Cyberinfrastructure Research at CMOP
CMOP has a commitment to integrate systems, education, research, and culture. We view cyberinrastructure as the substrate and catalyst for successful integration of multi-disciplinary Research, Education, and Knowledge Transfer, guided by an open data policy.
- eScience Appliances
Through a partnership with Microsoft Research, we are exploring an architecture for provisioning software and data based on appliances -- self-contained, specially-purposed hardware and software platforms that can be deployed as one unit.
We have two appliance projects underway.
The CMOP Data Appliance
The CMOP Data Appliance bundles a scientific workflow engine (Trident from Microsoft Research) with observations and model results from the SATURN observatory operated by CMOP to provide a vector for data delivery and systems research, a gallery of extensible workflows, and training materials.
The IOOS Appliance
What's the best way to distribute specialized software to organizations with limited IT budgets? We recommend this: Buy some hardware, load your software on it, and deliver the whole platform as a unit.
We are experimenting with this idea in three ways: First, as an inexpensive means of bootstrapping IOOS in the Northwest Region; second, as a rapidly-deployable and security-conscious method of providing real-time telemetry from vessels and platforms that we do not directly control; and third, as a vector for translating research projects in Cyberinfrastructure into tangible products without incurring the enormous software support costs.
- Gridfields: Algebraic Manipulation of Simulation Results
Simulation results in the physical sciences represent a continuous domain as a discretized grid structure. These structures are not amenable to management using conventional techniques such as a relational database. We develop a language for general construction and manipulation of these gridded datasets that can express complex transformations succinctly and affords algebraic optimization reasoning.
- Quarry: Dataspace Profiling
Through a collaboration with Portland State University researchers, we are addressing the problem of "bootstrapping" a data management application: How does one proceed from heterogeneous, unfamiliar data sources to useful knowledge, in hours rather than weeks or months?.
- Vistrails: Visualization and Provenance Management
We are collaborating with the Scientific Computing and Imaging Institute at the University of Utah to develop advanced 3D visualizations of CMOP models and data using the SCI provenance and visualization platform, VisTrails.
- Integrated Database
We are developing a database architecture to manage arbitrary oceanographic and biological observational data in a uniform manner. We use advanced features of the PostgreSQL Relational Database Management System to implement a bi-level schema: an integrated schema linked to multiple application-specific schemas. This approach is a realization of the "global-as-view" model of information integration. For more information, see the attached presentation given to the Portland PostgreSQL user's group.
- Service-Oriented Architecture and Rapid Application Development
A service-oriented architecture encourages reuse, lowers maintenance costs, enables platform independence, and, most importantly, empowers external users to access our data and build applications of their own.
In addition to developing our own protocols for accessing oceanographic data, we also evaluate proposed interoperability standards for their expressive power and ease of use. Currently deployed services include the following.
- Product Factory providing a system for serving dynamic plots over the web based on arbitrary queries against the integrated database.
- Sensor Observation Service providing access to our integrated observation database.
- Web Mapping Service (WMS) for forecasts and hindcasts.
- Profile Probe Service for accessing model data at arbitrary lat/lon coordinates.
Leveraging our web services accessing the integrated database, we are able to design, implement, and deploy new applications quickly and conveniently.
External Applications