE proposed governance structure here)

During the 7th e-Infrastructure Concertation Meeting, three governance

models were proposed, distinguishing between different types of

e-Infrastructures: \acp{nren}, Grids/Clouds and \ac{hpc} facilities

\cite{Sawley}. The governing bodies and mechanisms in different

decisions areas, some of which map to the five key IT governance

decision areas, are outlined in

Table~\ref{tab:3communities}~\cite{Sawley} :

\begin{figure}

\begin{center}

\caption{Proposed governance models for 3 communities} 

\label{tab:3communities}

    \begin{tabular}{ | p{2,5cm} | p{2,5cm} | p{2,5cm} | p{2,5cm} |}

      \hline

      & National \acp{nren}s & Grids/clouds & \ac{hpc} facilities \\ \hline

      Investment & National & Facility operator & Public grants, public/private partnerships \\ \hline

      Technology choices & Coordination bodies & Facility operator & Facility operator (mandate from the community, less general purpose) \\ \hline

      Access to the resources & Negotiation with contractual institutions & Voucher system & Community (peer review), facility operator for the implementation \\ \hline

      Interoperability & Maximal, Transnational coordination & Professional: Accreditation through standardizing bodies, SLAs & Not more than required by scientific needs \\ \hline

    \end{tabular}

\end{center}

\end{figure}

At a national level, the UK Research Councils' report on

e-Infrastructures promotes leadership and coordination along the lines

of the e-IRG. They recommend that ``[t]he leadership of the UK's

Research and Innovation e-infrastructure should be delivered by a

Director of e-infrastructure, overseen by a Strategic Board with

members drawn from areas closely related to e-infrastructure, such as

software specialists, research technologists, and specialists in data

management''~\cite{UKResearchCouncils}. This report makes evident the

need for collaboration between different e-Infrastructures and their

associated organisations at a national level. Countries attempt

cost-effective development of national infrastructures with the aim to

become leaders in the field. However, in e-Infrastructures funding

bodies can be diverse and do not necessarily coordinate their efforts

or align their objectives. Appropriate governance would provide for

the required synergy in investment, ensure interoperability between

the installed base and new e-Infrastructures, as well as maximise

re-use of available tools by different research disciplines, although

there is no one approach suitable for all cases \cite{Baxevanidis,

  eResearch2020, UKResearchCouncils}. Overall, these characteristics

distinguish e-Infrastructure governance from traditional IT governance

that focuses on internal governance arrangements. Other national

programmes of interest include the Dutch e-Infrastructure programme

and the legal framework developed in Germany \cite{Dutch, Dutch2,

  Karagiannis}.

\section{e-Infrastructure Governance and IT Governance}

\subsection{Governance Forms}

The eResearch2020 project (\url{http://www.eresearch2020.eu/}) was

funded by the European Commission to study the role of

e-Infrastructures in the creation of global virtual research

communities, from an organisational, collaborative and technological

perspective. The research consortium consisted of empirica

Communication and Technology Research in Germany, Fachhochschule

Nordwestschweiz in Switzerland, the Oxford Internet Institute in the

United Kingdom and the National Opinion Research Centre at the

University of Chicago in the United States. Their final report

acknowledges, among other things, the significance of e-Infrastructure

governance and its impact on research communities. Drawing on in-depth

qualitative interviews and archival analysis, the eResearch2020 report

presents, among other findings, information on the governance

arrangements of 18 e-Infrastructures and related

organisations~\cite{eResearch2020}. \footnote{The e-Infrastructures

  studied include: \ac{deisa}, \ac{eela-2}, \ac{egee}, \ac{geant},

  \ac{osg}, TeraGrid, Swedish National Data Service, \ac{c3grid},

  CineGrid, \ac{clarin}, \ac{d4science}, \ac{dariah}, \ac{driver},

  \ac{etsf}, MediGrid, \ac{nvo}, Swiss BioGrid and \ac{ogf} as a

  standardisation organisation.}

Similarly to the IT governance literature, it is recognized that the

diversity of e-Infrastructures creates a variety of governance forms.

Governance strategies emerge in different shapes, ``from the small and

informally organized (CineGrid is an example) to larger multi-tiered

and more elaborate and complex structure[s] (\ac{geant})''

\cite{eResearch2020}. However, it is noted that only large

e-Infrastructures distinguish between strategic oversight and

management.

The centralisation debate is part of the discussion on

e-Infrastructure governance as well. In the 7th e-Infrastructure

Concertation Meeting, grid and research network governance models,

especially that of \ac{geant} and \acp{nren}, were compared and

contrasted to \ac{hpc} models such as \ac{prace}. In the case of

\ac{prace} a fine balance is achieved between decentralised and

centralised arrangements, since decisions are taken at the lowest

levels possible, using local expertise and funding, but at the same

time there is a single contact point to ensure stability of the

infrastructure and coherent procurement plans based on pan-European

peer reviews \cite{Eickermann}.

