Satellite Data Processing: Making sense of extreme data volumes generated by satellites
During the lifetime of earth monitoring satellites from ESA or NASA large amounts of data is collected, which accumulates to the range of many 10s or 100s of terabytes. Analysing this data can be very compute intensive and require supercomputing resources. Applying several analysis methods to the same data results in it being read multiple times. Such applications will therefore benefit from being able to keep a copy of the data in fast I/O layers and the ability to pre-process this data while it is still within the storage system.
Time-to-solution of the analysis on the order of 100s of terabytes of satellite data can be significantly improved as it allows for efficient pre-processing and provides the resulting input for data analysis in a high performance tier.
The picture below shows a satellite data processing example ( “Radiative transfer” model results based on earth observing satellite data).