Should Linden Lab switch to OpenSim?

Today, Linden Lab is three companies in one.

First of all — and this is where most their revenues come from — it is a community company and content distribution channel. They charge people for access to this community through land rentals, advertising fees, commissions on currency exchange transactions and, of course, currency sales. In this, Linden Lab is most similar to Facebook or AOL. Like AOL, Second Life offers instant messaging, groups, social activities. Instead of an “AOL keyword” or prime placement on content pages, content providers and retailers rent land or buy advertising listings.

Second, Linden Lab is a software company, selling virtual world servers. In this, they are similar to Microsoft, with their Windows Server software.

Third, they make a browser for virtual worlds — like, say, Netscape or Internet Explorer.

Back in the early days of the Internet AOL was also three companies in one. In addition to their online community, they maintained their own proprietary back-end software, and distributed a proprietary viewer for their users.

Today, AOL still offers a proprietary viewer to its customers.

But, on the back end, the proprietary platform is gone. AOL.com now uses the same HTML as all the other Web pages out there and can be accessed by any Internet browser.

COST SAVINGS

By switching to OpenSim, Linden Lab will no longer need to invest in building virtual world software and instead focus on its core strength: community building. The money currently spent on development can be used for marketing, support, and improving the user experience.

The OpenSim platform is already technically ahead of the Second Life server. OpenSim is modular and scalable, supports mesh objects (with the modrex module). Additional modules can be plugged in to handle local currencies, integration with Web sites, back-end databases, and many other tasks.

There will be some substantial switching costs, as special modules will need to be rewritten to work with the Second Life environment. But, after that, Second Life will no longer need to worry about maintaining the platform — just in having the modules that keep it ahead of the competition.

By doing so, it will, in effect, be getting all the benefits of having IBM, Intel, and an army of volunteers working on their platform, and will automatically ensure compatibility with the rest of the 3D universe.

Instead, Linden Lab is now in a position of having to race to keep up with OpenSim. Full-region backups, real names, mesh objects, hypergrid teleports, server-side modules, megaregions and megaobjects are all currently available in OpenSim but not in Second Life. In addition, OpenSim grid managers can swap out physics engine — say, if they want their world to be in the zero gravity of outer space, or in the low gravity of the moon.

Keeping up with OpenSim will use up resources that could be better spent creating a better user experience, marketing, or community building.

CUSTOM DISTRO

Open source software is rarely used “as is” by enterprise clients. Businesses need to be sure that the software is stable, customized to fit their needs, and that there are support options available if something goes wrong.

The low-cost, self-serve alternatives are generally used by tech-savvy companies, resellers, and by non-profits with more smart people than money — like schools and colleges.

Linden Lab could find a niche for itself reselling customized, high-end, business-friendly versions of the OpenSim platform to enterprise users.

IBM, for example, charges $50,000 for its OpenSim-based Lotus Sametime 3D product — about as much as Linden Lab charges for Second Life Enterprise. And the IBM product offers more functionality and better enterprise integration. IBM can afford to pack a lot of goodies into the software since it doesn’t have to pay for core  development. (Though IBM, in fact, does contribute quite a bit back to OpenSim.)

BUNDLING

Microsoft was able to gain traction for its Web applications — Windows Server, Internet Explorer browser, and the MSN.com community site — by leveraging its existing distribution channels. The Internet Explorer browser, for example, comes pre-installed on all Windows machines, with MSN.com bookmarked and ready to load. Similarly, a company running Windows throughout the enterprise, and familiar with the Microsoft platform and management tools, may prefer to use Microsoft’s Windows server.

However, Linden Lab doesn’t currently have a distribution channel to leverage, unless it is acquired by Microsoft or Google, or strikes a deal with a major partner to embed their server software into their platform — as Microsoft initially launched MS DOS by signing IBM as a client.

Without bundling, there is no clear advantage to having their own platform.

Can Linden Lab be acquired by Microsoft so that Microsoft will have a hold on the new 3D Web?

OpenSim already runs on Windows, and is written in Microsoft’s C# language to run on Microsoft’s .Net framework. Microsoft is better off releasing their own distribution of OpenSim — say, with a graphical management console to replace the text-based one that currently comes with OpenSim. Microsoft could also offer integration tools that would let companies connect their virtual worlds with the corporate directories, instant messaging systems and social networks. OpenSim was built from the ground up to make this easy.

I would be surprised if a year or two down the line, Microsoft wasn’t offering a virtual world server alongside its Web server. The company already has a partnership with ReactionGrid and its own private world.

