When we first looked at OTOY about a year ago, the small company was trying to deliver a server-side 3D rendering technology that could allow modern video games to be played on basically any client. A lofty goal, for sure. Then OnLive was unveiled in March at GDC, and it sent ripples around much of the gaming world with a similar concept of cloud-based gaming (both good and bad). But OTOY believes it has a more lightweight and elegant way to do things, that is also more extensible. And it has a couple of key partnerships to prove it: EA and AMD.
Let’s run through some of the details quickly: OTOY is 100% browser-based, and works is all modern web browsers. All it requires is a broadband connection, and that will give you 720p (HD) graphics, with no plugins and no downloads. There is also a way to get 1080p graphics, though that’s a bit more intensive, obviously. But they key is that this is all done on OTOY’s servers and transmitted down from the cloud to run on whatever client you want, basically instantaneously.
Again, all of that may seem hard to believe in a world where the giant power-hungry systems like the Xbox 360 and PS3 are the only way to play graphic-intensive games. So the best way for you to see it, is to watch a video of it in action. Note that while in the video below, the player is using an Xbox 360 controller, he is playing it on his computer hooked up to his TV. Also note that this demo is being played from a server that is 400 miles away.
Pretty awesome, right? But we’ll understand if you’re skeptical. So here’s a video of our own Jason Kincaid playing a version of the hit title Grand Theft Auto on his computer, through his browser. [Note: At one point, Jason says "one GPU core per instance." But actually we're told that the technology scales so efficiently and cost effectively that you can allow anywhere from a minimum of 10 users per GPU, up to 100 depending on the application.]
And when I said this can basically run on any client, I meant it. Like perhaps your phone — at 60 frames per second. Stay tuned.
The videos both don’t work. They are apparently private.
Sorry about that, my changes must not have taken hold. Try it now, should work.
This is amazing.. i wonder how this technology would be used for online interaction other than gaming! it should be pretty interesting having an entire campaign built on this
The future of this will partly be in timed demos and rentable games. Play a demo version for 10 minutes per day, or rent a game instantly for a week, etc.
The “hit EA title Grand Theft Auto” is actually a hit Take-Two Interactive game.
So, we’re calling everything on the Internet a “cloud” now?
The internet is the cloud…
It’s a web, god damn it!
tubes!
Technically, the Web (short for World Wide Web) is a combination of a GUI browser and a managed encoding of URLs and router nodes. The Internet is the Network of Networks for data. The cloud, on the other hand is a collection of server hardware, supporting load balancing and other hardware, the telecommunications routers and switches that connect them and the software running and each with protocols that speak well with each other.
I get so tired of people not knowing the basics.
Neat.
In a good way, not like Martha Stewart.
Is it the YouTube video on the first one or did the gameplay seem a bit laggy?
Wonder how it will run when several client will connect simultaneously
Very interesting tough
The technology definitely has a future in gaming, but it will not replace our existing medium by a long shot, but instead will take a novelty “alternative” approach- You can either buy it in stores or play through your browser.
The tech to do this is *very* expensive and I am not sure how they plan to cover the costs. Keep in mind, the server thats powering your gameplay takes electricity, bandwidth and constantly updating hardware to support the latest games. If you buy one game for 40$, how long until you have eaten up enough electricity/bandwidth/hardware that they are losing money, not making it.
Another huge downfall of this system is the lack of “custom” mods- Meaning for games like counter-strike and the like, custom map, Player hosted servers and other unique player created items will be a thing of the past.
which again is why it will be a alternative, not a requirement.
I’m not sure I agree with you entirely, and I say this knowing very little about the gaming industry. One thing I do know is that Microsoft and Sony make nothing (or lose a little) on each piece of hardware they sell; they make the money back on software licensing.
So for your average customer, if they never buy any software, they equal maybe -$400 or more.
I think what this model begs is a subscription to gaming software. You’re assuming we’ll always go to the store and shell out $60 for a game. If music, movies, TV shows, iTunes tell us anything, it’s that buying discs in stores has a VERY limited lifespan.
Since bandwidth and server farms are pretty darn cheap when scaled, I think it’s more likely that we end up subscribing to a game for a few dollars a month, and just like the App Store is innovating, even support in-game purchases.
