I am trying to build a client-server OpenGL rendering environment in an ethernet-connected LAN computer. So the idea is to build a game and have one server running the application and rendering the scenes (plural) and send it to multiple clients.

It is basically a research project and my argument is since the game is basically played in a single arena and since multiple players means simply multiple cameras, I can put the required data into the server's memory and ask the (hi-spec) server to render 4-5 different images (or FBOs) and send them to connected clients. But I have yet to understand the limitation and technology that I will use to tackle this issue. In short, I want to localize the cloud gaming architecture (such as Stadia, Game Pass, PS Plus, GeForce Now) to observe whether it can decrease the delay significantly. Whether this is a good idea or not is a different matter. Right now I just want to know the technology/framework/library that I can use to tackle this and my limitation.

Any info/recommendation is most welcome. Thanks in advance.