The empirical work of the eResearch2020 project found more centralised

forms of governance with either a few central coordinators or a

governing body, usually in larger projects. Federated organisations

were also encountered as in the case of \ac{ogf}, TeraGrid, and

\ac{osg} with different coordinators for different tasks but it should

be noted that such arrangements exist also at lower levels in more

centralised forms (\ac{ogf} is included here as a standardisation body

coordinating the efforts of different e-Infrastructures, rather than

as an actual e-Infrastructure in itself). The authors of the report

conclude: ``What we see organizationally is a range from highly

centralized and hierarchical governance to more ?flat? or federated

governance which may have multiple centres'' \cite{eResearch2020}.

Although the variety of forms is recognised, there is no other

classification apart from the centralised/decentralised continuum, as

in the IT governance literature. Further study is required to

distinguish more nuanced e-Infrastructure governance forms.

\subsection{Governance Mechanisms}

Explicit identification of formal and informal governance mechanisms

employed in e-Infrastructures could bring to light and systematise

practices that are not already appreciated as governance efforts per

se.

The e-Infrastructures studied in the eResearch2020 project usually

employed advisory and/or steering committees to provide guidance or

guarantee that all interests are represented. This was defined as the

metagovernance layer or ``governance which coordinates the governance

of individual efforts'' across institutional boundaries. Again, there

are many variations in different cases ``informal and flexible (Swiss

BioGrid is an example) or highly complex, formally institutionalized

and hierarchical (\ac{egee} serves as an example)''

\cite{eResearch2020}. Governing bodies consist of researchers from

within the same, such as the \ac{nvo} or different disciplines

(Swedish National Data Service) and they are either permanent or

involve their members in ``episodic governance'' \cite{eResearch2020}.

Internal and external communication are frequently discussed as

important aspects of e-Infrastructures, but they are usually mentioned

as separate efforts, not as clearly defined elements of governance.

More emphasis is placed on structures such as governing bodies and

advisory groups, rather than on their actual functioning and

coordination. Although communication is taking place in various ways,

especially in larger e-Infrastructures, there remains the question

whether these are designed to be strategically aligned to the

objectives, investment efforts and other needs. It might be useful to

adopt the approach of the IT business literature, where internal and

external communication are described as explicit governance

mechanisms, part of carefully planned, strategic governance

arrangements.

With regards to internal project communication, the eResearch2020

project reports the use of basic mechanisms like teleconferencing and

emails as in the case of SwissBioGrid, but also finds examples where

more regular communication and extensive meetings are taking place.

Differences in culture, and more specifically field affiliation and

organisational identity, pose obstacles to effective collaboration

between internal partners and organisations working on the same

e-Infrastructure, while competitive behaviours reduce trust and

productivity. In addition, technological specificities can hinder

inter-organisational integration. Of course, the extent of these

problems depends on the level of involvement required by the partners.

In long term collaborations less friction is observed, while improved

coordination and communication (which can be attributed to successful

governance mechanisms) contribute to enhanced collaboration. Complex

arrangements such as those of TeraGrid are reported as most

effective \cite{eResearch2020}. TeraGrid is primarily governed by two bodies: the TeraGrid Forum and the Grid Infrastructure Group. Principal investigators from the 11 partners set governance and policy in the former, while the latter comprises of the management team, Area Directors and expert working groups to plan, manage and coordinate the work. These distributed governance arrangements provide for non-direct supervision of individuals in different sites \cite{eResearch2020}. 

External communication with users is deemed to be a more complicated

issue and one of the topics discussed in the e-IRG Blue Paper

\cite{BluePaper}. There is no uniform approach to external

communication in e-Infrastructures, although it is of great importance

to project sustainability and success. Some projects have

sophisticated user engagement plans even in their development phase

(e.g. \ac{clarin}, \ac{egee}, \ac{osg}), when others pay much less

attention to these issues \cite{eResearch2020}. Moreover, there are

challenges in intra- and inter-disciplinary collaboration between

developers and users, between actors in different research areas, or

even between developers in different domains \cite{Barjak,

  eResearch2020}.

Related to external communication is the process of user recruitment.

Although users determine the sustainability of e-Infrastructures, the

eResearch2020 project finds that user relationships are not always

prioritised. Reasons cited include: focusing on pure technology

development and delaying user engagement, perceiving the

e-Infrastructure mission as one that excludes take-up, as well as

estimating a great probability of project discontinuation.