If you remove the need to shell out $600 for a gaming console, suddenly you are also open to a MUCH broader market. Heck, it’s the main reason why I don’t own a PS3 or XBOX. I play games on my iPhone because I think the console price is ridiculous – and having a bluray player is not an incentive for me.
This won’t be an alternative. Remember, the gaming market itself is a niche market (though that’s changing). A billion dollar niche market, to be sure, but certainly not a mainstream industry. There’s no room for niches within a niche. This will either die, or replace.
Change the model and the economics work pretty well.
i don’t think that’s a fair statement. since july ‘07, over 50 million current-gen consoles have been sold (that does not include the ps2 or the incredibly fast selling nintendo ds or dsi). and sure, lots of people probably own two consoles, but that’s still possibly 25 million people that have video game systems. we’ve only got about 300 million here in the states. that’s 1 in 12. as time passes, more and more kid gamers are becoming adult gamers. gaming is not a niche.
brazilian fart porn is a niche. gaming is not.
I agree and the PC gaming market, though declining, is still larger than the console market. So make that 1 in 6 or 1 in 5. Really stretches the definition of ‘niche’…
I wouldn’t really consider gaming a niche market at $11+ billion solely in annual software sales and $22+ billion when you include hardware. Also, considering it is surpassing the music and movie industry, when exactly would you say that it is no longer niche?
Just about anyone can make a movie or song, but you really can’t make a movie and expect to see it in theaters (maybe on a few video sites, but most videos are under 10 mins or so, unless its professional, or indie) until sundance gets a hold of it, or some other major video event
video game’s on the other hand can be found on consoles via marketplaces, on the internet, on itunes, ect, there are far more venues for your video game to be released then your music or movie.
movies may have one or two formats, digital, or dvd.
videogames not only have the obvious markets, like ad or subscription supported browser games (more so ads, and sometimes freemium, like runescape) and the consoles, but the pc market, the cellphone market, the toy market (garage games, maker of instant action, had at least one game being sold as a net jet game (net jet is made by hasbro) and nintendo pulled the whole rob thing to keep in the market after the game crash)
You’ve gotta be friggin kidding me. Video games is a “niche” market???
Worldwide sales of Wii/Xbox360/Ps3 are around 100 million units, which DWARF sales of your iPhone… and unlike your iPhone, people buy these systems purely to play games.
For that matter, worldwide revenue of all video game software surpassed box office revenue about a DECADE ago.
Say what you will about OnLive/OTOY/Cloud, but get your facts straight.
[quote]You’ve gotta be friggin kidding me. Video games is a “niche” market???[/quote]
Um…He plays all his games on the IPhone; what did you expect him to say? Tetris is probably cutting-edge to to him.
HAHAHAHAHAHAHA !!!!
Love it !
“Another huge downfall of this system is the lack of “custom” mods- Meaning for games like counter-strike and the like, custom map, Player hosted servers and other unique player created items will be a thing of the past.”
That doesn’t make sense. You could host games, maps, or mods on their servers. In fact, OTOY would improve the mod experience. You wouldn’t have to wait for any custom map or anything to download. It would all already be there on the server.
How would you upload a map/addon u’ve created from scratch onto their servers ?
With their permission of course,
Im sure some type of developers area could be made, but hardware and everything else might require a premium account of some sort.
@dan : I would have to respectfully disagree. This technology (assuming it works as shown) IMO will be adopted extremely fast by many new game developers.
As a gamer I consider 2 things when buying a game: (1) will it work on my system (Console or … OS, GPU, CPU, System Resources) and (2) do I want to shell out that much cash for a game I have never played.
This setup can theoretically make both of the dilemmas I mentioned go away.
(1) No more graphics card “arms race” that all hardcore computer gamers seem to be in, and an immediate increase in potential customers who have older or cheaper computers which were previously considered inadequate gaming machines.
(2) Subscription fees make total sense for this business model. I shell out $20 a month right now just for WoW. I would gladly put up to $50 a month into a subscription service if it let me immediately play a few games in a queue on any device or machine with a web browser.
When you take those things into account and then add such benefits as the examples below, it makes this technology extremely compelling for every game company:
- Graphic quality can be upgraded for games without affecting client machines
- Support and patches will be reduced significantly
- One game could theoretically work on every web enabled computer, device, or console without the need to port to other platforms making it possibly a more profitable solution
Ok, first thing is first: THAT’S GTA RUNNING ON A MAC! Cool. Mac’s got games now, whether their publishing on it or not. It’s the only thing that keeps me on PC, including the ability to make that kind of artwork in packages of my choice.
