Facts and Figures

• 110 VM's are now hosted by our Openstack 'cloud' infrastructure
• Our Hadoop processing cluster now has 1969 cores and 6.27 TB memory available
• Thanks to tweaking to our scheduling, each user can still get hold of some resources in a fair amount of time.
• The Hadoop processing cluster now runs 6 operational workflows next to user submitted jobs.

Sentinel Data

Next to the PROBA-V data set, we now also offer a selected set of Sentinel 1 and Sentinel 2 data, over Belgium. This allows users to combine the Sentinel data with other data sets available on the platform.

Operational workflows

To support the needs of operational workflows, some features have been added:

• support for Docker containers, this allows users to create an isolated environment in which they can run their algorithms.
• support for the Airflow workflow system. This allows users to combine jobs into a single airflow workflow.

Future plans

With simple concepts like VM's and Notebooks with a direct link to an archive of analysis ready data, we've managed to simplify the use of remote sensing data.

However, these tools still require a lot of boilerplate work from the user. Especially when the datasets become large, and are updated frequently, this can become quite a bit of work. Therefore, we need to go a step further. One example is the on-demand timeseries web service.

If you are new to this web service, you can find a notebook demonstrating it's usage here.