[ENet-discuss] Fair peer-to-peer networking for gaming?

Adam D. Moss adam at gimp.org
Thu Feb 23 10:01:28 PST 2006


Hi all!

I've been thinking about the requirements of net-gaming
a lot lately, and I was wondering if there's any current
good-thinking in the area of peer-to-peer play amongst
several/many players.

There's a model where one of these peers acts as server
and then it's just a regular client-server model where (as
I understand it, correct me if I'm wrong) the peer which
is nominated as server is the authoritative source of
game-world state changes while the clients work on
extrapolated data / dead-reckoning until authoritative
decisions come in regarding requested state changes etc.

This seems like it would put a disproportionately high amount
of bandwidth and CPU load on the host acting as server, as
well as an unfair response advantage.  So, I was wondering
if there were methods in general use which spread the
authority for world state updates in a fair and fine-grained
manner.  It seems like it could be a hard problem where these
entities have to react with each other, e.g. dividing the
authoritative ownership of potentially-interacting physics objects
between peers, but I can hope.  Any links or such would be
appreciated.

Cheers,
--Adam



More information about the ENet-discuss mailing list