But how did you say EA made GTA? EA did not make GTA. EA has been trying for years to get at that franchise and a few pieces of it that interest them.
Rockstar Games is the polar opposite to EA in that one is a corporation employing artists and the other is artistic individuals employing the benefits of corporate endeavors.
Please, do take the time to learn about Rockstar Games and EA games, separately and their relations, because you will probably find great interest at the revolutionary things Rockstar does and has done, and be able to contrast and appreciate them with the opposite example set by EA. Look at things globally and not by a game by game basis. EA has made so many games and owns so many people they were bound to have some nice franchises – the dark side had nice Jedi’s too.
Thanks for the look – I oppose this type of delivery on most fronts, but I see it coming and don’t really mind. If it works, reliably, and I can still sell my license to the game/second hand, I’m happy.
You can’t cut out used games, it will never happen, so these services need to start considering ownership models for their consumers, because we will resist it once the people who profit in the used game industry start pointing it out.
yep my bad on the EA GTA reference, removed that. Still, yes, very awesome to see it running on a Mac.
suprise suprise… MG with wrong facts in his article…. ohhh better tweet this one ….
@Envy
I’m just curious what contribution snide remarks make towards the topic at hand. I’m sure we can all do without.
I will always snide MG, I hate his addition to techcrunch, and will protest by being a jerk in my comments to his postings…
I hope there is a Linux client as well. Then I would be able to switch completely
All you need to play this on Linux is a web browser like Firefox, Kmeleon, Konqueror, etc.
The real performance test would be trying out highly computation intensive, interactive Volume Rendering, through 3D Texture Mapping and Hardware Acceleration.
its like this onlive wanna be like the other one on http://onlivedream.com
as a serious gamer who spends plenty on his pc… this is awesome. Im sure employeers are going to love it
Of course GTA4 isn’t a hit… or an EA game. Good proof read, tards.
Yeah, only >13 million copies sold. Massive flop!
Sweet! Time to dump the ATI/Nvidia stock!
Why?? This is brilliant and cool, but no way this can scale. Think of the rendering power required to run one person’s machine. You can only have 1 server per user. How can they maintain 1,000 simultaneous connections without going bankrupt??
The article says they can support a min of 10 and max of 100 users per gpu depending on the game.
Sounds very scalable to me! These guys must be ridiculous wizards to come up with this insane engineering.
Depending on the game being the keyword. No way they can run GTA IV for more than 1-2 users at a time.
Unless they are streaming wolfenstein, they’d need a cray computer to live stream 100 high def users a modern platform quality game.
Business model is not scalable.
Cray computer?? Lol. You are attempting to talk about a subject you have no knowledge of.
Uh, no. Who do you think is going to be selling OTOY all their hardware?
Pretty amazing.
Would be really interested in the overheads to provide this service.
Problem is that unless there’s a huge financial reason for doing it instead of buying the game (which is unlikely for numerous reasons), PC/Xbox/PS3 users will stick with their existing experience (which will always be better or as good).
Phones is an interesting one, but fundamentally you can’t play a PC game on a mobile device unless it’s a very simple game. If it is, this would be cool because it gets around the different device issue.
Cool technology. Maybe ahead of the infrastructure at this point in time though.
Cool, but even on the demo video you can see a delay between the player moving the controller stick and the actual display.
So as much as this is cool, and definitely something that needs to be developed, in this stage no self-respecting gamer will touch it with a stick.
I actually got to see this in person at that exact location where the guy is playing. I think the footage or youtube isn’t catching the really fast frame rate.
This is truly spectacular technology and will advance today’s web experience by years.
I’m excited to see this technology grow. I hope it works for the masses.
Technology wise; is this something similar to what idsoftware has done with Quake 3? http://www.quakelive.com
it’s completely different
The numbers of Q3 are crunched in your PC… The numbers here are crunched in the cloud and you merely get a streaming video (that you can “control” on the fly)
OnLive (http://www.onlive.com/) is much further along with this tech. You guys should test theirs since they have sound, games etc.
I read on Venturebeat that Onlive’s technology cannot scale like these guys are able to.
