SkyFed:
Creating a Uniform Earth/Sky Datacube Federation
"Data is the new gold", and this gold keeps pouring down on us from sky, as Earth Observation satellites deliver multi-Terabytes of spatio-temporal data daily.
However, severe obstacles hinder us from optimal exploitation: transmission limits, processing limitations, incompatibilities, and several more.
SkyFed makes the state of the art in Big Datacube Analytics available on edge devices, specifically: on nano satellites.
To this end, the rasdaman datacube engine gets ported to a Raspberry Pi single-board computer
which then gets deployed on a nano-satellite built from inexpensive off-the-shelf components and using common IT standards.
This setup is used for demonstrating federated cloud/edge datacube processing connecting the satellite
with existing multi-Petabyte Earth datacube services, including Copernicus archives and weather forecasts.
Thanks to the strict use of open standards, a wide range of 3rd party clients
can readily tap into this integrated datacube space, allowing users and tools (like edge devices as consumers)
to remain in the comfort zone of their well-known interfaces, despite the heterogeneity of the underlying data.
With its innovative edge intelligence, SkyFed opens the door for integrated data mashups
spanning data centers, cloud, airborne and ship drones, and satellites
in location-transparent federations enabling users to see a single, homogenized information space
ready for zero-coding mix & match of Big Datacubes.
Project Partners
Funding
Runtime 2025/10 - 2027/09, EU EFRE funding support 748731€.
|
