I lowered the update rate to 2 for positions/rotations/scale(38 bytes) and am also sending 5(14 bytes) velocity updates a second and it feels pretty good...I'm also gonna need some updats for the two fire buttons. My position correction code needs some help though. right now it just snaps the position back if it's off by too much. It's especially noticable when the player makes sharp turns.

Even with these low update rates I'll only be able to support around 200 players on a 10Megabyte connection.

Can you recommend some good cloud hosting services?

Can you recommend some good cloud hosting services?

Well, the player is in full control of their character movement... only a couple checks to see if they're trying to illegally teleport or trying to fly. So it feels good as far as movement goes, but some players might get annoying tryig to shoot guys that suddenly snap a meter away if they are moving erraticaly.

I have no movement prediction yet... players don't need it for themselves but I could use a little better system for handling other players.

The snapping of spazzy players is a little annoying but now sure how to fix it.

The snapping of spazzy players is a little annoying but now sure how to fix it.

I've started sending the yaw rotation of the player 5 times a second and the snapping has pretty much vanished.

I think I'll end up scaling the update rates with the amount of players in the area with my current settings being the lowest it goes. Although I'm not sure I'll have to, the movement is smooth as silk on the remote systems. First time I've had good movement on games I've tried making... It's been the main thing keeping me from progressing a multiplayer game. Hopefully I can finish this one now.

Distributed server

For the cheating isse - who says a player who is the host has to be running the simulation for the zone they are in ?

If there are sufficient resources (for world representation needed for the server simulations) beside the players client state (a different world representation) the hosting one player is doing need not be the same as the one THEY are in and thus avouid most ofthe reasons players would cheat.

(you still have 14-year old mentality mind-sick griefers who want to screw up the game for anybody else and you would have to handle that case reguardless)

You could still try to archive something like (early) battle.net and Guild Wars 1 (as far as I know). The concept is, that you have a global lobby server (GW1 these are the cities) and the gamers team up(4-8) to play mission/quests in shared instances. The trick is, that the instances runs only on the clients. As already mentioned, you get in trouble with P2P and cheating, but the bandwidth/power requirements are really low compared to a real MMORPG server architecture and are still quite successful when looking at GW1.

You could still try to archive something like (early) battle.net and Guild Wars 1 (as far as I know). The concept is, that you have a global lobby server (GW1 these are the cities) and the gamers team up(4-8) to play mission/quests in shared instances. The trick is, that the instances runs only on the clients. As already mentioned, you get in trouble with P2P and cheating, but the bandwidth/power requirements are really low compared to a real MMORPG server architecture and are still quite successful when looking at GW1.

Even if you use global servers for the game instances it would be a lot cheaper than a proper MMO architecture since the instancing to that level gives you linear scaling. (Double the number of servers and you can support double the player count)

For the cheating isse - who says a player who is the host has to be running the simulation for the zone they are in ?