Gaikai: virtual worlds streamed as video
Gaikai is basically like OnLive, but for 3d MMOs.It uses Flash to stream video from a high-end gaming PC, and captures clicks and keystrokes and sends them back to the server.
The browser “sandbox” may curtail gaming, but it does not limit the streaming of video. As YouTube and a multitude of copycats have shown, streaming videos online in a web browser is fast, efficient and reliable. This is the key to our technology. When you play a game through our service, you are actually watching a video stream. A very high resolution, high quality, stream with stereo sound, but in essence no different to the last video clip you watched.
I guess this is an idea whose time has come — dumb client terminals that just display a picture. Everything old is new again; this is exactly how Prodigy worked back in the day. 🙂
43 Responses to “Gaikai: virtual worlds streamed as video”
Sorry, the comment form is closed at this time.
Gaikai makes a lot more sense, as most 3D MMOs are not twitch-heavy games (usually PvP is the real rule-breaker here.) OnLive was trying to show off action games that require precision control. Those kind of games require a minimal amount of latency, typically found when you’re directly connected to the system.
The whole thing still makes me nervous.
I believe WebTV was in development before Prodigy came online. I can’t find the book on my shelf right now, but I remember the WebTV interview in Founders at Work pretty much talking about how Perlman has been working on OnLive-related technology for a long, long time. (Or maybe I’m thinking of the TiVo interview…) So the skepticism about OnLive that I’ve been reading about seems simply dismissive.
Back in the day, there were LPBs and HPBs…
I feel like it’s an idea that still needs some more time as technology generally far precedes social acceptance, but I feel hopeful for the push this time. It’s sounding like that later kickstart of an ignition that finally turns the engine over in cold weather.
Looks like it’s for “any game without modifications, even if it was never designed for playing in the web browser” – the OnLive concept without the hardware.
I was doing this kind of stuff a couple years ago with VNC and WoW, wondering when it’d take off. I don’t think streaming video alone is the ultimate answer, though. Maybe something along the lines of a UI overlay on a video stream to get around the lag.
Exciting times!
What they aught to do is market this as a software piracy prevention mechanism. If the software never exists on the clients’ machines, it cannot be duplicated.
This subject strikes both sides of the coin with me.
On one side it excites me greatly. The progression of data transfer and compression technologies, bringing us ever closer to my sci-fi dreams.
On the other side, it tends to make me uneasy. I’ve read too many sci-fi nightmares where central control of any large portion of data turns out badly.
Either way. It will be fun to watch unfold.
I really like this one because the quality of youtube videos is often not even the quality you would expect for a game.
But I guess (and hope for them) that if both Gaika and OnLive are on beta now this is not a coincidence and that their strategy must be reasonable enough.
This totally alienates people with download limits which I do not foresee going out of existence for sometime. Further it requires that a user is online to play a game which maybe offline. I do not think games are the correct platform for this technique and the compression mechanisms maybe better suited to streaming video.
“If the software never exists on the clients’ machines, it cannot be duplicated.” Piracy will always find a flaw in a technique, by not having a physical medium you can not also sell games you no longer require or buy these games from third parties.
Games in the Cloud? More like heads in the cloud.
Biggest catch with any of these ideas
is bandwidth.
While it’s nice to dream that everyone is on fiber optic or alone on cable model it really isn’t true. I know just as many people that are still on DSL or live in areas where their cable modem bandwidth is split among a whole neighborhood or apartment.
@luke: Add in the wireless router in-home and drop outs get worse.
All the work done to decentralize the network being undone by the model that business can only be supported through centralized control of distribution. Tear the hierarchies down, they rebuild themselves in a generation. Bonobo blues…
And the line between a PC-Bang and an 80s college lab full of VT220s gets a little blurrier.
The main advantage of such technologies is that it will drive a long sought-after bandwidth increase in ‘final mile’ solutions. The games industry is in such a strong, dominant position right now!
