When I try to draw a relatively large data set (e.g. 36,000 points) in Portal for ArcGIS at a small scale (e.g. all of North America at about 1:40,000,000 scale), it's so slow it's just not usable. However, I can draw 10-million+ points in a 90,000 km² area (at about 1:1,700,000 scale) in Portal without any issues at all. Why is that? Why does scale seem to play such a big role in GIS drawing performance? I can also draw 200 points of locations all across North America without any performance issues at all...so quantity of points is (obviously) the main part of the performance issue. But why can I draw larger data sets at large scales much more efficiently than moderately smaller data sets at small scales?
What's happening on the server that slows this performance? On my desktop (ArcGIS Pro - 2.4.3) the same data stored locally draws very quickly too. It's only when it's running on our server from our federated Server/Portal servers at small scales does large data sets seem to be an issue.
Normally I would just cache the data and be done with it...but the user wants to be able to make small edits to the data in future discussions and meetings and of course editing cached data is a problem (I'm surmising that this is probably because one cannot cache a feature layer, only a map image layer?).
What do other people do when they want to display large data sets at small scales using Portal for ArcGIS?