Hi all!

I have tried various virtual world software (moove 3d, ActiveWorlds, SecondLife, realXtend) and decided that OpenSim based ones are the most interesting to me.

But I am still curios - would it be possible to integrate some minimal server and peer-to-peer part into a viewer? I'll try to explain why.

Let's say - I want to have my own private piece of world where I can do everything.I do not want to buy any regions on the server nor upload anything. For an average user it wolud be a bit complicated to join his/her PC to the grid (even if it would be allowed to do) and it would not be good idea anyway because this PC is not always running the server software. But if the viewer has capabilities to join the grid sometimes then the user can have his world. The central server just keeps the simplified world mesh and displays everything blank when the user is offline. When the user gets online, the central server detects it. At this time everyone who wants to visit this private world will connect to the world owner directly (peer-to-peer) and download the textures and objects directly from the owner.

Is it possible with the current OpenSim based servers/viewers?

Indeed loginGrid GAS may help although I think in our own project track I could point out a little more.

One of the things that LL does well that we are yet to attack in the project is the delivery of the scene. I'm not so much talking about how well they send it to you or congestion but the method.

Giving you what you need to see when you're facing that way, texture layering, u touch and object and it gives it greater priority etc.

From my understanding so far, the streaming delivery of the scene on a whole is more efficient that if all textures for examples were urls pulled in from different places. Getting intelligent priority based on user activity is probably another one.Then I guess there's a rather large swift of other factors to consider, IP, security, implementation of DRM?

Internally we've talked a bit about peer-to-peer but have yet been able to come up with 'enough' positives to decide to implement. But discussion is a good thing :)

I'd write more but I'm a little pressed on time atm.

Sakai Openlife