[ENet-discuss] ENet package broadcasting
Espen Overaae
minthos at gmail.com
Wed Feb 25 12:29:00 PST 2009
On Wed, Feb 25, 2009 at 8:41 AM, Boris <bstih at zootfly.com> wrote:
>
> Hellow again, thank you for such quick reply...
>
> Please elaborate your first statement.
>
>> / Because of our fixed time step, we wait every client to send its input
>> to
>
> />/ every one. When data arrives, we update next frame.
> /
> that is a bad idea,
>
> Why do you think this is a bad idea, do you have any other idea. I know that
> our idea is not cheat safe. But we see only this, as a major problem in our
> idea. And ofcoarse our code is for LAN only. Internet is another story.
For LAN it's ok, it makes things simpler and you usually have very low
ping values on a LAN, but it's a bad idea because you can't use it on
the internet. Also, if your LAN is either wireless or crowded, packet
loss could occur, causing delays.
You mentioned Halo, so I assume this is an FPS game or at least a fast
action game (as opposed to turn-based strategy game). The alternative
I suggest is to use a fixed timestep on the server and the clients,
but not demand everything be fully synchronized before incrementing
the time. This introduces a lot of new challenges in regards to
consistency. The site Erik Beran links to in another email talks about
this, and there are a lot of other resources about this on the web.
It's not my field, but I believe John Carmack has written some good
stuff about this, he wrote most of the engines for the Doom and Quake
games, and pioneered multiplayer FPS games.
Espen Overaae
More information about the ENet-discuss
mailing list