Outback online's P2P gaming protocol

Hi, just some news i thought would be of interest to you:

http://www.3pointd.com/20070610/outback-onlines-peer-to-peer-gaming-protocol/

"His team broke the problem down into three parts. First, they tackled
the challenge of indexing users in space, designing a spatial index
that allows the various clients to discover users in the 3D space
around them, without having to have all that presence information
contained on a single server."

We can find similarities in the solipsis project, about the
proximity-based networking:
http://solipsis.netofpeers.net/wiki2/index.php/Peer-to-Peer_Algorithms

"Secondly, they tackled the problem of interacting, using
multithreading techniques (among other things that got lost in a poor
connection) to optimize communication between clients."

"Third came a security solution that obfuscates users' IP addresses
while still allowing clients to transmit the necessary information
across the network."

Interesting: i recall some discussions about anonymity here.

"That system of three modules is optimized for MMO applications, and
then sitting atop that is an API that is designed to work with
standard game engines."

This is a novel approach: they position themselves as unified
networking stack only; the intermetaverse market is expanding. Outback
Online will use this product.

Regards

Florent

 Florent THIERY wrote amongst other interesting things :
"Interesting: i recall some discussions about anonymity here."

There is a little feature I recall in one of the tutorial movies that
might have made folks less scared to play with me,
one way portals.

I go to this meeting space through a one way portal and neither I
nor other members of the meeting see the way back into my virtual world
or computer.

Before the tutorial movie,
I hadn't consciously noticed that one way portal
as I leaped through mirrors in my wifi home
set up.

That massive multiplayer online gaming requires more than Oreilly's
introductory tutorial on threads intrigues me.
I believe I read, in my library, that multithreading was advanced stuff.
This too is driven by gaming.
In particular, I believe you have to send signals between threads
to keep feedback "rational" and not driving you crazy in VR.
That's sophisticated stuff, if I recall right.

I fooled with such readings doing stuff with NSTimer in os x derivative
of the Next Step os driving shutterglasses and Maya PLE, so
seeing folks do such stuff is like meeting old friends.

I'm sure my stuff was much simpler than their stuff,
but, like Newton, I felt like I was dipping my feet in a vaste ocean
and didn't feel bad about that.

Thanks for the link.  Unfortunately I couldn't find any papers that  
describe their approach in more detail.

I think that the modular approach that they took makes sense, and  
will work for Croquet too.  For example, Croquet has not yet taken a  
distributed approach to user discovery, but such an approach is  
perfectly compatible with Croquet.

I'm not sure how a P2P network can obfuscate IP addresses, unless  
it's a two-layer setup where only superpeers know the real IP addresses.

Josh

On Jun 12, 2007, at 6:16 AM, Florent THIERY wrote:

> Thanks for the link.  Unfortunately I couldn't find any papers that
> describe their approach in more detail.

Nope, the website is opaque... :(
But their product have to be free to use, otherwise they won't be
"the" overlay network...

> I think that the modular approach that they took makes sense, and
> will work for Croquet too.  For example, Croquet has not yet taken a
> distributed approach to user discovery, but such an approach is
> perfectly compatible with Croquet.

Do you think that croquet routers, as well as users, could be
referenced using this network ?

Florent

The reBang blog has mentioned Outback in relation to Croquet a couple of times. The most recent is Layers of Peer-to-Peer Worlds which links back to another post in Feb. I haven't gotten to following the trail of links(the one going thru Robert Scoble has a good number of pingbacks) so I don't know if there's any details to be had but it might be a start.

Laurence

On Jun 12, 2007, at 12:33 PM, Florent THIERY wrote:

>
>> I think that the modular approach that they took makes sense, and
>> will work for Croquet too.  For example, Croquet has not yet taken a
>> distributed approach to user discovery, but such an approach is
>> perfectly compatible with Croquet.
>
> Do you think that croquet routers, as well as users, could be
> referenced using this network ?

Yes, definitely.

Josh

>
> Florent