Don’t believe it. Why not???
It’s the same thing really!
the engineering has to be different if what they say about the results are true. this is a whole new frontier…..who knows
I can understand how this works under Flash, which can render high-definition streaming video. Ditto for Java, which it also supports. But the commentary in the video about switching to HTML/AJAX mode perplexed me — anyone have any idea how this would work without using Flash or Java?
it wont. you need a type of a video player to play it.
You’re wrong. You only need a browser that supports the HTML5 video tag (and necessary codec, most likely h264 in this case). If you listen to the video, he was told to use Safari. That’s because it supports HTML5 video tags. Go here (in Safari 4) if you don’t believe me that you can play video without a plugin installed:
http://webkit.o...-media-support/
I can watch the video on that page on my iMac (not sure about PCs) and if you look at the source, you can see that it’s not using any plugins but instead uses the HTML5 video tag.
YouTube also has a HTML5 demo page here:
http://www.youtube.com/html5
Works fine in Safari 4 on a Mac. It doesn’t look like the video tag is supported in Safari on the PC yet.
Love it, I’m eager to learn more about the scalability and what type of customization goes into getting a game setup and working for this model.
Is this legal?
ISPs are going to love it. Video was already killing their networks, this is even worse : you cannot imagine any caching strategy, of any kind.
That said, this is really neat. Server side rendering allows all kind of new apps. (We are using the same kind of techniques here at stupeflix.com)
” you cannot imagine any caching strategy, of any kind.”
If a large ISP wanted, they could probably make a deal to place some servers in their own network. I bet they could make a lot of money by leasing their technology to others.
It would be a win-win for everyone. The bandwidth would be “free”, so the profits even greater. OTOY could split the income with the ISP and both parties would make money.
Yes, you are right.
That’s the trick dynamic/interactive IP TV providers are already using.
But it is far more work than just running your own datacenter, and the machines are still quite specific with dedicated hardware, large energy consumption etc.
Why not, it may not be that hard.
Otoy in yo mouth!
What a redundant idea.
lag. licencing issues. broadband requirement. frames/sec. not to mention occasional connection interruption right at the cliff’s edge.
and above all.. WHAT NEED DOES IT SOLVE? Where is the problem? Don’t we all have an adequate GPU? Don’t we all have strong enough PCs? and if we don’t, we really have killer broadband?
Really? game rentals? (throw in micro-payments? mmorpg in the browser? ajax? Adobe AIR? any other buzzwords) … So you mean instead of selling the game on a CD for $60 in store and be done with it, a gaming company will need to buy and host strong servers, pay for electricity, cooling, maintenance, and gazillions for bandwidth serving those games to the users for hours straight straight? Ye! I can see the business model! Real economic winner!
Oh. But it uses the “cloud”.. oh.. so its a brilliant idea. So web2010.3.0
Next.. a company developing streaming solution for Tivo.. so instead of buying a Tivo, you will use a Tivo on the cloud and stream it home. No sound available at the moment.. but what the heck.. it uses the cloud!
(don’t copy the idea! its MINE!)
- business models : in game advertising, virtual goods, premium access, sell new levels, etc… The games will become brands and evolve constantly. No more GTA 1,2,3,4 but GTA.com
- Using the could is not a new idea per se, but who’s actually able to develop this technology?
- powerful PC : I only have a good two years old laptop and I can’t play real games on it.. This is true for about every non tech people I know. The era of powerful boxes is over, people are buying ultra light netbooks, smartphones, and be able to play complex, beautiful games.
- “Tivo on the clouds” – please check hulu.com and other small startups like youtube
in summary, you’re an idiot
absolutely amazing. i really believe that this is the future of gaming. cant wait!
can’t wait? you mean for the next frame to arrive?
I’m with Ron. It’s cool but I just don’t see what need this technology solves, and it’s rife with potential technical problems and added cost.
The game industry isn’t like Web 2.0 – they’ve already figured out how to make money by selling shiny discs at $50 a copy. Sure, they may be upset with Gamestop hogging the secondhand market, but that’s small beans compared to the proposed infrastructure needed to deliver a ‘cloud rendered’ gaming experience with any decent graphics.
How did Otoy get rights to the Valve and Rockstar IP to even do this demo? Their EULA’s would seem to preclude streaming a PC version of their games for this type of use. They certainly do not allow hosting or rental use.
