Ability to use your own PC as a SIM

imagine if everyones PC could be a SIM this way 2nd life would not need to rack mount hundreds of servers. Say for a reduced cost you can go in a section of second where everyones PC was part of a grid and you move from one persons PC to another in a peer to peer type fasion. Maybe charge $x per month to add your PC. You can then sell use of your PC to others for $y per month....

Four words: "Reverse engineering" and "user fees". LL won't do it. Wouldn't be financially prudent, and would almost certainly lead to their server being broken apart.

The feasibility of a peer-to-peer architecture for a platform like SL is extremely low. Maybe one day when everyone has a supercomputer in their closet connected to a fiber-optic line, but a 1.4Ghz PC on a DSL connection is not happening.

On the reverse engineering note though, the full protocol spec including sim to sim communication is already available. If one was determined enough they could write a server emulator (local sandbox) just like the one's available for World of Warcraft. The difficulty depends on the amount of complexity you want. Sending a couple packets to tell the client you're a server is easy but storing land information, prim info and coordinates, running an asset server as well, all of it adds up and to convince the official SL client it's talking to a "real" server would be a lot of work. I guess my point though was the information is already out there if one wanted to do this, releasing a client-side sim would only make the work very slightly less difficult.

If you want to know why there isn't one available yet from LL, think to yourself about return on investment from a corporate standpoint.

While it is appealing to think about think about the massive amounts of traffic a sim attracts per day. That would take not only an extremely dedicated connection to handle, but what if that sim is popular? What happens if this computer shut down itself? Our own computers cannot avoid the same problems the Linden servers have been going through with all these attacks.

I am not sure we need to worry about reverse engineering etc this is not required as it is a LL request not a rougue hack project. It costs the Lindens a lot of money to add a sim they need to add a Blade (I assume) or a CPU board. Ram network etc) and then they need to support it. In a peer to peer world you could only connect your PC as a SIM if you paid Linden. No pay no connect. So Linden would still generate profit just not have the costs. You could then sell your sim to someone else to use. The price you sell it for would be dependant on performance. ie supply and demand.

I think Peer to peer virtual world is the future (grid) and I could not imagine there is a lot of traffic between the SIM and the user as people use this service today over dialup. If the Lindens don't do this some smart open source group are going to create a grid VR project that will grow very fast cause it will be free..

I expect that long term LL and SL may well be overtaken by an open source project. But the solution for LL isn't to just turn their own business to open source and effectively throw away their livelihood.

I think this proposal is logical and is perhaps the way it should have been done from the beginning. Would surely cut down on some of the hassles LL deals with every day. They could start focusing exclusively on programming and a whole lot less on grid maintenance and crashes. It would also allow users to have a TOTALLY dedicated server (not four sims sharing the same hard drive/databus/network access). It would of course, likely require major and extensive re-writing of code, but I think it's arguable that at SL's current growth pattern, they're eventually going to get to the point they simply cannot support any more new sims. Then what?

If we own a local-server sim, it crashes, we're back up in 5 minutes. If the grid crashes, we're down for hours. Every time SL revamps software, it takes out an entire Wednesday morning, whereas if sims were client-local, they could download the software revision, install, up in 15 minutes.

So Nink's suggestion bears consideration. I'm sure LL has considered it extensively and negated it, but it does make a lot of sense. I mean, how many of us play solitaire by dialing up MicroSoft? No, we load it on our home computer and go with it. Client-side sims makes a lot of sense. LL could still charge a monthly license and maintenance fee, but they'd significantly drop their overhead costs and downtime.

Of course, on the downside, there would always be the issue if a problem happens of "Is it LL software or client system?" Answer to that would be the same as every other software company in the business: how widespread is the problem? If widespread, likely central server/software issues. If not, likely local.