go read it now, i'll wait.
okay. back? great! let me add a few notes.
in Second Life's early days, it was pretty clear Linden thought of it's creation as an online service for the creation, display and sale of 3d content. back then, you heard a lot of people at the lab saying things like "Your World, Your Imagination." i've heard people say uncharitable things about Linden's founder Philip Rosedale, but seriously, whatever his faults, he was VERY good at internalizing and communicating Second Life's fundamental message: "This is YOUR playground, Linden is here to make sure the plumbing works."
throughout the early-mid 2000's, the SL community rambled along, doing cool things, showing them off, building an early community. after the hype cycle started kicking in, corporate america showed up and a few of their denizens thought (and i paraphrase here): "oh hey, check it out. it's like WebEX except i can build 3d doo-dads to demonstrate concepts and since it's a 'game' i can get away with looking like something other than a corporate drone."
as IBM and Cisco and other corporate heavy-weights started looking at SL as a tool for business operations (rather than just marketing,) you started to hear a lot of talk about SL as a "platform." up 'til this time people had used the word "platform" synonymously with "service." but now the big guys wanted to run a server or two behind their firewalls so corporate secrets wouldn't have to leave the corporate intranet. there's absolutely nothing wrong with this concept; but it's not what SL was initially built to do.
a couple "one off" solutions were built for various corporate customers, but their features were never integrated into the main-line SL service offerings. Linden management was caught in a dilemma: their casual consumer customer base was exploding in 2006, but corporate users seemed like a great market. Linden's challenge was how to increase stability and capability for consumer users while adding features for corporate customers.
around 2006 Linden, IBM, Cisco, Intel and a few other corporate heavies started talking about interoperability in virtual worlds. this eventually led to the Virtual World Interoperability Forum. the VWIF was supposed to produce detailed functional descriptions of how virtual worlds work so there was a good chance virtual worlds operated by different companies could talk to each other. the VWIF eventually sputtered, giving rise to the Second-Life focused: Architecture Working Group (AWG), Open Grid Protocol (OGP) and eventually the Virtual Worlds Region Agent Protocol (VWRAP) working group in the IETF.
but please note, when a crowd of network interoperability people use the term "platform," they don't mean the same thing as when second life users (and even most Linden management.) Second Life management had used the term "platform" to mean, "hey, here's a service you can use to do cool stuff." they did not mean "hey, we have a technology platform you can download, extend and deploy."
but throughout the latter 2000's when Linden was working on their "behind the firewall solution" for corporate customers, there was a small kernel of Linden employees that did intend to make a real-life, honest to goodness "technology platform." the idea was to corral linden's server software into an appliance customers could plonk down in the middle of their data center. viola! instant virtual world populated only by corporate users.
for a lot of different reasons, this turned out to be harder than initially suspected. one reason (certainly not the only one) was the difficulty of maintaining two distinct code bases: one for corporate, behind the firewall use and another for the consumer solution. because Second Life was always considered a "service," there were a wide variety of assumptions the code could make about network architecture and peer service availability. and it turns out that some of the assumptions you make for a huge virtual world with 85,000 concurrent users are not the same as the assumptions you make for a couple simuators and a smattering of "behind the firewall" users.
whatever the reasons, the project to convert the Second Life "service" into a behind the firewall "platform" failed. (though it should be noted a lot of very skilled Lindens did a lot of awesome work to shoehorn Second Life into an appliance.)
about the same time Second Life/Enterprise (SL/E) was being discontinued, Linden's official support for an open virtual worlds interoperability protocol evaporated as well. sic transit gloria munde; at least there's still OpenSim.
but what relevance does this have to the modern world?
does it mean there's no market for "behind the firewall" virtual worlds?
does it mean that Second Life as a "platform" is an architectural dead end?
what about content? and the network effect?
let me offer my two cents about these questions.
first off, Second Life is not and nor has it ever been a "technology platform." Second Life is a "service." it was coded by people who made certain assumptions about their environment (like where the asset servers existed, that peer services were "trustworthy" and what features end users demanded.)
so my first assertion is, the failure of Second Life/Enterprise (SL/E) and VWRAP does nothing to refute the existence of a market for a well-developed, well-supported "behind the firewall" virtual world solution. SL/E's failure simply means it's damn hard to cram the Second Life code base into a small number of servers. in fact, it's so hard that the NRE and support costs are waaaay out of line with any income the service could generate.
i also believe that Second Life, as a technology platform, is dead. as a service? sure, it's still cranking; there are a bazillion people (myself included) who log into SL to socialize and play. what's more, i think Linden also thinks SL as a technology platform is dead. recent actions from Linden management make it clear they're putting all their resources into making SL as compelling an online service as possible.
Tateru Nino speculated that a team of crafty engineers could rebuild something like Second Life in about 6 weeks. i'm more inclined to say it would take 18 months before you get something close to SL's performance and concurrency. were i to develop a business plan for a startup doing more or less the same thing Linden's doing, but with a better, cleaner architecture and the ability to deploy as a real technology platform, i would estimate about 15 man years to beta launch. taking a completely arbitrary figure of $200,000 per man year, you wind up with a $3 million price tag.
so this has got to be how the VCs, angels and midnight engineers are looking at it: given Linden's flat subscriber growth, why would someone risk $3x10^6 on a new virtual world platform?
how 'bout something like a "Social Virtual World" that uses identity information from Google, Facebook, LinkedIn, Twitter (or even your own corporate LDAP server) ??? that might be compelling.
or even a "virtual world technology framework" that niche application developers could easily extend. that seems like it would be a good idea.
maybe even a "content framework for 2d and 3d experiences" where you could easily develop some kind of 3d model using blender, maya, etc. and have it automagically uploaded to the cloud where virtual world servers could easily import it. bonus points if you work through the intellectual property issues.
so there are a lot of interesting things a VC or angel could fund and an entrepreneur could build, but rebuilding Second Life? no thanks, i'll pass.
No comments:
Post a Comment