It’s sad when you think that there have been technologies to push much more data down standard telephone lines for years — there’s just not been enough necessity, or market pressure. Perhaps, in the years to come, there will be.
I think almost every technological leap (in the home at least) will now be dictated by game architectures, designers, and development.
As every programmer knows, there’s a tradeoff between CPU usage and memory usage. A programming goal can often either be achieved by using heaps of CPU and very little memory, or heaps of memory and very little CPU, or someplace inbetween.
Likewise, there’s also a tradeoff between bandwidth and local cpu/memory. If you have infinite bandwidth, you don’t need much local cpu/memory, and vice versa.
The question becomes: At which point in the spectrum does the bandwidth vs. local cpu/memory make (business) sense?
Aren’t a lot of MMOs being developed (and some already released??) with this kind of streaming video now?
What kind of cheats can be eliminated with this? I’m sure that macros and scripts wouldn’t be affected, but what about the programs that read things the player isn’t supposed to see, or substitutes art for easier identification/location, etc? Anything else I’m not thinking of in regards to cheating that can be eliminated?
My questions are on the business side.
How in the world are you going to make this a viable business that can scale? How do you for example set up enough back-end compute power to be able to simultaneously render and stream to 10’s or 100’s of thousands of people at once?
What is the consumer demographic you’re trying to reach? People with high speed internet who want to play hot games, and who haven’t the time, knowledge or money to invest a bit in a nice video card? Unless you can make this DIRT CHEAP to the consumer, why in the world wouldn’t they just invest a bit of money into their PC and make it capable of rendering nice 3D?
And who are you competing with? NVidia for the graphics card upgrade market? Dell for the “and for $200 more you can get a machine that plays games really nice” market?
Honestly once you solve the problem of how to cheaply host a massive amount of render capability in your server, you’ve essentially figured out how to make really cheap 3D rendering systems — and you’d make more money selling the tech to cellphone manufacturers and others.
Because Gaikai and OnLive are totally responsible for your ISP’s policies? ISPs are reactive organizations. They’ll follow the money. Policies will change.
If they don’t, that’s between you and your ISP, not Gaikai or OnLive. I mean, do you also hold Netflix and Xbox LIVE responsible for “alienating” you?
Not a problem.
You’re talking about broadband adoption. These technologies are incentives. Broadband is also not necessary to be competitive in online games. Plus, a slice of the broadband market is enough to sustain a startup for years regardless.
Not a problem.
Server farms in every major city around the world.
Not a problem.
Ask that question again, only in the context of consoles versus PCs.
Not a problem.
Wrong question. Right question: who are viable partners?
Not a problem.
I’m with Morgan here. The only possible issues here are technology based ones on the end user’s side of things (or that ‘final mile’ at the very least); latency needs to be low, but if that can be achieved, the system will work just fine.
And most if not all of the potential techonolgy issues will be solved as new networking tech replaces the old, so even if there are problems with things like latency, they’re not ones that we don’t already see the other side of. If you can shake hands across the Atlantic, you can play video games on a server across the country. It’s only a matter of time until the entire network is that consistently fast; once that happens there’s no reason why you wouldn’t do exactly this.
Getting in at the ground floor now, when there is a sizable number of users that already have what they need to achieve this is a damn smart thing.
Because if you could make it work, I wouldn’t need to upgrade my computer ever again.
This will fail though. Case in point: mods.
Since the actual game isn’t stored locally, you can’t add mods. If WoW did this, every single player-created mod would break, because all you are getting on your end is a streamed video. You can’t even do something like parse the chatlog, because there is no separate log. It’s just a video.
You couldn’t also modify the in-game assets, for machinima or aesthetic puroses. UI hacks? Not possible, the game is stored on some cloud server far away.
You also just busted things like security tokens, because you can’t use them-the machine that accesses the wow servers isn’t able to put your specific token in to verify.
Keep in mind also, you are running vent alongside this, so you are adding even more lag to the streaming aspect. There’s a lot of ancilliary issues that they have yet to think of which will hurt this idea.
