I have a Story upon a question that is bothering me some time now.
I use NR as Mediacontrol for various devices in a number of Projects.
Most devices are common PC's which I start, stop and observe to create Multimedia Content.
The most of that is done, with default nodes Pallet. Some are from the community like SSH or Pjlink or simple pinging.
All those Protocols that rely on a TCP connection underneath are causing a slow in/out reaction time somehow.
If I ping more than 40 devices, I get a delay on Telnet requests that I use for volume control in a latency I can feel when stepping down, volume takes from one step to the next 2sec. But as soon as I deactivate the pinging, I can control my audio volume seamless.
Same in another Installation where I have loads of KNX actors that are switching regularly upon PIR sensors while I measure my flow takes up to 0.1s to roundtrip inside NR the actor actually switches after 2s.
I would think this is because the KNX-Gateway has that limitations, if I wouldn't know that I have changed the system from openHAB to NR, and it worked instantly on openHAB on the same Host.
This leads me to a more broad questioning about what on the side of TCP dispatching could be blocking my usecases.
