Breaking News: Blue Mars Launching Cloud-Based Option This Year: Low Lag, High-Res 3D Virtual World Without High-End Hardware -- Can Second Life Technically Compete?

Update, 4/15: Linden Lab replies about cloud options for Second Life here.

Blue Mars will offer a cloud-based service this year, beginning perhaps as early as this quarter. I confirmed this rumor in a chat tonight with Jim Sink, CEO of Avatar Reality, the Honolulu-based company behind the immersive 3D virtual world that recently launched in beta with a couple content teams who originally began in Second Life, and ex-Linden CFO John Zdanowski. (Sink just posted an official announcement of the cloud service on the Blue Mars blog.)

What's this mean? For technical background on cloud-based game/virtual world systems, read this article I wrote about OnLive, one of the services Avatar Reality may use for Blue Mars. (The company has not announced its partner yet, but will do so very soon.) Other contenders include OTOY, which I speculated about last week, and a Japanese offering.

What's this mean in practical terms? Well, look at the Blue Mars waterfall screenshot posted above -- cutting edge 3D graphics with dynamic shadows, a viewing distance of 2000 meters, and other ambient eye candy.

Now imagine it displayed on pretty much any computer: say a three year old laptop, a $300 notebook, or even an iPad. Imagine getting a world like that on a small browser-type program, displaying about as quickly as you'd play a YouTube video. It gets better: Imagine simultaneously sharing that waterfall space with 3000 other Blue Mars users, logging in from all over the world, with little or no lag. Getting similar graphics quality in Second Life would require a high end system costing several thousands of dollars, and even if your SL client was connected to the best broadband in the world, lag caused by other users would be unavoidable.

"I can confirm that we’re going to offer Blue Mars as a cloud service," Sink told me today. When I first wrote about Blue Mars in 2007, it was planned as a huge downloadable client. However, Sink tells me, the company has always been investigating a cloud-based service:

"We've been experimenting... from the very beginning," he said. They've run Blue Mars on some of the leading cloud systems, and came away impressed -- they were even able to run it from Hawaii connected to a mainland-based cloud server, "and it's very satisfactory." Or as he puts it, "The ability to play Blue Mars on a thin client, on a Macbook Air, is really strange."

Or an iPad. Here's a look at Apple's new tech fetish item running Crysis, a game which shares the same 3D graphics engine as Blue Mars:

If anything, Jim Sink is even more excited how a cloud-based system solves the problem of lag caused by other users. With a cloud system, all the users connect to the same server before their interactions are rendered in-world, so "they're all on the same machine... the problems of concurrency are very easy to solve when they're in the same room."

This announcement, needless to say, presents an interesting competitive challenge to Second Life. As a world, its unique market distinguisher is dynamic content creation. However, a cloud-based system is not an ideal way to handle dynamic content changes. (Updates to Blue Mars content are not dynamic, but first made on a local computer, and only then sent to the development server.) If you've ever tried to edit a Google Doc with several other people, imagine how much more difficult that would be if you were all editing 3D content. Indeed, it may be technically impossible to adequately run Second Life in a cloud as it's currently architected. (Ironically, nearly all of Second Life's most popular places do not leverage dynamic content creation, but remain relatively unchanged by their creators for many months or even years.)

Blue Mars will also have a downloadable client, but for Jim Sink, making a cloud-based option available is part of making Blue Mars (and by extension, immersive 3D virtual worlds in general) a mass market product. A category enjoyed not just by hundreds of thousands of people, or even several million, but tens of millions.

"I'd certainly love to share Blue Mars with 50 million people," as he puts it. "[But] we're not going to be able to do that if they need a high-end computer."

Comments

[Hypatia Callisto]

smugly looks at people with a big "I told you so" look. Especially about OTOY.

[Net Antwerp]

A lot of private information would be passing through OnLive hands, like business talks and private (IM) chats. An unnecessary middleman added to the complex VW arena.

Might work with general visits to Virtual Museums etc but in general, OnLive etc is just another unsecured proxy, when it comes to Virtual Worlds.

Second Life Killer? Highly improbable. Blue Mars just doesn't seem to have the same premise as Second Life did - before CEO Kingdon abolished the "Your World, Your Imagination" motto.

[Ann Otoole]

Doesn't change anything. It isn't even newsworthy. It still is not user generated content and it sucks. There is no mature content allowed. It is a failure out of the gate. Oh wait they are suckering rich people into giving them loads of cash to maintain the hono surfer dood lifestyle so it is a success.

Oh wait they hired an EA suit to further ruin it totally. Can't miss.

[Jon]

Actually you can generate user based content as we speak, on a much more detailed core level than any other virtual world offers at this time. It is not the same platform as Second Life, as a Mac is not a PC they are similar in what they set out to do but architecturally different. The staunch Sl'ers will wave their flags in protest while others will benefit from both platforms. EA is a very successful games publisher. I would think an ex employee would have some valuable input to give Avatar Reality. Good luck to AR...their success will determine the users' success and so forth.

[Markus Breuer (Pham Neutra)]

Hamlet, a cloud-based gaming platform, like OnLive, could be applied to Second Life without any problems - including dynamic, collaborative content creation. For all the games it supports, OnLive runs the CLIENT versions of theses games on ONLive's own, specialized servers which are connected to the actual user via a kind of video connection. The servers (sims) could even stay in Linden Lab's own data center in such a setup.

It is just a matter of Linden Lab agreeing to such a deal and OnLove (or OTOY) expecting an audience which is big enough. Such a deal could be cut - if it makes sense for both parties - within a few months.

SL performance might become better with such a system, if the CPUs running the servers and those running the clients were arranged in the same local net - with more bandwidth and less latency. But in principle, the current software used by Linden Lab and us residents could be used nearly "as is".

Such a setup would NOT necessarily lead to a significantly improved performance on the server side. For example, the maximum number of avatars on a server would NOT be greater and the servers would still become laggy when too much is happening on a sim or too many avatars are present - (not taking into account positive effects resulting from more bandwidth and less latency between server and client).

Cloud-based gaming platforms are NOT a silver bullet for curing all problems typically found on MMO platforms. They substantially reduce the hardware requirements for "optimal client systems", though.

[Ehrman Digfoot]

I'm so excited, I've been waiting for this news for nearly 6 months now!

Hamlet, thank you so much for investigating this.

[Arcadia Codesmith]

I think Markus' comments are a pretty good summary. This does nothing for server-side performance issues, but is a great option to have for systems that aren't monster gaming machines. It doesn't preclude dynamic content creation. I'm looking forward to seeing for myself how well it handles on some of the more demanding games.

As for Blue Mars... I've been resistant, but if they can deliver what they're promising here, it might be worthwhile to jump through the extra hoops necessary to create content for it.

But if I did, I'd be pushing them to add comprehensive, accessible, integrated building tools. I'm annoying like that.

[Mary Damer]

I had thought that Macs couldn't run blue mars -- Has that changed?

[hm]

So... you run the client on a server somewhere and it feeds you a video stream of the game.

How much, per minute, is that going to cost?

[Not You]

OnLive has some of the same minds that created the original video compression technology that lead to what we currently have today.

Want tech specs? Check out OnLive's Youtube channel, packed full of information about how it all works.

OnLive could work really well with Second Life, considering the client side rendering is almost always the main issue for people, offloading this onto servers that can just handle it would be an awesome way to migrate that massive issue with SL.

From the setup each OnLive server is custom made with Graphics Cards (not a server staple), each server runs Windows to increase compatibility, publishers take their Windows game, make some slight modifications, and then package it to run on OnLive servers.

Only thing though, in the demos, the big question was asked (what sort of latency is there) and the two presenters choked for a few moments, it is the big question, without it being "Lagless" it will fail for most gamers, but for a virtual world like SL it would not matter as much.

This is Blue Mars only option for mass adoption, the client requirements are HIGHER than SL, and SL is pretty demanding to begin with, using the Crysis Engine was a mistake from a marketing point, considering the sliver of people who can run it.

SL could level the playing field considering this is like releasing your app onto a console, the app is ready to go! The only issue being, is if these cloud services will deliver a "Local" experience.

[TheShadow99]

First let me say I've signed up as a developer for Blue Mars so I could kick the tires and take it for a spin. Right now it's in very early 'beta' or maybe late 'alpha' stage. But if you want you can take a look.

Ann, first to you... It is user created content in Blue Mars and it is Adult content. I've talked to the Blue Mars designers and eventually it will have adult content, far more basic things are in the works right now (IM is almost brand new and iffy for instance). Also with no 'builder' tools integrated into the viewer for Blue Mars getting developers to use outside tools and then important them through the developer client will be a bit of effort. For instance building clothes (which are 3D objects) requires photoshop and Maya/3DS Max unless you want more hassle on your part.

Now some other notes... My 4 year old machine can tun Blue Mars fairly well... Maybe not 'great', but enough for a social virtual world... I can even play mini-games... Getting 14 FPS isn't that big an issue in something that's not a FPS...

I'm neutral to Blue Mars though in general and I've handed them a hefty list of things they need to work on... But if they can work out some of the details and get enough developers (including the fairly high number of SL developers I've seen in world and on the forums for Blue Mars). Then they will succeed at least as well if not better than SL...

Oh I should point out the Mac market is the biggest loser if they don't create some tactic like this... CryEngine2 is DirectX based, Linux can emulate DirectX calls and pass them to OpenGL, Mac's don't come with this sort of functionality by default... Though since CryEngine2 has been ported to consoles... their is a possibility of console support at some point without 'cloud computing'...

[Hamlet Au]

BTW, I'm checking with the Lindens if Second Life could be deployed in a cloud, and more key, whether that's in their plans.

[Ignatius Onomatopoeia]

Cloud-based VWs that are platform agnostic, running in a browser or some sort of basic client, would be a win-win here. This year, 40% of the students came to campus with MacBooks and use wireless exclusively to connect. So no Mac OS, no interest on our campus by the two faculty who teach with SL so far.

So far, it seems that Avatar Reality has not shown much interest in the edu market (on whatever platform), but that could change.

It's beyond the skills of our faculty do design the type of UGI Blue Mars requires, but that's not a show-stopper. After all, high-end Web coding is beyond the skills of those colleagues who learned HTML in the mid 90s and are not coders by inclination.

We cannot build in Heritage Key, yet that is becoming a great venue for education.

Looking forward to what Avatar Reality announces!

[Hitomi Tiponi]

This IS a game changer - unlike Viewer 2 or Shared Media.

[Dwayanu Weyland]

Re SL going to cloud, TANSTAAFL !!

All the processing still has to be done, just would be more of it done on the servers, and less on the pc's. This would require Lindens to majorly increase their server farms, just to support the existing max number of concurrent logins.

And for the comment about most SL complaints being about slow rendering on the client, I'm on a hefty PC, and *my* complaints are mainly about lag from server slowness ;-)

Don't see any possibility of Blue Mars having the cloud server capacity to support a significant number of simultaneous users (say, 10,000 ?), any time in the next few years. But what the heck, having Blue Mars be a pioneer here ("pioneer" == "full of arrows") is always good for the rest of us, to learn from their mistakes.

Also re Blue Mars, I personally find their top-down content creation constraints, to be the antithesis of SL's freedoms. ;-P

[kanomi]

The Internet cannot even reliably deliver 2D flat web pages 95% of the time without some sort of broken script hourglass loading nonsense, and we are supposed to believe real-time, multi-user, dynamic, streaming 3D environments are next year's cookies?

Not a chance. Not a chance in this or the next 10,000 parallel worlds - not until you have light-speed internet fibre optic from server to terminal.

These gregarious salesmen are like 70s arcade game operators: "Hey! look at our amazing vector graphics! So realistic, you'll think you're that Luke Farwalker kid!"

[LittleLostLinden]

I have been to Blue Mars.

There is nothing to there, so I will check back in another year.

[Corcosman Voom]

I looked at the OTOY vid on YouTube several months ago and was very excited.

I read an article a couple of weeks ago about the cost and cooled down considerably.

[Komuso Tokugawa]

Suggest you check High speed broadband speeds/QOS and penetration rates before you start salivating and hyperventilating over cloud based video game/vw services. A great idea, a little ahead of time you might find.

[Sidney Smalls]

It's easy enough to buy a better graphics card. Getting a fatter data pipe can be very expensive to impossible. I could be wrong, but I don't really see this panning out.

[Dennis Bacsafra]

I've always wondered why Avatar Reality built a VW with a client requiring so much hardware power. This cloud thing can definitely make it appeal to a larger audience. I would have to see it for myself though to really get the pros/cons.

[William Burns]

BlueMars is ignoring two very important factors which are not immediately apparent in the beginning of such an idea:

1. What is considered "high-end" computers today will drastically improve over time as per Moore's Law. If they had the slightest understanding of computers and the industry they would instead be looking for more intelligent ways to lower the bandwidth while keeping dynamic fidelity.

2. They will never effectively trade the Achilles Heel by centralizing on a cloud to stream the content. They are trading one brick wall for another, but it won't be apparent until they get the bandwidth bill from the cloud service.

3. Cloud streaming is a short sighted approach to a long term problem. It'll be great in the beginning, but there sure as heck won't be tens of millions of simultaneous users - not unless BlueMars is alright with charging people $50 a month for access to cover that bandwidth bill they are going to rack up.

Sorry, BlueMars... I'm not following your hype machine off the cliff.