Sorry for the additional post, but this also struck me. Wouldn’t this make MMO’s impossible to mutlibox? Considering you’d have to stream multiple video images on one pipe to one user?
DBlade, the mods issue is actually fairly solvable outside of a small subset. Just load a set of standard mods onto the basic game installs that you’re streaming, and then leave an open section where players can request mods outside of the offered ones. It’s been a little while since I last dealt with my web host too much, but I’m pretty sure they do almost exactly the same thing with site back end software. People who wish to make mods or test new mods will probably still buy a box set for their own computer. Chances are, however, that it won’t even dent their final user base, which is most likely people with exceptionally portable, but not very powerful laptops.
The security tokens can be an issue, but overall there aren’t many games that actually use them. If the developers partner to some extent with Gaikai on the other hand, they can potentially create security tokens that assist a large cross sections of games rather than only their own. Just make security tokens for the Gaikai service.
Vent may or may not effect this. There may be a large segment of players who use a standard install and vent, but also maintain a Gaikai account so that they can, for instance, play from their laptop while on business or out and about. If it works for anything with a working browser, there is also plenty of reason to believe that people may use it for playing from their iPhone or Wii relying purely on mouse control for games that support it. Another issue, though, is it’s not exactly like your adding latency to your network on top of having the game’s network traffic running. In this case it would be instead of the game’s normal network latency, and for some people with very good connections this may actually open the option of running both the game and Vent instead of closing it.
OnLive has some interesting issues to deal with for many of the games they’ve advertised as working through it. Gaikai on the other hand is in an area where I simply don’t anticipate them failing. There is too large a cross section where it provides a clear advantage, and a great many games where a little bit of additional lag won’t effect play in the slightest. Not to mention if it can run at small enough resolutions for things like the iPhone, it may finally give EVE players their skill change on the go feature they’ve been wanting for so long.
But the game is being run elsewhere. I’m thinking in this model that everything is done serverside, so they would store your saved games and your mods (that don’t violate the TOS of your game).
The client is no longer in the hands of the enemy!
This service is not for every PC-owner. I am sure that anyone could easily admit to that. But going from there to saying that it will fail because of “mods” is a completely different thing altogether.
All of the cases that you mention like heavy modding, machinima and multiboxing seem to lie more on the “hard-core” side of the gamer spectrum. These are likely players that already own a pretty expensive gaming-rig to begin with. I find it hard to believe that these people would settle for anything less than the optimum experience which should still be running the game on your own monster machine.
I find it very unlikely that the “On-live” service will even try to target these people. Instead I believe that they will have much more luck trying to attract an audience that is a little more “casual”. Like for instance people that still want to play today’s PC titles but aren’t willing to pour money into hardware on a regular basis. There should be a big enough market to start there.
Prognostication is always a risky business. But it occurs to me that if OnLine and GaiKai enjoy some modest success (and I’m fairly certain that they will), the cable industry will take notice. If cable acquires one or both of these services, along with the associated technology, they’ve got the resources and infrastructure to take it mainstream big time — delivering video is their specialty. ISPs get on the bandwagon or they render themselves irrelevant. That old buzzword “convergence” gets dug out of the dumpster and polished up.
Shortly thereafter somebody comes up with cheap holography and all of our old-timers’ screen savers are Carrie Fisher leaning over a droid and saying, “Help me, Obi-Wan, you’re my only hope”.
Seriously, this would be huge for people who are heartily sick of wading through registry entries, video drivers, DxDiag, etc. every single time a company releases another poorly-tested title with an overburdened and largely clueless support team.
It may not displace the current model, but it has the potential to expand and suppliment it nicely.
