Architectural solution for virtualized processing of big earth data

The Earth Observation data repositories are increasing each day by several terabytes. There is a need to provide efficient techniques to store and process this huge amount of data in order to provide end-users with valuable information and knowledge. The architectural solution proposed in this paper is based on cloud virtualization and aims to provide a flexible and adaptive method to extract and highlight knowledge from the huge data of Earth Observation images. The problem of data processing is solved by virtualizing algorithm objects which are placed in key positions on the hardware level. The users could describe and experiment complex use cases and take advantage of the improvement on execution performance provided by the flexible description and the adaptive processing on high performance computing platforms.