How effective would an Internet server be for a complex game?

Yes. Physics based eye-candy can be rendered per machine without having to worry about it. Different machines would need to share positons and forces for the vehicles, and probably uniform variations to terrain. The data transport requirements don't need to be much above the norm.
 
london-boy said:
All i can see from Motorstorm - the equivalent we will get on the PS3, regardless if it looks as good as the CGI demo - is lots of particles and smoke... Car physics will be good but hardly something i'd want an offline server to do instead of the PS3 i have at home...

Shifty Geezer said:
Yes. Physics based eye-candy can be rendered per machine without having to worry about it. Different machines would need to share positons and forces for the vehicles, and probably uniform variations to terrain. The data transport requirements don't need to be much above the norm.

Ahhh ok. Thats good to hear then. I really need to brush up on the specific needs of a Game Server and whatnot. What type of game would really tax an Online Server anyways? (just MMO's?)
 
Anything where the world events can affect each player. In the case of mud, you can render these per player. But in the case of falling rocks, say, that impact players and vehicles, every machine needs to be in sync. Otherwise in player 1's game, player 3 gets knocked off the cliff by a boulder, while in player 3's game the boulder passed close to his left but missed...

In that same avalanche situation, you could process the boulder physics on the server, and for the little pebbles and dust, calculate those as special effects on each console because they don't affect the players.
 
Shifty Geezer said:
You'd be better off running an offline simulator and record the object's motions through a series of keyframes.

It is not a "full-on or nothing at all" proposition. Physics simulations can be very rudimentary (computationally fast and light) or very detailed (not particularly amenable to realtime applications) PLUS anywhere in between. You scale to the kind/degree of data you need to drive the event you intend to create, and no more (i.e., don't be gratuitous with physics simulations just to say you are doing physics simulation).

What you are describing is essentially "pre-baked physics", which is more useful than nothing at all, but can potentially fail in situations that involve an adaptive/dynamic component. If you can "afford" the computations, than why not use it? That way the event will be correct for the incidental conditions of the environment rather than just a single orchestrated environmental state.
 
Mr. Hanky said:
What you are describing is essentially "pre-baked physics", which is more useful than nothing at all, but can potentially fail in situations that involve an adaptive/dynamic component.
What wrote was only in response to randycat's idea. If you're going to use prebaked physics you would pre-calculate it, rather than have animators hand-design it. Not a system anyone would want to use though.
 
Shifty Geezer said:
In that same avalanche situation, you could process the boulder physics on the server, and for the little pebbles and dust, calculate those as special effects on each console because they don't affect the players.

I think this is the situation Im most curious about. Game World physics as opposed to Special FX physics is the distinction Havok PR has made for their SDK. Just hypothetically using the MotorStorm trailer again as an example, take the scene where the ATV and a Dune Buggy completely wipeout into that wooden cabin where it realistically breaks apart and affects the results of the crash. I suppose that could go under both categories. Do you think some form of server side physics will be required for network play as these physics engines advance and affect gameplay? Or will devs hit a brick wall and have to "dumb things down" (for lack of a better term) for online play?
 
In that s[ecific example, I think it'd be on a per console basis as the broken bits don't affect play. If they did, like hittin motorcyclists, it'd need to be synchronized. And I don't know how to achieve that (hence my opening this thread!). I don't know how realtime server-side calculations can be used for a high-speed arcade game given network bandwidth limitations and lag.
 
Shifty Geezer said:
In that s[ecific example, I think it'd be on a per console basis as the broken bits don't affect play. If they did, like hittin motorcyclists, it'd need to be synchronized. And I don't know how to achieve that (hence my opening this thread!). I don't know how realtime server-side calculations can be used for a high-speed arcade game given network bandwidth limitations and lag.

Gotchya. To use that example again, another problem could be with the cabin basically destroyed it potentially changes the terrain in a major way by leaving an open pathway through that specific area. Hypothetically another vehicle could be just a couple seconds behind the aformentioned terrain deformation, so I suppose latency in that situation could really throw things for a loop as well.
 
Back
Top