Bundling with Google’s new operating system is another alternative — unless Google decides to go the OpenSim route as well. Integrating OpenSim into their platform will be cheaper and quicker than integrating the Second Life server software.

Meanwhile, there are other proprietary virtual world platforms out there that use more modern technology than Second Life does and offer better graphics and physics engines.

BOTTOM LINE

If I was starting from scratch building a virtual world, and I decided that I wanted the same level of graphics quality as Second Life, I would opt for the OpenSim environment over Second Life, due to easier customization and significantly lower costs.

There’s an exercise that some business executives do, where they ask themselves: “If I weren’t in this business today, would I enter it?”

There’s no benefit in sending additional money into a project simply out of intertia. But switching platforms is a difficult and costly process. I’m not recommending that the Lindens unilaterally make the jump tomorrow.

NEW GRIDS

One way to make the transition smoother would be to start new grids using alternative platforms.

Educational institutions, for example, require mixed-age environments, a PG-friendly world, real names for avatars, and the ability to make backups of their builds.

A new grid specifically designed for educational institutions could help address all these concerns. By using OpenSim as the back-end server platform, Second Life would be able to offer regions at a lower price point, making them more attractive to schools and colleges, while leveraging the Linden Lab brand recognition. In addition, the Lindens can make the move easy for current customers. This would enable them to compete against upstarts like ReactionGrid. In fact, they could create a separate grid for each new school district.

Similarly, the Lindens can create a new grid for business customers looking for a training and collaboration platform. Companies looking for a retail experience and access to Second Life’s population would still prefer to be on the main grid. But companies looking for privacy, security, real name avatars, and backups may want to have a more business-friendly environment. And, as with schools, a PG rating would enable them to do away with age restrictions, which could be useful for companies using student interns or who want to use the grid to hold focus groups with underage customers.

Switching to an OpenSim platform for these new grids would allow Linden Lab to serve new markets without alienating their current user base. It would also allow them to offer land at lower price points, and roll out features like mesh support as soon as they’re available in the OpenSim universe.

maria@hypergridbusiness.com'

Maria Korolov

Maria Korolov is editor and publisher of Hypergrid Business. She has been a journalist for more than twenty years and has worked for the Chicago Tribune, Reuters, and Computerworld and has reported from over a dozen countries, including Russia and China.

  • Zauber Paracelsus

    If Microsoft buys out Linden Lab, I'm quitting Second Life completely. I don't want SL becoming more screwed up than it already is.

  • Blogland Oh

    Reaction grid is microsoft. Maybe you should check channel 9 a microsoft channel for Reaction Grid. It's all .net development running on microsoft technology.

    Microsoft can never buy out linden labs. Check sl there's no microsoft sim anymore.

    I don't see the point in jumping to another grid (unless it's mica astrophysics n-body experiments) The point of opensim is running your own and experimenting. Second life is the leading competitor so why settle for anything less than the real deal?

  • Rob Danton

    Interesting idea. I think this may be the way that SL was heading under Philip Rosedale, with repeated promises that the server code would be released and if that happened it would make sense to incorporate the technical advances made in OpenSim. But mention of opening up the server code has become much less frequent now to the point that it sees not to be on the horizon at all, just as cross-grid interoperability seems to have been dropped (at least in public) by LL in the pursuit of a more proprietary approach, locking people into the LL world. I think that's backward step that places Second Life under threat from Open Sim and of course the other contenders. I'm skeptical of Blue Mars as a platform but I do think that if efforts to build a Unity-based client for Opensim come to fruition then the universe of virtual worlds will get very shaken up by a combination of great graphics and freedom.

  • "By using OpenSim as the back-end server platform, Second Life would be able to offer regions at a lower price point…"

    I don't understand the argument behind this point, since LL don't have to pay themselves a fee to use their own technology.

    On the whole though, this article made more sense that I initially expected it to from the title. The physics in SL is vastly superior to OpenSim because LL have bought-in the Havok engine. But, imagine if the Havok engine, and other commercial modules, became optional features to buy when setting up a Reaction Grid sim. Adding all of the modules would probably take the cost to equivalent or above what is available from LL, but the choice would be a vast improvement.

  • Tinker LaFollette

    One thing the SL server has that OpenSim *may* lack, is the ability to run at the huge scale that the Linden grid requires. It would require some serious engineering work to test, upgrade, and certify the software for suitability. Which is not out of the question; the result might be the equivalent of the highly customized build of Linux that Google uses on its servers.

    Linden's commitment to its behind-the-firewall enterprise customers may be an obstacle as well; clients paying big bucks for SL server software may not like the idea of swapping server software out from under them, even if would nominally be an improvement. So at a minimum, Linden Lab would have to foot the bill for maintaining the old server software, plus customizing the new platform, at least for the duration of existing contracts.

