Governance Considerations for Southern African e-Infrastructure

By Bruce Becker, South African National Grid, on behalf of SAGrid JRU: Prof. C. J. Wright, Prof. J. W. A. Cleymans, Prof. J. Tolmie, Mr. A. Juyn, Mr. A. Vorster, Mr. I. van Rensburg, Dr. Z. Z. Vilakazi, Prof. F. Joubert

South Africa stands as a potential scientific leader in the Sub-Saharan region, thanks in part to its position on e-Science and related e-Infrastructures. Since these areas are rapidly evolving, technology can leap ahead in terms of functionality and quality of service. However, there are also associated risks: choosing immature technologies, for example, can lead to irrelevance or isolation. How then, are South Africa’s e-Science investments best governed?

CENTRALISED VS DISTRIBUTED
Perhaps a defining feature of e-Infrastructure is that it is distributed. Distribution can be seen to be synonymous with “dilution”; however, it is precisely this feature that provides key strengths: elasticity, resilience and adaptability. A distributed infrastructure is “governed” as opposed to “managed”. Instead of one point of governance with a strict hierarchy, it has a flatter structure, reflecting the nature of activities that e-Infrastructures support: predominantly collaborative, cross-disciplinary and integrative. e-Infrastructures are also enabling, providing researchers who would otherwise be excluded with access to remote and interoperable resources. This is acutely significant in Sub-Saharan Africa, where often there is no such research infrastructure. Such exclusion limits research output and subsequent innovation, negatively affecting society, quality of life and GDP. e-Infrastructures tend to level the playing field, lowering barriers to entry and providing access to more researchers from a variety of disciplines. Thus, a sophisticated governance model is required to recognise these users and the special role that e-Infrastructures play in their activities.

WHY GOVERNANCE?

• Governance structures and policies exist to address issues such as:

• sustainability and certainty of future activities

• inclusion and fairness of use

• quality of service, as well as corrective or preventative action

• engagement of new user communities, to ensure relevance of these services and resources

• long-term development and improvement of services and resources

• appreciation, protection and proper exploitation of research output value, and the encouragement of the production of such

CHOOSING A GOVERNANCE MODEL
Perhaps the best model for the governance of e-Infrastructures is that of a federation, with a (possibly rotating) central component and several confederates providing a similar level of resources. Such a model highlights the cohesive nature of governance. However, we may also use the metaphor of an ecosystem, which highlights the collaborative nature of the infrastructure. Focussing on cohesion and collaboration as fundamentally positive aspects, we reach a scenario where:

• divergence is naturally avoided, since each confederate (organism) is dependant on the others to provide a whole that is stronger and better equipped than could be achieved individually or in any other combination
• a large degree of freedom is permitted to the confederates (organisms), since there is no single point of failure or control; the departure or temporary shortcoming of one does not imply failure for the federation (ecosystem) as a whole, providing resilience

• so long as a minimum level of participation is respected, further participation is encouraged and beneficial, since it adds more resources for a minimal investment by the confederate (organism), relative to the total investment of the federation (ecosystem)

• sustainability is encouraged due to the mutually dependant co-investments made by the confederates (organisms)

• strong coordination and reliable execution of activities by the central component encourages – and stems from – an appreciation of the benefits of collaboration and resource sharing

• well-managed and credible technical coordination encourages uptake of services by user communities, which in turn justifies the initial investment, favours expansion, and contributes to sustainability of the infrastructure

• the self-interest of user communities promotes self-organisation and hence more efficient exploitation of the infrastructures

We have of course painted a biased and rosy picture of the effects of governance in a specific model. Counterpoints not considered here, given the space constraints, are the separation of responsibilities for governance activities, including the need to consider conflicts of interest or expertise.

EXECUTION, EXPLOITATION, EXTENSION AND MANAGEMENT

Governance of public research infrastructures is generally guided by two groups: those providing the service and those using the service. Yet this is not a one-way producer/consumer relationship:

• the consumers often fall under the same institutional boundaries as the producers and are sometimes the same group of people

• the infrastructure is generic, widely appealing, and therefore of great interest across scientific domains

Research support structures exist to satisfy the needs of research communities. Thus governance, no matter its form, should take its cue from these users. An organised and responsible User Form is essential to the ultimate success of e-Infrastructures, as is mutual understanding between the users and providers of e-Infrastructure services.