How would you know what Valve and Rockstar allows? Do you think Otoy would invest huge sums of money without checking first? The EULA is for end users, and Valve is free to give exceptions to whoever they want.
Very, Very cool.
This is just damn HOT! o0
i agree with RON this technology sounds like its going to eat through my monthly cap.
Thou i can see ISP’s etc trying to entice users to there service by offering this technology as a subsidised service. If ISP’s get onboard and give this bandwidth free so that they can convince users to use there service Ill definetely consider it.
BUT i love my console and I love knowing its just there for me to switch on even without an internet connection. Looking forward to seeing where this goes BUT it aint replacing my 360
Urbach (of Otoy) is a wizard. Last I heard this stuff will scale ridiculously.
Most games have under 10gigs of unique content (models/textures/maps/video/audio/music). Decompress that and maybe 20gigs.
An individual gaming session will add some transient data (like: where your character is, where the bad guys are, how much health and ammo you have, and maybe some environmental destruction, etc.), but since most game consoles have under 512 megs of ram — and most of that is used for storing the current maps and so on — transient state is usually tiny, like 40-100megs at the biggest, typically a lot less.
High-end servers (given partnership with AMD, assume an 8-way quad-core opteron box, or something similar) can easily carry anywhere from 32-256 gigs of ram.
That’s enough to store the entire game contents — every model, map, sound, music, movie, texture, etc. — decompressed in ram, plus transient state for hundreds of players (and possibly thousands — depends on the game).
Only real issue is: is it possible to render scenes fast enough for that many players on a single box?
Answer appears to be yes: grapevine says otoy is converting the polygon models to voxels, then using gpu-assisted raytracing to render; done right, this strategy scales far higher than you’d expect if you’re coming from a rasterization background, and at the scale in question raytracing’s attributes work in its favor. Also keep in mind that if you’ve got 20gigs of game data and 100+gigs of ram you can store a lot of helper structures — spatial indexes, lookups, etc — and also keep in mind that amd-style hypertransport means the gpus can be given insane memory bandwidth.
End product will be much like ec2, but for games: a publisher has to make some tweaks so their game runs smoothly on the infrastructure (though keep in mind, otoy probably has an opengl implementation that can do an ok job with pc games not specifically ported to otoy infrastructure, at the expense of not scaling as well), and then they can start up as many instances of their game as they want (ramping up/down as demand increases/decreases).
If there’s any kind of workable persistence option mmos could become commodified-and-improved in short order: imagine a universe with better-than-wow graphics accessible on any modern web browser.
Whether non-gamers will sign up seems debatable — and whether gamers will sign up also seems debatable — but if this fails it’s not going to be for technical reasons; it’ll fail b/c no one’s needing what they made.
nice comment
Insightful comment, thanks!
“Answer appears to be yes: grapevine says otoy is converting the polygon models to voxels, then using gpu-assisted raytracing to render”
Even assuming they had a rendering method far more efficient than anything anyone else in the world has been able to come up with (there is no “economy of scale” for rendering different viewpoints of the same game world, it would simply have to be faster) they could not just fit it in at the D3D/OpenGL level. A game engine doesn’t pass “polygon models” to D3D/OpenGL, that level of abstraction is only present inside the game itself … the representation of the game world is also completely different for almost every game. To retrofit a different rendering method the entire rendering engine would have to be rewritten, for every game.
Your grapevine is passing bullshit.
woah, I disagree strongly with almost everything you have to say. I am going to examine your post paragraph-by-paragraph, pointing out what I believe to be technical inaccuracies, gross simplifications and ignorance of the server market.
“An individual gaming session will add some transient data (like: where your character is, where the bad guys are, how much health and ammo you have, and maybe some environmental destruction, etc.),”
The context of a typical PC game is 1-2 GB of RAM per instance. You cannot simply share data between game instances, as the code does not support it. It may well be that the same texture and level data being replicated heavily across instances, but there is no way to avoid it without a major re-engineering of software that developers have no financial motivation to do.
“but since most game consoles have under 512 megs of ram — and most of that is used for storing the current maps and so on — transient state is usually tiny, like 40-100megs at the biggest, typically a lot less.”
