SORASCS (Service Oriented Architecture for Socio-Cultural Systems) is a project to develop a coherent, flexible, extensible data-to-model service oriented architecture for socio-cultural modeling and analysis to support military intelligence and modeling community. Enables reach-back cells, military analysts, and field operatives to rapidly make use of tools and models from diverse sources. Supports model re-use through sharable reports and configurations. Supports rapid terrain assessment. Supports strategic planning by reducing gap between field/ethnographic data and models.
Over the past decade there have emerged a large variety of tools supporting analysis of large volumes of heterogeneous information to understand complex real world relationships and trends, including natural language processing, network analysis, simulation, and what-if reasoning. Most analysis ensembles assume a relatively limited set of model types and input sources. They are wired together with special purpose, tool-specific and ad hoc integration code, making them brittle and requiring low-level systems expertise to reconfigure them in new ways or add new capabilities to the system. The underlying problems stem from the lack of a coherent and flexible architectural framework that will allow tools and models to be seamlessly incorporated and composed by end users. Having such a framework would ideally allow information analysts to bring together tools from various vendors and research groups, dynamically compose their analyses in new ways, store and reuse previous workflows and settings, ingest new forms of information, and interactively fine tune analyses, simulations, and report generation through consistent and uniform forms of interaction.
We are developing an architecture that will be useful with both web-based and non web-based models and analysis tools. As part of this effort, we are populating the architecture with existing thick-client systems and to creating a web-application version of several analysis and modeling tools, for use by combatant commands. The architecture will be a light, extensible, community based system serving as the glue for linking a set of heterogeneous data sources and tools needed for modeling and analysis to support the needs of the military. We initiate this architecture with a limited number of tools and to engage others to add in their tools. These tools will include thick and thin versions of the CMU CASOS tools, ORA, AutoMap, and Construct. The basic architecture is a necessary step for an overall system that we expect to grow incrementally. Additional common services, analysis and modeling tools will be added in the future by CMU and others. The goal here is to provide the system into which these items can be integrated with minimal effort on the part of future developers, and with an emerging ontological base.