Information Week has an article up about new improvements coming to Second Life.
In the article, claims are made about upcoming improvements to the Second Life platform. Obviously, the upcoming voice support is mentioned. There also appears to be instant messaging between Second Life and the web on the horizon (this already exists through a third party), photo sharing (screenshots), third party access to Second Life registration data for service based websites, and prettier skies in-world. The latter being brought to the Second Life universe via a recent acquisition of Windward Mark by Linden Lab. Windward Mark specializes in atmosphere and environmental effects in virtual worlds. Here’s a screenshot gallery of some of their work. Linden Lab released a first look client that incorporates the Windward technologies. And who can forget that sculpted prims were recently added to the list of improvements to the world.
All of this is really great, but there’s a certain 2,000 pound gorilla sitting in the middle of the virtual world that is begging for attention, grid resident concurrency. There is mention in this article of new upgrades and a next generation architecture that will increase the concurrent user number to 400,000 (this is a much lower number than what was estimated in an earlier Information Week article of “tens of millions of simultaneous logins“). That’s a huge increase from the current roughly 40,000 concurrent user cap (and virtual life gets really annoying as the user counts approach this number, usually on the weekends).
I truly hope that Linden Lab is able to make this transition quickly, as remaining at the current 40,000 (actually less than) user concurrency cap any longer is going to damage the attractiveness of the Second Life platform for business, education and the general residental population.
There was a mini survey posted to the Second Life Educators listserv this weekend asking what one thing people thought needed to be “improved” about Second Life. I replied with the following to this inquiry:
“Distribute the servers, bust the grids into pieces. Linden Lab would provide the links between grids and a central database for objects and L$. It’s the only logical next step in the evolution of the platform. A contiguous land mass holds no value when everyone teleports everywhere anyway. The only purpose a contiguous land mass serves is to give a size of the virtual world that can be compared to real life cities. “Twice the size of San Francisco” holds no value if you can’t fit the population of Cupertino (over 50k) on the main grid concurrently. A concurrent user count of 1.5 million would be roughly “twice the population of San Francisco”.
Linden Lab cannot survive with a 40,000 concurrent user ceiling on the grids. Blocking free accounts from logging in, and other “band aid” solutions to ease pressure on the grids in peak times will not allow Second Life to become the “next big thing”. Plateauing at 40k users for any significant length of time will bury Second Life as a “good idea” that was poorly executed.
It remains to be seen if this problem will be solved, or if it’s a hard coded limitation of the infrastructure that is not fixable without a full rewrite.
All other issues are moot in comparison to this single improvement that must be made in order for Second Life to survive.”
June 10, 2007 @ 10:00pm PST: Someone from the listserv has compiled the responses to the survey and have posted them here.
While the response may seem a little harsh, I think it’s necessary that the user community lets Linden Lab know that they are unhappy about the grid issues that have plagued Second Life for the past six months. The “band-aides” starting being applied back in February to ease the grid congestion issues. All too often, discussions about Second Life and Linden Lab only focus on the positive, many are hesitant to criticize Linden Lab, or to make their feelings known in a public forum. If they are going to improve the platform, and move to the next level, the residents need to take an active role in helping to shape the future of the platform and that can only happen if we give constructive criticism of this platform.
My method of addressing the grid issues (by distributing the grids) is only one possible solution. It would appear from another Information Week Article on “next generation architecture” that Linden Lab is attempting to throw more processing power at the problem instead of following the Internet model of distributed servers. I was somewhat shocked in reading this article that their database is based on MySQL?!?. Here’s the quote from Joe Miller, VP of platform and technology development for Linden Lab:
“MySQL allows the server farms to scale horizontally, by adding large numbers of low-power servers as needed, rather than vertically, which would have required Second Life to run on a few, powerful systems, Miller says.”
I’m sure other people in IT will share my surprise at that revelation. Linden Lab is also looking at reworking the Linden Scripting Language (LSL) using a version of Microsoft.Net. This seems like a very positive move on their part, as there is a significant increase in performance over the current LSL.
These grid issues have already gone on too long, if significant improvements are not made in the very near future the viability of the entire platform will come into question. I personally hope that these latest moves fix the problems and that things can move on to the next level. We are definitely at a critical juncture in the history of this platform.