PCs, however, do not run the same binaries as consoles. Console code is designed around a trade-off of RAM capacity for bandwidth. While consoles do not have gigabytes of DDR RAM, they did have significantly higher bandwidth RAM than the desktop PCs that existed when they were launched. An Xbox 360, for example, has 22.4 GB/s of bandwidth from its GDDR3 alone (not counting the embedded DRAM used for framebuffer and some postprocessing effects). A Core i7 with triple-channel DDR3 can offer similar bandwidth today, but PC ports are not programmed with that target as the norm. They instead make programming trade-offs in the opposite direction, hence the 1-2 GB of RAM per game instance I cited before. Future native PC games will no doubt aim higher, as it not uncommon for PC enthusiasts to have 4-12 GB of RAM.
“High-end servers (given partnership with AMD, assume an 8-way quad-core opteron box, or something similar) can easily carry anywhere from 32-256 gigs of ram.”
Cost of cores and RAM does not rise linearly in the server market. A four-processor motherboard in the server market typically costs over $1000 while a single processor desktop board can be had for less than $100. Higher density RAM needed to reach the 32+ GB range scales in price well beyond linearly as well. When you consider rackmounts, vendor mark-up and the like you are talking systems well over $10,000, for which you can have 10 high-end gaming desktops or more. Furthermore quad and higher socket motherboards typically only have one 16 lane PCIe socket for a single video card. There is a reason that small render farms who are cost sensitive typically use multiple single-socket boards rather than going multi-socket and high spatial density like enterprise.
“That’s enough to store the entire game contents — every model, map, sound, music, movie, texture, etc. — decompressed in ram, plus transient state for hundreds of players (and possibly thousands — depends on the game).”
So what. Context is not shared between instances. Game clients aren’t written that way. You would only improve load times by stuffing that data on a large RAM disk.
“Only real issue is: is it possible to render scenes fast enough for that many players on a single box?”
Not for modern games. No. You need something like the power of an entire Radeon 4770 to get 60 fps on at 720p on a modern game like Call of Duty: World at War. CPU needs are typically not as great as GPU needs for modern games, but by default every modern engine is running at least 1 CPU core at 100% to assist rendering. A handful of games do require significantly more CPU power — Supreme Commander can tax even a modern Quad core CPU.
“Answer appears to be yes: grapevine says otoy is converting the polygon models to voxels, then using gpu-assisted raytracing to render; done right, this strategy scales far higher than you’d expect if you’re coming from a rasterization background, and at the scale in question raytracing’s attributes work in its favor. Also keep in mind that if you’ve got 20gigs of game data and 100+gigs of ram you can store a lot of helper structures — spatial indexes, lookups, etc — and also keep in mind that amd-style hypertransport means the gpus can be given insane memory bandwidth.”
Utter balderdash. All these companies are doing is a latency cheap compression algorithm like Motion-JPEG or modified MP4 without CABAC on 720p video and streaming it, while getting back control data from the client. Ray tracing is only beneficial for performance when the number of polygons grow extremely large (think more than one polygon per pixel); modern games do not have such extremely large polygon counts. Nobody is intercepting DirectX draw calls to convert polygons to voxels (or do anything else other than profile). Nobody. If you were in a close partnership with Microsoft, maybe, but the software engineering necessary is well beyond a small start-up. Also, don’t forget your “insane memory bandwidth” has to be shared by every client on your box. Nobody is going to be doing software rendering, anyway, as high-end GPUs are 30x or more faster than the fastest CPUs at rendering. The “insane memory bandwidth” from NUMA on a quad-core hypertransport system is tiny compared to the 141 GB/s a single GeForce GTX 280 video card can give you for rendering.
“End product will be much like ec2, but for games: a publisher has to make some tweaks so their game runs smoothly on the infrastructure (though keep in mind, otoy probably has an opengl implementation that can do an ok job with pc games not specifically ported to otoy infrastructure, at the expense of not scaling as well), and then they can start up as many instances of their game as they want (ramping up/down as demand increases/decreases).”