A parallel to IT governance can be found here in terms of how IT

departments in companies communicate with business units who are

usually the users of IT systems and infrastructures. Such user

engagement might contribute to avoid technology push and similar

problems. User recruitment can also be encountered in corporate

settings, when IT departments attempt to persuade business units to

use certain technologies rather than requiring the development the new

tools or purchasing their own solutions. Appropriate positioning on

the governance forms continuum as well as the adoption of effective

governance mechanisms would further the objectives of

e-Infrastructures in terms of user recruitment, engagement and

sustainability.

External communication is also taking place between different

e-Infrastructures which coordinate their work and standards to develop

interoperable platforms \cite{eResearch2020}. One example is provided

by \ac{geant} which ``facilitates discussion between the co-operating

\acp{nren} and with the responsibles for the equivalent

e-Infrastructures outside of Europe to ensure a harmonised,

interoperable evolution of the networks, management and monitoring

tools and end-to-end services'' \cite{ConcertationReport}.The

\ac{ogf}, as a coordinating standardisation body, employs a

``\emph{formal} liaison function'' with individuals from member

organisations monitoring developments in other bodies where

standardisation work evolves in parallel or competing areas.

Furthermore, \emph{informal} relationships and collaboration play a

major role in the development of interoperable specifications and

standards, in order to provide seamless access to distributed

resources (e.g. \ac{nvo} and similar astronomical infrastructures).

However, provider interoperability is not always feasible through such

mechanisms. In cases where e-Infrastructures have developed distinct

software or middleware, higher levels of technological complexity need

to be overcome with common tools or portals (e.g. ``Science Gateways''

in TeraGrid)~\cite{eResearch2020}.

Absence or inadequacy of governance mechanisms might exacerbate issues

such as limited communication and understanding between technical

developers and research communities, creating problems in eliciting

user requirements, something commonly encountered by the eResearch2020

project. Further, governance uncertainty is one of the reasons data

sharing remains restricted in repositories and similar infrastructures

\cite{eResearch2020}.

\subsection{Drivers of e-Infrastructure Governance Forms}

As criteria on e-Infrastructure governance forms remain uncertain, the

business literature on multiple contingencies can provide a useful

starting point when adapted towards this direction.

As discussed in Chapter~\ref{chapter:3}, single contingencies cannot

adequately explain the adoption of certain governance forms. The

theory of multiple contingencies relates IT governance to drivers

grouped in the larger categories of corporate governance, economies of

scope and absorptive capacity. In e-Infrastructures, especially

smaller ones, corporate governance might overlap with e-Infrastructure

governance as a concept and practice and therefore, it cannot be perceived as a determinant of e-Infrastructure governance. Of course, an overlap between IT and corporate governance can also be encountered in other organisations apart from e-Infrastructures

\cite{Grembergen2003}. The size of e-Infrastructures might be another

factor correlating to more mature e-Infrastructure governance,

similarly to what we see in IT governance, but it would not predict a

tendency towards more or less centralised forms. CineGrid and

\ac{geant} are both large e-Infrastructures with significantly

different governance arrangements \cite{eResearch2020}. Economies of

scope might be relevant to e-Infrastructures when looking at how the

infrastructure and its tools are used by more than one research

communities or fields, leading to either decentralised or centralised

decision rights. In addition, absorptive capacity relates to the

expertise of the members at the various levels of e-Infrastructures

development and use. Less technically sophisticated users might favour

centralised modes of governance, whereas in highly technical fields

decentralised forms might be more relevant.

However, these hypotheses are difficult to validate with the data

currently available and further research could highlight the multiple

interacting drivers behind e-Infrastructure governance forms and

mechanisms. Drivers that are e-Infrastructure specific and might be

more relevant to explain and further recommend appropriate governance

forms include funding sources, research objectives, research fields,

interdisciplinary collaboration or partnerships between academia and

industry etc.

\subsection{Further Similarities and Differences}

When Weill and Ross discuss governance arrangements in governmental

and not-for-profit organisations, they recognise similarities with

for-profits but also acknowledge that there are difficulties,

especially in assessing performance and value, as well as in

justifying IT infrastructures and investments. Within a political or

similarly restricting environment, there are usually mandates over and

above business value and cost-benefit considerations, such as value

creation in terms of public good or other objectives~\cite{WeillRoss2004b}.

Funding sources are inextricably linked to e-Infrastructure governance

arrangements. The governance panel at the 7th Concertation Meeting

discussed how the public funding model restricts options and how

further complexity stems from the interconnection of local and

national systems and networks, which might already be centrally

controlled. Other points raised include the need to emphasise on

return on investment, especially when commercial involvement is

desirable~\cite{ConcertationReport}.

