Title of Presentation: Collaborative Virtual Sensorweb Infrastructure: Architecture and Implementation
Primary (Corresponding) Author: Prasanta Bose
Organization of Primary Author: Lockheed Martin, ATC,
Co-Authors: Neal Hurlburt - Lockheed Martin, ATC, Palo Alto; Doug Shimokawa - Lockheed Martin, ATC, Palo Alto; Ankur Somani - Lockheed Martin, ATC, Palo Alto; Peter Fox: UCAR
Abstract: Geoscience and space science research involving exploitation of insitu sensor networks and remote observing instruments offer an unprecedented opportunity for coordinated observation and science data processing critical to understanding and timely prediction of climate and (earth and space) weather changes . Current efforts in specialized virtual sensors or observatories targeting the atmosphere, ocean, carbon management, space weather and etc. domains raise the common challenge to create a collaborative sensorweb infrastructure that coordinates sensor resources and data products from individual observatories for effective and robust science processing and timely weather prediction. The challenges to creating such an infrastucture include model-based (model-predictive) integration of processing services and data spanning multiple sources, intelligent push of the data in an efficient, timely manner and closed-loop decentralized management and reconfiguration of the sensing resources to meet dynamic science and prediction needs. Semantic descriptions (via OWL-S, SensorML) of services and resources via ontologies and semantic markup languages provides a basis for the development of middleware infrastructure enabling intelligent and flexible (goal-driven) management of the services and physical resources. Our NASA funded Collaborative Sun-Earth Connector (CoSEC) project and Virtual Sensor Web Infrastructure for Collaborative Sensing (VSICS) project has been developing a common scaleable architecture and agent-based (event-driven) coordination (e.g. workflow management) and control (e.g. planning and scheduling) services for such an infrastructure for application to the space science and the geosciences domain respectively. This paper will present features of the architecture, and current implementation of the infrastructure applied to collaborative sensorwebs.