First, few games even feature an OpenGL rendering path these days! In addition to magically sharing resources, do you also claim OTOY is converting the DirectX command stream to OpenGL? If this could be done by a small company, all those DirectX 10 for Windows XP projects would not have come to naught. Second, console game developers and publishers are already complaining that development costs have grown too large this generation, and the biggest PC developers are going multi-platform with bogus complaints of piracy to cover lack of sales. No one is going to want to spend even more development effort on an unproven market (except Nintendo who has a good idea of what part of the gaming market lies untapped). The question is, frankly, “Who needs OTOY (or any similar service)?” If you have a legitimate interest in playing games, why don’t you have the hardware already? If you have the hardware, what is stopping you from buying the software? If the answer is lack of money, then a $50/month service on top of broadband costs is not going to attract customers. If the answer is lack of technical know-how, consoles already exist to fill that niche. If you can use a browser, you can use a console.
I suppose OTOY could be marketed as a game rental service (as PC games cannot be rented in most of the world because of piracy concerns, and many console gamers only own one console and therefore cannot play exclusives to other consoles). This might work if OTOY could get a huge installed base of content on their service (which, of course, would largely consist of titles from publishers’ back catalogues that will never see any re-engineering to be more resource friendly). Try-before-you-buy may not be popular with some game developers, however, as many games have experiences that can be measured in single digit hours. What is your incentive to buy a game, if you can beat it in a weekend on a game rental service?
“If there’s any kind of workable persistence option mmos could become commodified-and-improved in short order: imagine a universe with better-than-wow graphics accessible on any modern web browser.”
The only technical advantage I can see for MMOs would be if they could be hosted physically at the same site as the game server. This would allow Ethernet-grade reliability and speed of connect from client to server. This could lead to better MMO experiences, but I don’t see how there would be any graphical improvement. Graphics are almost entirely dependent on the client, and whether they are high-end or scalable is only a function of the designs seen to be marketable. Age of Conan and WoW look remarkably different by design and target market, not because of bandwidth reasons. I am skeptical that OTOY would cannibalize the large segment of the MMO market with mid-range or lower graphics cards. Are you going to want to pay MMO fees, plus OTOY fees plus broadband fees, when you can’t afford a decent gaming PC in the first place? The answer is likely no, you are already paying as much as you are willing to pay to game.
“Whether non-gamers will sign up seems debatable — and whether gamers will sign up also seems debatable — but if this fails it’s not going to be for technical reasons; it’ll fail b/c no one’s needing what they made.”
You can’t separate technical issues and business decisions so easily. What is technologically possibly is largely influenced by how much money is available to make it happen. If there is no market, there is no capital, and if there is no capital there is only what can be done cheaply with existing technology.
Does anyone remember the Phantom console? It was a technologically mediocre product with an ill-defined target market. There is little to make me think OTOY or any similar service will be any more successful.
No thanks! I have absolutely no problem playing with my precious ps3. And is it just me or does that guy suck at L4D?
Like the guide mentions in the video..this will not work well for FPSes (250ms delay)…LOL (and they haven’t added sound yet). FPSes are the genre that needs the most computing power. I remember in Quake 3 being hacks that people could jump further if they increased their FPS by faster GPU/lower settings.
Games like Warcraft, single player GTA 4 could be useful.
Have you ever tried the PC version of GTA 4? It absolutely requires a quad-core CPU to get more than 20 frames per second.
Granted, that’s because it’s a crappy port which doesn’t rely on the video card at all, but still, FPSes are not always the genre which requires the most computing power.
WTF is it with Google Servers in Loading the Video.. The Wait Time SUCKS BIG TIME!!!!! WTF GOOGLE MORE POWER!!!!
Ok, so why couldn’t this have been a connection to some optimized VNC server running the game? Wouldn’t that essentially be the same thing? Sure you could optimize it so multiple game instances on one server share common data and what not, but it looks like VNC to me.
The hardcore gamers play games for the experience. If you have lag in a fps, they are not gonna play it. If you have lag in the US, think abou the rest of the world. Australia, parts of Europe, India, China – these guys will have no option till quite a few years. And in the US, the ones who can shell out money for a broadband connection AND play games like these will not need to play it over the cloud. They’ll play it on their consoles. Till 5-6 years, this cannot become a reality. With a revolutionary technology, you’ve got to give time to the masses and the infrastructure to adapt. We have to have a middle technology for people to jump over, until cloud gaming removes the lag completely – which will never be possible. Ask any gamer – any real gamer – he’ll walk away from a game the moment he feels even a little bit of lag. The reason we still stick it out on mmorpgs is coz there’s no alternative. But with games which you can play on the console – why play on the browser. And the people who want games for free – and dont mind the lag – are not gonna pay for these games – so Otoy’s partners will have to do with advertising. If they can scale efficiently with their ad revenue model, they might survive. But I dont see this thing taking over this world.
