Optimization Tools Using Virtual Reality

Research Overview

Research Team:  Jurgen Schulze, Assistant Project Scientist, Calit2; Tom DeFanti, Research Scientist, Calit2; Falko Kuester, Associate Professor, Structural Engineering, UC San Diego

This group of computer scientists and their colleagues in biology and geoscience will target the complexity of system-of-systems architecture and integration and resource management within the GreenLight Instrument, applying domain modeling to capture the power utilization as well as all other functional and non-functional requirements; infrastructure architecture and design; implementation and deployment and quality assurance. A loosely coupled service-oriented architecture (SOA) will integrate multiple hardware and software components of the instrument, allowing specific policies to be enforced by pluggable entities enabling on-demand monitoring and control of the GreenLight Instrument. SOAs rise to the integration challenge; they use standards-based infrastructure to forge large-scale systems out of loosely coupled, interoperable services. The SOA middleware separates scientific applications from the underlying physical infrastructure, creating a series of new web services that enable secure replicated data storage, and portal interfaces to the investigators' data, workflow distribution tools, and visual analysis. SOAs can further create software systems-of-systems by mapping existing systems into services, then orchestrating communication between the services. This group is developing SOAs for the recently established NSF Ocean Observatories Initiative (OOI), CAMERA Ocean Microbial Genomics, and the large-scale Proteomics projects.

To properly monitor and control the energy use of the GreenLight Instrument, the domain model will take advantage of a wide range of physical infrastructure sensors available in the BB, and applied by Rosing's group along with the internal measurement capabilities of the PCs (e.g., CPU voltages, operating frequencies, memory timings, enabled peripherals, CPU load and memory usage levels, disk activity, network interface activity, etc.). Portals will expose attributes of the GreenLight Instrument to domain scientists to help tailor their experiments to resource expenditure. Also, requests for additional computational, data, or communication resources could dynamically be used to reconfigure the GreenLight Instrument to provide the requested resources in a less power consumptive resource budget.

Technical White Papers

Virtual Reality 2011
Virtual Reality 2010
Virtual Reality 2009