The 8th IEEE International Workshop on Trusted Collaboration

Research Article

Accurate Weather Forecasting Through Locality Based Collaborative Computing

Download577 downloads
  • @INPROCEEDINGS{10.4108/icst.collaboratecom.2013.254178,
        author={B\ae{}rd Fjukstad and John Bj\`{u}rndalen and Otto Anshus},
        title={Accurate Weather Forecasting Through Locality Based Collaborative Computing},
        proceedings={The 8th IEEE International Workshop on Trusted Collaboration},
        publisher={ICST},
        proceedings_a={TRUSTCOL},
        year={2013},
        month={11},
        keywords={weather forecast distributed computing collaboration peer to peer},
        doi={10.4108/icst.collaboratecom.2013.254178}
    }
    
  • Bård Fjukstad
    John Bjørndalen
    Otto Anshus
    Year: 2013
    Accurate Weather Forecasting Through Locality Based Collaborative Computing
    TRUSTCOL
    ICST
    DOI: 10.4108/icst.collaboratecom.2013.254178
Bård Fjukstad1,*, John Bjørndalen1, Otto Anshus1
  • 1: Department of Computer Science, University of Tromsø, Norway
*Contact email: bard.fjukstad@uit.no

Abstract

The Collaborative Symbiotic Weather Forecasting (CSWF) system lets a user compute a short time, high-resolution forecast for a small region around the user, in a few minutes, on-demand, on a PC. A collaborated forecast giving better uncertainty estimation is then created using forecasts from other users in the same general region. A collaborated forecast can be visualized on a range of devices and in a range of styles, typically as a composite of the individual forecasts. CSWF assumes locality between forecasts, regions, and PCs. Forecasts for a region are computed by and stored on PCs located within the region. To locate forecasts, CSWF simply scans specific ports on public IP addresses in the local area. Scanning is robust because it avoids maintaining state about others and fast because the number of computers is low and only a few forecasts are needed.