Методы распределенного выполнения аналитических запросов над большими объемами ретроспективных географических данных
The paper describes methods and technologies that for the first time enabled the real-time access to large volumes of retrospective climate data for numerous concurrent clients. With graphical 3D interface, a user is able to navigate intuitively in time, visualize data and interactively issue queries for calculating statistical values. The required data for visualization and computing are not stored on the client machine. They are delivered in real-time over Internet from computer cluster. Provided experimental results for several thousand users evidence the effectiveness of the approach. The methods presented in this paper are implemented in client-server system Climate Wikience. It is capable to transfer and visualize in real-time data from ERA-Interim climate reanalysis archive, MODIS L3 satellite data as well as build isolines based on these data. Climate Wikience will enable significantly enhance the understanding of large volumes of retrospective geographical data and provide unprecedented opportunities for Earth science researchers.