In January, I was stunned by the news that LL was opening up the source code to the SL client, far earlier than I had predicted. At that time, it looked like they’d keep the server software under wraps for awhile though…which only made sense. I mean, how else could LL make money?
Fast-forward to today. Jaw, meet floor. Again.
In a report from the VW07 conference, Joe Miller of LL states that they will “be open-sourcing the back end so sims can run anywhere on any machine whether trusted by us or not.” He further adds, “SL cannot truly succeed as long as one company controls the Grid.”
Now, this begs the question: What exactly do they mean by this? Will these newly-independent sim servers be part of their own Grids, or will they all coordinate with the master asset servers and master “Grid map” at LL HQ? If the latter, there’s a business model for LL, similar to that of a domain registrar: for some annual fee, a sim server can enjoy connectivity to the world and the greater economy of Second Life. Or, put it another way, it would work like private islands do now, except LL wouldn’t need to supply the additional servers for these new sims; it would just “point to” the sim in its master geography database, and hand avatar connections off to it as they TP there. Yet private grids are still an option, assuming there’s a way for someone else to build the necessary infrastructure around these sim servers (asset server, login server, etc.)
ZDnet Social Web speculates that companies like IBM and Google will be interested in setting up shop with their own sims. Might we have an “IBMland” and “Googleland” continent to complement (or eventually supplant) the four LL Mainland continents, and sit alongside Dreamland and the rest? Will there still be room for the little guy in all this? (For example, could I create a single-sim “EMinds Island” running as a process on EMinds’ present server?)
More to the point: How will this affect the stability of the world we have now? Heaven knows we’ve been seeing a lot of instability recently, much of it asset-server-related (TP issues, inventory problems, L$ balance problems, etc.); will the strain of shooting assets all over the world prove to be too much for the asset server, even clustered? Can anything be done with the asset server architecture to scale this better? (There’s some indication from the original 3pointD report that LL is working on this.)
This isn’t as dramatic an announcement as the client opening was; when the announcement was made, you could download the client source practically right away, if you wanted to. Here, there’s no fixed timeframe for when we’ll see the server code. But we’ve just moved that much closer to the vision of the distributed Metaverse that LL has been chasing after all along.
“You’ve taken your first step into a larger world.” – Obi-Wan Kenobi, Star Wars: A New Hope