We enjoyed success with our POC. Ne we need to know if the solution scales. Thus, we are planning the final deployment steps.
Background
Our Sun Ray setup used 1024 x 768 resolution and required nothing fancy to serve all the desktops throughout both offices and across the site-to-site VPN tunnel.
With FHD (1920 x 1080) resolution, that generates more pixels to manage per user session. Our HP thin clients with ThinPro OS plus the monitors both handle this resolution.
Questions
1.Is there anything special I need to do with the Linux desktop the TL server provides to pass-through this resolution from the Windows RDP session launched in the Linux desktop back to the monitor, in its native resolution?
2. Thinking with how TL handles video in a session, is there a difference in resource utilization when launching a RDP client session via "tl-single-app"configured via the TL server profile page versus launching an RDP client from within a served xfce desktop session, for example?
3. Do you have tips around extra resources or configuration settings the TL server might need to handle the increased pixels for the 30 concurrent sessions?
4. Do you have suggestions on how much LAN bandwidth I might estimate per TL session for this type of setup?
No, aside from running the ThinLinc client in full screen, you should not need any special setup. The RDP client in the session will see the virtual screen as a regular screen with a resolution matching the window size of the ThinLinc client window.
Thinking with how TL handles video in a session, is there a difference in resource utilization when launching a RDP client session via "tl-single-app"configured via the TL server profile page versus launching an RDP client from within a served xfce desktop session, for example?
Probably not much difference, but in theory there could be. tl-single-app will only launch openbox alongside the application, which means there are fewer processes running per user.
Do you have tips around extra resources or configuration settings the TL server might need to handle the increased pixels for the 30 concurrent sessions?
Not much. If you decide to run a full desktop environment on the ThinLinc server instead of tl-single-app I would recommend disabling the “compositing” setting.
The biggest impact on perceived performance is likely to be found in tweaks to the RDP connection, I don’t expect ThinLinc to add much overhead here.
Do you have suggestions on how much LAN bandwidth I might estimate per TL session for this type of setup?
It mostly depends on which RDP client you use and what the users do in their RDP sessions. If the RDP client correctly only draws the parts of the screen that change, ThinLinc will be able to only send those pixels. This means that there is a considerable difference in required bandwidth between users watching full-screen videos compared to users writing in a text editor.
It is therefore difficult to give you such estimates. Perhaps you can do some measurements in the POC environment with expected usage patterns?