  • Matthew —

    Just because you're using in-house technology and don't have to pay yourself anything, doesn't mean that the product is free. It costs a lot of money for Linden Lab to develop and maintain its server software. If the software never changed again then you could argue that it's been paid for — why not use it?

    But the platform continues to evolve. Under pressure from not just OpenSim but also Blue Mars and other virtual worlds, Second Life will have to continue innovation.

    Take mesh objects for example. The Lindens promised to have mesh imports this year. That means upgrading the server software, and upgrading the viewer, to support meshes. On the OpenSim side, this is already available (with the modrex module and the realXtend viewers).

    Keeping all these guys employed is a lot of overhead.

    Meanwhile, business and educational grids will need to have Vivox voice support — which is an extra cost to grid operators. But the free physics engine that's available in OpenSim is just fine for these users.

    AOL faced all these issues when it became clear that HTML was the way to go. It had invested a lot of money into its own proprietary technology. And HTML — at least at the beginning — looked really bad by comparison.

    If the Lindens switch over to OpenSim it won't happen overnight. Meanwhile, the OpenSim developers are continuing to work on scalability and load balancing issues. Intel in particular has been putting a lot into this — last week, they were able to run 1,024 regions in a single sim modeling Yellowstone National Park (come and visit — look for Geography regions on ScienceSim grid). Keep in mind that OpenSim is still officially "alpha" software — it hasn't even reached the 1.0 release yet. However, despite all its problems and issues — and despite the fact that the developers remind us of its alpha status on every occasion — hundreds of grids are already running the software. Based on the growth of the top 40 grids alone (not counting all the uncountable private grids out there) I'm estimating a 177% annual growth rate in OpenSim regions — compared to just 6% growth for Second Life.

    Article here: http://www.hypergridbusiness.com/2010/01/opensim-

    Even if OpenSim grows at a steady linear pace and does not accelerate (which is highly unlikely) it's on track to pass SL in regions by early next year.

    — Maria

  • Maria this is a good observation and would be a smart business decision for Linden Labs. Novel owned the network server space for many years but quickly evolved to embrace unix/linux and continues to thrive today. At some point Linden needs to realize the race is over and start to redirect their efforts towards Opensim. This would accelerate the growth of the 3D internet, allow Linden to continue to grow market share, reduce their development and support costs and put linden in a key position. After all today they own 75% of the total grid.

  • "Educational institutions, for example, require mixed-age environments, a PG-friendly world, real names for avatars, and the ability to make backups of their builds."

    and the rest of the world doesn't?

    Second Life? “C’est magnifique, mais ce n’est pas la guerre”

    (saddling members with silly names ensured that Second Life would never be seen as anything other than a niche fantasy game by the majority of internet users, no matter how much real money is spent there)

  • Just like AOL in 1997, LL is "too cool" (in there minds) to want to be "just an isp" so while the idea makes sence -big picture- its too late. LL is locked into the same tech biz models that befell AOL. YAHOO, and others.

    And the only BIG "business" hope they had- PAYPAL plus protection- has been lost by the utter disdain they have shown to 3rd party IP owners over the years.

    The new SL viewer will determine their future.. either another 3-6 years as a quasi-game-platform entertianment- or another forgotten AOL, Compuserve,within the next 2 years,

    Other Futures I really dont see.

  • Pingback: Episode 138 – iPawned | Dogear Nation » Podcast()

  • plain and simple: Linden Lab's "Second Life" has 70000 online users at any given moment. Second Life has an economy. Second Life has Linden Lab.

    Open Sim (which is also Linden Lab) is not Linden Lab's Second Life.

    Here's an analogy:

    Second Life is to Open Sim as Facebook is to a Ning group.

    i don't know how to explain it any better than that.

  • Pingback: What are people saying? – « SLookable – Second Life Social Media Guide()

  • Anonymoose

    If LL were to switch to Open-Sim, they would have to spend 5 years rewriting Open-Sims LSL implementation, to actually work right. Not that it would take that long, just thats how long they would take. I have a feeling, most of their programmers do nothing 6 days a week, and work for one hour on the seventh.

    Also, the LL platform is programmed in C++ and ASM. Providing LL with the best performance on the low power systems they use to host. If they switched to Open-Sim they would spend all of the money they "saved", upgrading every system in the data centers. And They would again RAISE sim purchase costs to cover the higher end machines.

    [WORDPRESS HASHCASH] The poster sent us '0 which is not a hashcash value.