koobas.hobune.stream
MMO games in the Web Player: lots of players and game performance.
I was wondering how many human players is possible to bring togheter in a MMO genre game, running within the Unity Web Player, before the game starts to drop under 30 FPS.
My idea was some kind of simple game (absolutely not an RPG) where user only have to get the highest number of points while playing togheter (via online multiplayer) in the same scenery, track, or arena. For instance, think about a race like the Paris-Dakar: thousands of cars (the users) races one against the other for the podium.
In order to get the best performance, the game will be designed to run at 30 FPS, using low-poly models, small number of bones, a very small number of variables to handle, without heavy effects for the CPU/GPU, etc.
For whose of you who have some experience with Unity Web Player and online gaming, how many players is possible to get thogheter, approximately?
Your issue most likely won't be with graphics (though old computers may have issues with that many separate meshes if you're not careful), it'll be with the networking side
Every player will have to be sending every other player their current position if they're close together (be it via the server which sends the data to other players, or via peer to peer, which is basically impossible for thousands of players). The amount of data required is actually pretty staggering - you'd be looking at a 1 gigabit upload speed for the server, or a couple megabit or so for peer to peer.
On top of that, the cpu usage for handling it is pretty severe - you'll need to synchronize moving the player positions with unity, which throws threading out of the window for added "fun"
Both of those are unreasonable, so it's more a case of how well you can code it. You'll need to make sure you're barely sending packets to cars outside the nearest 10 or so, and then add heavy interpolation/extrapolation for the rest.
I'm not saying it isn't possible (It most likely is, though you'll have scary issues at the starting line) , but it'll take a hell of a lot of work to make it work, and a pretty hefty server on a backbone, which will be rather expensive.
Thank you very much for the answer, very explanatory. I really understimated this networking problem! At this point, I would say taht $$anonymous$$$$anonymous$$O games are really difficult to develop, so the best thing would be a simplier Onlne $$anonymous$$ultiplayer mode, with a dozen of players... I will think about it. Fortunately this $$anonymous$$$$anonymous$$O sistem was only an insane idea in my $$anonymous$$d and nothing more concrete. Thanks again!
Holy return to sanity Batman! But yeah $$anonymous$$r. Drayton, as I commented on a similar past question, if anybody needs to come to Answers with a $$anonymous$$$$anonymous$$O question, they likely don't have the resources to pull it off. It is much better to focus on something that you can actually complete.