I dont think its a lame idea to try to make money, but Id still rather play on my rig. There arent enough good games out for me to even consider paying more towards my FiOS bill.
How can people do this?!?!? Are we humans incredible or what?!
Great, now we can simulate murdering people and be desensitized to violence even more efficiently.
Ok, there is no way this is not going to be less than 100 ms of lag. That makes for really bad gameplay.
Looks cool on videos.
To everyone who is instantly writing this off and doubting the technology: You’ve seen TWO demo videos for a still in beta product. Why not give it a chance instead of pretending you are an expert on the subject at hand?
“There is no way…”
How the hell do you know?
Well, they’re testing it on the Fusion Render platform, which they developed it with AMD. Its not a question of being in beta. The lag is there due to the quality of the stream. If they compromise on the quality, they lose somewhat their credibility since their objective was to provide cinematic quality rendering. If they dont, they keep the 100 ms lag. And that is pretty substantial. If they still have that much lag with a one petaflop GPU-based supercomputer, there’s not much you can expect out of this.
Look at the main demographic who’s gonna use this: the facebook crowd. They still play shit games despite there being better alternatives outside of fb. But they want to be where everyone is, and still have a little fun. If you ask them to pay, they’re gonna laugh at you. If you tell them you’ll stream movies at $10 a month, they’ll still laugh. They can download them pirated for free overnight anyways. The only viable thing is advertising, which will *not*, imho scale that well, since they are building a very expensive server. They need to have the numbers to make a profit out of this.
The sun is shining. Somewhere behind the cloud.
It is interesting tech with some challenges. But how about getting MG a nicer camera, less blurry and stable
This technology looks very nice.
One feature that’d be nice, is that this stuff could make possible to access your games anywhere.
I guess lots of people these days are talking about cloud stuff and at the same time, like games, talking about some sort of “store” stuff too.. I see the future of cloud and its evolving toward “cloud store”.. so anyone can purchase and enjoy games, contents, services, etc. With app store? it’s just an open market place and you need to “download” things anyway..
Goodbye proprietary systems… Soon games will run on hardware ideally suited for them, while being enjoyed by people using any possible platform. I may be able to play TF2 on my mac or even on my (future) crunchpad.
Does this stop here? Of course not. It should be relatively easy to use this for just about any program that isn’t available in the cloud yet. Think photoshop, illustrator, …
And this means a new way of paying for software. Pay-per-use, subscriptions… That removes the need for hundred dollar licences. It also removes one of the greatest reasons for piracy.
Really looking forward to this.
It doesn’t just remove the reason for piracy it prevents piracy. All the other benefits are secondary to the big game developers.
This is freaking AMAZING.
Thanks for sharing!
Seems the technology is a step ahead of OnLive. Hope they have a similar business model (I recall it kinda rocked).
If they made a device that hooked to the enternet, hooked to the tv, had a 360 like controller, and booted into the sites main page (a small thing, like the tvix device from korea I use for my media http://www.tvix.co.kr/Eng/)... I would buy this over a game system…
Nice technology but one problem: In europe we have download limits. For example right now i got only 25 GIGs per month. How much will this new technology be using anyway??
@tom: In Europe we also have flatrates – and if there is a cool service worth it, you get one.
I think this really makes sense for all the console manufacturers: Sell one cheap box with an embedded piece of software that plays all your games forever and do the rest online. Never worry about user base having to upgrade to your new (expensive) system, having to support old systems etc.
It also makes sense for consumers – I can play my games, wherever I want. I have a flat and two houses – when I start playing a game at one place, I can´t continue at the other. Profiles are hassle, multiple copies, too. So this is a cool thing.
Plus you could even offer a game flatrate, giving you a huge library, without the costs of having to do expensive hardware upgrades for specific games.
Plus (even if it´s unlikely) they could also offer you to play ALL games (no system exclusives, no need to own all nextgen consoles), although that´s unlikely as there will probably multiple services that want to distinguish themselves with different game profiles etc. Still you need only one piece of hardware – that´s cool!
One bad thing for hardcore gamers: no internet, no play. But that´s like an MMORPG.