[ENet-discuss] Questioning a server without connecting to it
Stefan Lundmark
stefanlun at hotmail.com
Sun May 27 01:25:38 PDT 2012
I'm sorry but that's not how it works. There are very good papers out
there (from the makers of the Unreal Networking Engine, Source and TNL)
where you can get a good idea of different design choices that you have
to make.
In any case, you don't have to make your choice now. You can change your
tickrate at any time, if you're not happy about it.
Your 20 * 32 figure is incorrect, of course. You assume each client only
is within scope of itself. The real figure is much larger.
But I think you made up your mind even before you came asking, so it
doesn't matter.
On 2012-05-27 06:47, manu.n02 wrote:
> An additional thought :
>
> a classic FPS server runs at 20 frames per second ; with 32 clients,
> that gives 20 * 32 = 640 packets / second.
> So I doubt that my 60 packets/s are any close of an "insane network rate".
>
> And the real limit of a FPS server is the bandwidth... More info can
> be found here : http://www.cstrike-planet.com/tutorial/1 .
>
> > Message du 26/05/12 06:13
> > De : "Lee Salzman"
> > A : "Discussion of the ENet library"
> > Copie à :
> > Objet : Re: [ENet-discuss] Questioning a server without
> connecting to it
> >
> > Which is why 60Hz is basically an insane network rate. You're
> still drowning in headers no matter how small you make things.
> Really, you don't need 60Hz. For objects with predictable
> trajectories you don't really need more than 20-30Hz. The changes
> in trajectory don't need to be sent at any particularly rate, and
> thus are sent out immediately, there doesn't have to be a delay
> involved.
>
>
>
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at cubik.org
> http://lists.cubik.org/mailman/listinfo/enet-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cubik.org/pipermail/enet-discuss/attachments/20120527/c5cf6985/attachment.html>
More information about the ENet-discuss
mailing list