In addition, dependency on external stakeholders and beneficiaries to

use and contribute to the infrastructure, dataset or service may mean

increased input from these parties in decision-making structures, such

as e-Infrastructure governance forms or mechanisms. In this domain,

it is observed that top not-for-profit performers clarify how

their organisation creates value, recognise the capabilities required

and strive to achieve alignment with governance. To this end,

guidelines drawn from their five key decision areas and adapted for

effective e-Infrastructure governance include~\cite{WeillRoss2004b}:

\begin{itemize}

\item making joint user and technical decisions around

  e-Infrastructure principles and investment

\item viewing e-Infrastructure decisions as strategic decisions and

\item avoiding feudal, autonomous decentralising models for research

  communities needs

\end{itemize}

Although governance forms differ between e-Infrastructures, it is not

clear how the same e-Infrastructure might use different governance

structures and mechanisms. A nuanced classification of governance

forms that can be applied at the same time within one e-Infrastructure

to address issues in different decision areas might be a useful step

forward. As in IT governance, an organisation might use centralised

models to decide on IT principles and decentralised models to solve

business applications needs.

In addition, IT governance benefits from processes and control

frameworks like \ac{cobit}, \ac{itil} etc. It might be possible to

adapt these frameworks to serve the purposes of e-Infrastructures or

to develop other specific tools to support decision-making and

monitoring in a formalised manner. Furthermore, larger

e-Infrastructures might lead the way in creating maturity stages and

models of e-Infrastructure governance, against which new or less

developed projects might benchmark themselves. Related to the problem

of user communication, balanced scorecard have long been implemented

in organisations to measure user satisfaction and strategic alignment.

Such a practice might be useful to e-Infrastructures and their

communication with the usage base.

Relational mechanisms such as direct informal contacts, common locations and ``virtual meeting points'', lobbying, joint performance incentives, shared learning or more formal senior management involvement as mobilised in IT governance (see Chapter~\ref{chapter:3}) might prove helpful in collaboration and expertise transfer in e-Infrastructures,

most of all in what relates to tacit differences between fields and

specialisations. Formal and informal communication of strategic

objectives and associated governance mechanisms would also contribute

to problems currently encountered.

\subsection {\ac{eric}}

Governance lessons from business and managerial areas might contribute

to e-Infrastructure governance. However, more issues remain to be

solved, such as the inadequacy of legal frameworks for

e-Infrastructures, which has an impact on governance, among other

things.

Different European legal forms for e-Infrastructures have been

discussed in the first e-IRGSP2 Legal Issues Report \cite{SP2ReportA}.

Among these the~\ac{eric} model has been described as a legal

instrument to facilitate the establishment and operation of

e-Infrastructures in Europe. Together with specifying a legal

personality and tax benefits, \ac{eric} also includes provisions for

e-Infrastructure governance. A significant amount of flexibility and

adaptability is provided to respond to the needs of European

e-Infrastructures. In terms of internal governance, the creation of

only two bodies is mandatory~ \cite{SP2ReportA, Thies2009}:

\begin{itemize}

\item the members' assembly with full

decision-making rights, also with regards to budget decisions, 

\item the

director or board of directors, which would be appointed by the

assembly of members to legally represent the e-Infrastructure and

function as its executive body. 

\end{itemize}

Although these two governing bodies seem to impose a centralised form

to e-Infrastructure governance, their actual implementation and

complementation with other structures and mechanisms in different

layers would determine the final result.

According to the eResearch2020 report, \ac{eric} does not address all

obstacles and is primarily suited to non-commercial entities.

Therefore the authors recommend further policy development to overcome

remaining issues. How the legal framework will be adopted and in what

ways it will work in practice to enforce governance arrangements still

remains to be seen.

\section{Conclusion}

This Chapter presents e-Infrastructure governance as discussed in the

e-Infrastructure community and draws on IT governance literature to

find parallel developments and divergence regarding the adoption of

appropriate forms and the mobilization of relevant mechanisms.

Although the IT governance field can provide e-Infrastructure

governance with useful ideas, these may not be directly applicable but

require contextualisation and adaptation in different cases. The

nature of e-Infrastructures and their technical and organisational

characteristics may diverge from more clearly delineated internal

organisational IT systems. However, it should be noted that the

literature on IT governance also refers to and derives from empirical

data and research in multinational conglomerates and business

ecosystems that might present more similarities with

e-Infrastructures. As in the case of different companies where the

same governance model or mechanism may bring opposite outcomes when

implemented in different contexts, specific e-Infrastructures need to

find appropriate solutions that work within their own domain of

opportunities and limitations, while realising that governance

arrangements and needs are not static but evolve and grow. Towards the

effort for effective e-Infrastructure governance, the IT literature

contributes its emphasis on the importance of clear and coherent

planning, while at the same time providing a wealth of previous

experience on forms and mechanisms to draw on.

%%% Local Variables: 

%%% mode: latex

%%% TeX-master: "governance"

%%% End: