IN31C-1769
EarthCube as an information resource marketplace; the GEAR Project conceptual design

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Stephen M Richard, Arizona Geological Survey, Tucson, AZ, United States
Abstract:
Geoscience Architecture for Research (GEAR) is approaching EarthCube design as a complex and evolving socio-technical federation of systems. EarthCube is intended to support the science research enterprise, for which there is no centralized command and control, requirements are a moving target, the function and behavior of the system must evolve and adapt as new scientific paradigms emerge, and system participants are conducting research that inherently implies seeking new ways of doing things. EarthCube must address evolving user requirements and enable domain and project systems developed under different management and for different purposes to work together. The EC architecture must focus on creating a technical environment that enables new capabilities by combining existing and newly developed resources in various ways, and encourages development of new resource designs intended for re-use and interoperability. In a sense, instead of a single architecture design, GEAR provides a way to accommodate multiple designs tuned to different tasks. This agile, adaptive, evolutionary software development style is based on a continuously updated portfolio of compatible components that enable new sub-system architecture. System users make decisions about which components to use in this marketplace based on performance, satisfaction, and impact metrics collected continuously to evaluate components, determine priorities, and guide resource allocation decisions by the system governance agency. EC is designed as a federation of independent systems, and although the coordinator of the EC system may be named an enterprise architect, the focus of the role needs to be organizing resources, assessing their readiness for interoperability with the existing EC component inventory, managing dependencies between transient subsystems, mechanisms of stakeholder engagement and inclusion, and negotiation of standard interfaces, rather than actual specification of components. Composition of components will be developed by projects that involve both domain scientists and CI experts for specific research problems. We believe an agile, marketplace type approach is an essential architectural strategy for EarthCube.