We tested AppStream CPRS at Valley Coastal Bend TX VAMC (one of the first VAMCs to migrate VistA to the AWS cloud).
One particular 'hotspot' of latency for clinicians was loading a complex multi-layered TIU note for annual checkups.
In their locally installed (Windows Desktop) CPRS it took over 29 seconds to load.
When they used the remotely installed (AWS cloud-hosted CPRS) and then streamed it to their desktop using AWS AppStream, it took 1.2 seconds.
This proved our hunch that it was indeed VA's networking infrastructure as the root cause for CPRS latency.
When we eliminated all the layers of routers, gateways,.etc) and connected CPRS directly to VISTA within the same AWS cloud,
all of this network latency disappeared.
What we actually "rediscovered" was the original archtecture of VISTA/CPRS, which were designed to be physically closely
connected to each other (i.e. both running within the same VA hospital). As VA moved the Vista servers away from the
medical centers, it introduced latency.