Closed csmoore closed 6 years ago
@csmoore Can you provide some more information. What can be done about this issue? What is the proposed solution? What are the limitations, if any?
I think we need to only request the tiles within the bounding box of the input features. I don't know, but I suspect that a bigger area is being pulled down for the analysis.
It would likely involve first extracting the server data locally (Note: this option must be enabled on the server data for this to work): http://desktop.arcgis.com/en/arcmap/10.3/tools/server-toolbox/extract-data.htm
This would still be a client-side solution with a several minute lag before processing would start (depending on clip size) - so we might also want to set a smaller maximum range also.
This is like the case outlined here: https://github.com/Esri/visibility-addin-dotnet/issues/194 Still think this is important.
@lfunkhouser @dfoll @ACueva I feel like we should close this and show all the work in the issue that came in from support #194.
I think we should calculate based on the extent based on the information put in by the user. The challenge is when a user selects two areas that are not close to each other, but are still small request areas.
LLOS and RLOS are slow using image services. Even with reasonably sized image services (such as county-wide data).
Example run times (depending on service/AOI size): LLOS takes 1-20mins RLOS take 20mins-2 hours
There is a warning that pops-up before you run (when you press OK to execute)
Repro steps:
Originally noted in #184 but I guess we never opened a new issue specifically for the slowness.
Tested with data from: Rhode Island: https://maps.edc.uri.edu/arcgis/rest/services/Atlas_elevation/Mosaic_RITopoBathy/ImageServer Maryland county data: http://lidar.geodata.md.gov/imap/services/Allegany/MD_allegany_dem_m/ImageServer