I've explained the BZ2 networking model countless times before on previous iterations of the BZ2 forums. Rather than dig all those old posts up, I'll point to a 2-part (so far) set of articles on how other games do similar things:
Part 1
Part 2
Mainly FYI, and a good set of reading on things from a programmer's point of view. Like SupCom in the first article, BZ2 has a sim layer and a user layer that handle smoothing out the local user's viewpoints, etc.
-- GSH
Battlezone II networking model - not unique
Moderators: GSH, VSMIT, Red Devil, Commando
Re: Battlezone II networking model - not unique
And you are only one man working on the beast from hell.
Re: Battlezone II networking model - not unique
Before reading the article I can probably make a few points.
What makes it workable:
Deterministic AI processes and physics and logic.
What makes it hard:
Non-deterministic factors that can't be made deterministic, like humans.
What happens when it de-syncs:
You send the ENTIRE GAME STATE from whomever is the master to all the children.
How to mitigate the human factor:
Send over as much data about the humans as possible. It has to get to every client and the server. Send partial game states that are basically 'player states' for the purpose of fixing the very likely and common de-syncs (warp).
Am I right GSH? I will read that article tomorrow.
What makes it workable:
Deterministic AI processes and physics and logic.
What makes it hard:
Non-deterministic factors that can't be made deterministic, like humans.
What happens when it de-syncs:
You send the ENTIRE GAME STATE from whomever is the master to all the children.
How to mitigate the human factor:
Send over as much data about the humans as possible. It has to get to every client and the server. Send partial game states that are basically 'player states' for the purpose of fixing the very likely and common de-syncs (warp).
Am I right GSH? I will read that article tomorrow.