Do you plan to implement server switching so that netcode servers can be fault tolerant?
eg: in rare cases, the server may crash, so all clients must switch between servers and continue playing.
some sort of a secondary server that connects to the primary one and all clients at tick x and stays for a few other ticks until it makes sure all ghosts are in sync before changing its state to primary.
That’s something that only MMO games do, AFAIK, and even then only some (maybe EVE Online).
What other games do you know that can survive a server crash and migrate players over to another one?
Many authoritative multiplayer games specially those using Google Cloud or AWS are using spot instances which costs 10 times less than on demande ones. But they have to be fault tolerant so they can switch server at any time with delay of few minutes.
Are you making assumptions or talking from experience?
Such games don’t use long running processes. A server instance only exists during a match/session and the process is killed afterwards. The matchmaking/server browser itself is a separate service, not a running “server”. If a game needs persistence it will write out state to a database that can be loaded when spinning up a new server instance.
Can you please re-read the thread description ?
I am not talking about MMOs and long running servers.
Thanks