Petascope - too slow to open connections to SECORE for loading CRS GMLs

[Bang Pham Huu]

In case with many coverages in different CRSs (e.g: Sentinel 2 UTM zones), petascope starts and takes minutes or more than that to send HTTP requests to SECORE for getting the CRSs GML.

That is not good as it takes too long time for the result can be shown in WSClient. When the CRSs are cached in petascope, it returns in seconds though.

Either SECORE and Petascope should be enhanced for this case with simultaneous CRS requests.

[Peter Baumann]

would it make sense to retrieve the root = all defs in one request?

Independently, for the case when we operate a local SECORE, could we establish a tighter coupling without http requests?

[Dimitar Misev]

SECORE used to be separated into a

1. library that handled everything, and
2. a frontend that parsed http requests and forwarded to the library

I'm not sure if this is still the case now after the refactoring with spring and maven. But it would be great to use the library directly in petascope. In this case one could specify secore_urls=embedded in petascope.properties for example.

[Peter Baumann]

yes, exactly some solution along that line I had in mind. Hope it works out.

[Bang Pham Huu]

It is due to BaseX library for SECORE cannot handle requests in parallel. Duplicate with: ​https://rasdaman.org/ticket/2328