So as i understand, if server meshing all goes to plan, then you’ll be able to segregate servers based on areas in game, like a planet or a building.
1) Are server “ticks” linked to the clients inputs? (Like Valves games for instance)
2) Will the AI be improved with a higher tick rate?
3) Is it possible to have a higher tick rate server in places with a concentrated amount of FPS combat on a planets surface, or in a bunker for instance?
1) No, the servers run the simulation as fast as they can up to a maximum of 30 frames per second. The packet rate sent by a server is determined by its frame rate, and how many entities are updating and have new data to send to the clients. Neither the frame rate of a client or the rate at which it sends packets to its server have any influence on this.
2) AI logic is run entirely on the servers and the higher a server’s frame rate the faster the AI “think,” leading to quicker reaction times. Their ability to plan ahead isn’t really affected by frame rate though so faster frames won’t make them smarter. Some recent bugs with the AI have been due to desync issues, where what you see on the client no longer reflects what they are doing on the server. Good old fashioned bug fixing and continuing to flesh out the AI systems are where the improvements will come from.
3) Yes it would be possible but this is not something we’re currently considering. Running a server at a higher frame rate means it can’t update as many players and other entities. Since servers cost money to run we want them to simulate as many entities and players as possible. The final version of server meshing should allow us to balance the load between multiple servers so that we can maintain a consistent frame rate on all of them. Concentrated FPS combat in a bunker between a fairly large number of players may warrant that bunker being given its own server. The same number of players spread out over the surface of a moon may mean that the entire moon can be simulated at the same frame rate by just one server.