[Opensim-dev] oddities with asset storage

[Frisby, Adam]

I would argue that the majority of hosts are transient and not permanent. Hence in casual operation some assets will vanish from time to time. The only way to do this is by having a few rock solid core nodes who have everything between them.

However honestly, I would say that region assets are the responsibility of the region host, and user inventory assets are responsibilities of the user host.

Adam

> -----Original Message-----
> From: opensim-dev-bounces at lists.berlios.de [mailto:opensim-dev-
> bounces at lists.berlios.de] On Behalf Of Eugen Leitl
> Sent: Friday, 20 February 2009 2:42 AM
> To: opensim-dev at lists.berlios.de
> Subject: Re: [Opensim-dev] oddities with asset storage
>
> On Fri, Feb 20, 2009 at 05:32:00AM -0500, Frisby, Adam wrote:
> > It's for use in a trusted environment only.
>
> How much can you trust 10000 random strangers
> who can run OpenSim servers in a hypergrid and do?
>
> > You can't have an asset cluster in an untrusted environment
> > without replicating every asset across every host -- otherwise
>
> I recommend you look at what Tahoe does.
>
> > someone could setup, run for a while, then deliberately corrupt
>
> I really recommend you look at what Tahoe does.
>
> > the hard disk. Given enough machines doing this, you could
> permanently wipe assets.
>
> It is not realistic to assume that most hosts are malicious.
>
> Notice that this is the time where bad design decisions could
> wind up cemented for good.
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev