Via Jeanricard Broek comes this demo of the high-end 3D game Crysis running on an iPad. To be more exact, Crysis is running on OTOY, a new cloud-based game service, which processes and streams a fully rendered, lag-free version of Crysis on an iPad. It's so impressive, Mr. Broek wonders, "Is Blue Mars or Second Life next?"
Given that Blue Mars runs on the same 3D engine as Crysis, I'd say it's quite possible (and likely) it will go into the cloud. But can Second Life follow suit? There, I'm not sure. Last year the CEO of OnLive told me of his own cloud-based service, "You could design the next Second Life on this!" But it's quite possible he doesn't know all the special rendering and processing challenges inherent in a fully-streamed, user-generated world like Second Life. (I imagine it'd be kind of like creating a 3D version of Google Docs that thousands of people can simultaneously edit.)
I put the question to all the tech experts reading this -- is going into the cloud a realistic option for Second Life?
Blue Mars has already experimented with it too.
"lag free" depends on how you define it. What you get on the client is a video stream that you can interact with like it is the game/program itself. But because it is a video stream, there is some delay in interaction. All depending on your available bandwidth.
Such a delay would not be a problem for SL or Blue Mars which are more social places, where twitch razor sharp responses are hardly needed.
So yes, a service like this, would be excellent for Second Life and Blue Mars. It will not only bring them to mobile clients, but also to other less powerful mainstream clients, like a 5 year old pc. Most people don't have a system to run everything with the highest graphics settings, with a service like this, people would be able to experience SL and 3D games on the maximum eye candy settings.
I don't see insurmountable hurdles on getting SL on such a system, they would need to be able to run the client and have it connect to the LL services. If LL was smart they would make a deal with such a service to have a big pipe open between the LL backend and the service, so the world would load as fast as possible. And give premium accounts a discount to the cloud service.
One thing that LL might have to focus on is making the Client rock-solid, so it wouldn't crash. And there would need to be a path to reboot the client if it did. But that is something the Cloud services have to consider for all games too.
Extra super duper bonus you get from this, is that through this there would be far less vectors of possibility to pirate any content. You don't have any of the 3D models and textures locally.
I can see a platform like this become the distribution platform of choice for 3D games/programs, etc.
Posted by: Frans | Friday, April 09, 2010 at 04:29 PM
Isn't HTML5 the more obvious route? It's potentially possible to write games in HTML5 and play them right in a compatible web-browser - like this demo http://www.youtube.com/watch?v=XhMN0wlITLk
So write an HTML5 version of SL and play it on your iPad or future iPhone or any other HTML5-compatible device with enough horsepower?
Posted by: Jovin | Friday, April 09, 2010 at 04:54 PM
If it is written in HTML 5 it would still need to compute on your physical device, it would need lots of memory, a graphics processor with all the latetst features, etc.
The point is not only that it is on the web, but that all computing is done on the server, and then streamed as a interactive video stream. That stream could then potentiality be embedded in html5, of course.
Tada we have SL on the web, with web on a prim in SL which play SL on the web, in infinity.
Posted by: Frans | Friday, April 09, 2010 at 05:40 PM
i dont know if you 've heard of the 'vollee' experiment back in 2008. it worked just fine on a symbian device. check this:
http://naturalbornkiavik.blogspot.com/2009/09/sl-mobile.html
Posted by: Definitely Gibbs | Saturday, April 10, 2010 at 01:17 AM
Interesting stuff, personally I think SL on the web is a bit of a blind alley but I realise it's probably inevitable we all get dragged down it.
Coincidentally the iPad/iPhone/iPod touch ecosystem just got its first fantasy MMO (I think it's the first?) with Pocket Legends - more here: http://www.spacetimestudios.com/
Posted by: Jovin | Saturday, April 10, 2010 at 04:51 AM
Wonder how cost model will work. Will it be expensive for the operator (continual video streaming to the client, intensive GPU/CPU processing in the data center etc).
Would that affect in-world pricing to the end user, don't think they could charge everyone on subscription basis like OnLive/cloud gaming guys are looking at.
Curious to see how it will shake out.
Posted by: Simon Newstead | Saturday, April 10, 2010 at 10:28 AM
OnLive has been in beta since last fall and they're going into public release this June with a bunch of games. (Public release means in the 48 contiguous states. Why not Alaska and Hawaii? Read on...)
The lag between the time when you press a button and the time you see its effect on your screen is known as the round-trip time, or "latency". It's basically the time it takes the signal to travel from your computer to the server + processing time at the server + the time it takes the video signal to get back to your computer + video decompression time on your local computer.
Modern computers are so fast that the longest of those times is the travel time from computer to computer (through the internet). It's limited by the speed of light (a constant of nature) and the speed of intermediate routers.
You can crank up bandwidth as high as you like, but you can't increase the speed of light. You can make the highway wider, but you can't increase the speed limit.
Suppose my sim is simulated and rendered in Dallas, Texas. If I live in Dallas, the latency will be low and I'll be happy. If I live in Australia, the latency will be high and I'll be unhappy.
So, you say, why not have a server in Australia? Okay! Go ahead and move the server simulating your sim to Australia. Now the Australians get low latency but the Americans get high latency. You can't make everyone happy.
This problem is mitigated for sharded worlds and games, because everyone in a shard can be geographically close to the servers simulating their shard. Second Life isn't sharded.
Still, I like the idea of server-side rendering. The graphics would look great. Textures and 3D models wouldn't be sent to client computers so they'd be harder to copy. As Frans said, lag doesn't matter as much in social worlds.
It sounds like Blue Mars will test the waters. I know I'll give it a try when they do.
Posted by: Troy McConaghy | Saturday, April 10, 2010 at 10:58 PM
Web.alive has been running on cloud for some time now. Check out the MellaniuM Dome at
http://wa-184-73-220-81.projectchainsaw.com
Posted by: Tele3dworld | Sunday, April 11, 2010 at 02:31 AM
OnLive isn't true cloud computing so much as a subset of cloud computing that started as a parallel technology known as "virtualization." It is ironic to use the word "virtualization" when talking about virtual worlds because they have absolutely nothing to do with one another.
With Virtualization, all of the computing is happening at the server side (you could say in the cloud) so both the server AND a virtual client are all up in the cloud and the only thing that needs to be transmitted to the remote user is the screen updates... like watching a live machinima of someone else logged into the world.
This concept is superb for non-UGC experiences like authored video games and what not. But when it comes to a platform where users are generating content actively and they create as well as consume, the concept behind OnLive becomes an inherent mismatch.
There are smart people who totally get this idea like ENTERMETA that have a massive running start and figuring out how to do things things correctly. The bottom line is SL's architecture is too old and not scalable enough to take advantage of these technological advances. Blue Mars might have a compelling gateway for say mobile devices using OnLive and maybe SL could tap into that but again that is virtualization only. But like I said a few smart people have already figured out how to make this work and they are light years ahead of both SL and Blue Mars.
Posted by: Money | Sunday, April 11, 2010 at 06:22 AM
Entermeta has a one-page website where you can send them an email. Google searches for Entermeta turn up nothing else. No blog posts. No news articles. Nada.
I did more digging. I checked the WHOIS domain registration information for etermeta.com. It turns out that domain name was first registered Dec. 27, 2009 and they only registered it for one year!
Yet Entermeta somehow has a massive running start? Forgive me if I'm not convinced.
Posted by: Troy McConaghy | Sunday, April 11, 2010 at 11:19 AM
Most problems wins. Eg - the bottom is curved so if you are using it on a table it will tilt and spin and you can't touch type.
Cho Yung Tea
Posted by: smithevne | Sunday, April 11, 2010 at 10:23 PM
Money really has no idea what s/he's talking about. Blue Mars is not figuring out any sort of cloud computing at all.
Second Life is very different from Blue Mars on the backend too. Putting them into the same sentence is like saying
Apples and bananas don't have a head start on evolving into mammals. Well, now does that make sense? NO.
I'd almost say Money is just spamming for some kind of startup that s/he's got some kind of financial involvement with. Stay away from this person.
Now, OTOY is real and of any of the cloud based services, was on the scene first. They seem to have a patented streaming compression technology that in tandem with AMD, may in fact be a massive contender. Blue Mars can run on this system, just like ANY CRYTEK BASED GAME CAN.
Duh. With a resounding THUMP.
Posted by: Hypatia Callisto | Wednesday, April 14, 2010 at 01:28 AM