We are facing the problem that we have multiple projects with multiple high-resolution orthophotos from drones (1 photo > 2GB). We are using Geoserver with the integrated GeoWebCache. As a storage container we are using the ImageMosaic with a temporal enabled config (time-series) storing the references to each ortho (granule) in a PostGre table. We enabled a TIME filter (dimension) for the GeoWebCache, so we can query different timestamps. Projects are created every now and then, so the amount of orthophotos increases over time.
The data is only consumed via the GWC (WMTS, TMS) on different clients.
The problem we are facing is that even though our VM has 32GB of RAM with 8 cores (Windows Azure VM), processing multiple photos in high resolution takes a lot of time.
So my idea was to create some kind of GeoWebCache Docker container which can get spawned multiple times depending on the amount of orthophotos to be processed.
Our whole infrastructure is on Azure so I would like to use Azure Container Instances for the Docker Container and a Storage Account (File Storage) as the shared Data Container which is mounted on our VM to GeoServer as a NetworkDrive and to the GWC in the Docker Container.
My questions now:
- Does the whole concept make sense?
- Has anyone already done something like that?
- I have found some posts about starting multiple GeoServer Instances - would that be a better solution? (For me only multiple GWC make sense as I only need to create the caches)