<html>
<head>
<style>
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
</style>
</head>
<body class='hmmessage'>
Excellent work, John! How far from implemented is this?<BR>
<BR>
Two immediate points;<BR>
a) This seems to shift the model for throttling from push to pull - ie, the data servers are no longer in a position to throttle bandwith or cut texture transfers depending on client needs[1]. What is your stance on that?<BR>
b) I guess the core of my suggestion is what you call the "OpenSimFrontend" - is that documented anywhere?<BR>
<BR>Best regards,<BR>Stefan Andersson<BR>Tribal Media AB<BR> <BR>
[1] In the current region-centric asset distibution model, we have a one-to-many diagram where the region fetches one copy of the asset and transfers it to all the clients, ie 1*n connections and transfers for the asset server, where n is the number of assets. In a 'client-centric' senario, we will have a many-to-many diagram where each client connects for each asset, ie m*n connections and transfers, where m is the number of clients. This shifts the asset connections and total asset bandwith use from the region to the asset server(s).<BR>
<BR>
> From: john.hurliman@intel.com<BR>> To: opensim-dev@lists.berlios.de<BR>> Date: Thu, 4 Dec 2008 15:52:06 -0700<BR>> Subject: Re: [Opensim-dev] Decoupling Was: [Opensim-users] Blender Exporter for OpenSim<BR>> <BR>> Take a look at the flow diagrams for the distributed asset server at http://opensimulator.org/wiki/AssetServerProposal/ClientDocs<BR>> <BR>> The only thing left out of that diagram is that the distributed asset server is also compatible with the OpenSim protocol today by running with the OpensimFrontend extension, acting as a replacement for the current asset server. But instead of a single monolithic service, you can run as many as you want (in your diagram the region asset cache could be an asset server running local to the region) and the data can be sourced from anywhere (such as the AmazonS3+CloudFront extension in SVN).<BR>> <BR>> John<BR>> <BR>> > -----Original Message-----<BR>> > From: opensim-dev-bounces@lists.berlios.de [mailto:opensim-dev- <BR>> > bounces@lists.berlios.de] On Behalf Of Stefan Andersson<BR>> > Sent: Thursday, December 04, 2008 12:57 PM<BR>> > To: opensim-dev@lists.berlios.de<BR>> > Subject: [Opensim-dev] Decoupling Was: [Opensim-users] Blender <BR>> > Exporter for OpenSim<BR>> > <BR>> > Ok, so this has been rotting in my draft folder for so long, I'm just <BR>> > going to send it. Consider it just kicking mental box-walls.<BR>> > <BR>> > One of the biggest (but not necessarily hardest) changes to opensim <BR>> > that would really get some balls rolling, would be to change how we <BR>> > think of assets from server-centric to region-centric. And, with that, <BR>> > introduce urls as the 'real' assetId.<BR>> > <BR>> > The assetId of today is, at its core, a thing between the region and <BR>> > its connected clients.<BR>> > <BR>> > Today, the region already fetches the assets with an http call.<BR>> > <BR>> > So, if we had a model<BR>> > <BR>> > client <-- assetId --> Region Asset Cache <-- assetUrl --> binary <BR>> > asset http resource<BR>> > <BR>> > we could still use the 'old' assetService by just fabricating a <BR>> > standard url, just like we do today<BR>> > <BR>> > BUT<BR>> > <BR>> > we would also be able to refer to asset resources by url... web style.<BR>> > In practice, killing off the need of the monolithic asset server as it <BR>> > stands today (being all synchronous and stuff)<BR>> > <BR>> > this would not be any problem with libomv, as the protocol would still <BR>> > be the same; the difference would be that some entity, quite possibly <BR>> > even the clientview/clientmanager would have to keep track of what <BR>> > asset url is connected to what assetId.<BR>> > <BR>> > It's SUCH an low hanging fruit, it would be really cool if somebody <BR>> > would just play around with that asset/url substitution and see where <BR>> > the actual problems show up.<BR>> > <BR>> > Best regards,<BR>> > Stefan Andersson<BR>> > Tribal Media AB<BR>> > <BR>> > Join the 3d web revolution : http://tribalnet.se/<BR>> > <BR>> > <BR>> > <BR>> > <BR>> > <BR>> > <BR>> > ________________________________<BR>> > <BR>> > <BR>> > <BR>> >> Date: Mon, 22 Sep 2008 11:42:36 +0530<BR>> >> From: shreekumar@hp.com<BR>> >> To: opensim-users@lists.berlios.de<BR>> >> Subject: Re: [Opensim-users] Blender Exporter for OpenSim<BR>> >> <BR>> >> Diva Canto wrote:<BR>> >>> If possible, it would be better if the objects would be exported to <BR>> >>> an external representation, instead of the MySql DB. Ideally, the <BR>> >>> external representation would refer to the textures by URL <BR>> >>> (including the local file system), and the right thing would happen <BR>> >>> when OpenSim would parse that. So no cryptic info.<BR>> >>> <BR>> >>> <BR>> >> +1. That's a downside of my exporter : it writes directly to the<BR>> >> database. Not pretty.<BR>> >> Issue is that I need to restart my opensim instance every time I <BR>> >> export the scene.<BR>> >> <BR>> >> The "load-xml2" command doesn't seem to fit these requirements too.<BR>> >> <BR>> >> -- Shree<BR>> >> _______________________________________________<BR>> >> Opensim-users mailing list<BR>> >> Opensim-users@lists.berlios.de<BR>> >> https://lists.berlios.de/mailman/listinfo/opensim-users<BR>> > <BR>> ><BR>> _______________________________________________<BR>> Opensim-dev mailing list<BR>> Opensim-dev@lists.berlios.de<BR>> https://lists.berlios.de/mailman/listinfo/opensim-dev<BR><BR></body>
</html>