This is the best comment I have read so far about this technology. It was posted on HN about OnLive, but I think it works for this company as well:
It is doubtful that this will perform well on the load that one would need to service the video gamers who would want to use it, read … everyone. Let’s see 140 million PS2’s, 20 million PS3’s, 45 million Wii’s, and right around 27 million xbox 360’s. I’ll be generous and not count the kids still gaming on the original xbox. That makes about 233 million possible customers. Let’s be nice again and suppose that only 1% of those people are interested in worry free, console free, multiplayer high end gaming via the net. I know, the real number is at least an order of magnitude higher, but I’m being kind. That brings us to 2.3 million people hitting these servers a night, all of them pulling down a 720p stream. Nightly bandwidth cost? Now consider the computing horse power needed to generate and then compress all of those streams. Amazon EC2 is not going to get you there.
But let’s suppose all of this works as advertised!
There is no way Microsoft, Sony, and Nintendo would NOT compete against you in that market, the only question is when will they make their entry.
Or maybe the sleeping tigers. NVidia, Intel, and to a lesser extent AMD. They come along and put extensions in their drivers that work only on their clouds. Programmers have to use NVidia’s drivers anyway, they do not need to use this company’s SDK however.
But by far the greatest concern of this company should be Apple. By making their product hardware and server software based, they are putting themselves on terrain that Apple knows FAR better than they. In fact, Apple knows it far better than anyone in the gaming industry. This is exactly the mistake the music industry made. I remember video of a guy from Sony and Universal laughing at the first iPod. What could Apple do? Well they could come along and revolutionize your revolutionary technology. Like, say, building an Apple gaming cloud that only talks to the iPhone, iPod Touch, and the new iTVGameConsole. And vice versa. They could even make an online store to buy games for their cloud. They might call it like . . . I don’t know . . . iTunes Media Store, or something like that.
I am not sanguine about this company’s future. But the future for technology like this is pretty bright. I’ll put my money on an Apple partnership with NVidia.
“If I’d asked my customers what they wanted, they’d have said a faster horse.” — Henry Ford
Adapted: “If I’d asked my customers what they wanted, they’d have said another fantasy MMORPG.” — Raph Koster
Bandwidth is cheap.
Gaikai is led by David Perry. OnLive is led by Steve Perlman. What bothers me most about the so-called critics is that they presume that these entrepreneurs are naive and haven’t ever considered these issues before.
Steve Perlman worked as the principal scientist at Apple earlier in his career, helping lay the foundations for Apple’s entire multimedia business. Most people know him best as the founder of WebTV, but he also created the Catapult modem for consoles and the Moxi DVR set-top box.
@Morgan
I don’t get your point.
Is it your assertion that since Steve Perlman worked for Apple, that Apple won’t come after them?
Or is your material point that Perlman already knows that Apple is coming?
I don’t consider company founders naive. I consider the *investors* naive.
Bandwidth, cheap? Amazon S3 is at $0.10/Gb out for top tier, that’s 18 cents/hour for 5mbits stream out. Of course, the user has to pay to receive, but we’ll pretend they don’t notice that their current cable cap puts them at a 100 gb/month limit which they’ll reach after 55 hours of play. But back to 18c/hour. We then have processor time. EC2 puts the cheapest processor at 10c/hour. Of course, Amazon is getting a cut and you could get cheaper yourself, but I’m figuring you want at least as much profit as amazon, so you are faced with charging a floor of $0.30/hour for your service. Let’s say I’m a light user playing 2 hours a day, that’s 60 hours (oops, there went my cable cap) And a minimum $18/month fee. Of course, we haven’t paid capital costs on all that custom hardware, dealt with peak usage problems, or given any royalties to the actual game developers.
There are also much better targets for this streaming gaming technology. Consider inside a house, I already have a high end machine, but it might be in a different room than my television. If you had a highspeed hardware VNC card with cheap software decoders you can put on every surface, there’s a big demand *within* people’s houses. And the bandwidth inside a house is unmetered and minimum Wireless G, possibly gigabit. You then also have a natural partnership with NVidia or ATI to throw this on the card itself – imagine a video card with an extra plug for the network.
As for publishers “signing up” with them. Well, duh. It is a no brainer – their whole point is that the developer doesn’t have to change anything. With no cost to “sign up” it becomes a non-choice, because there is no loss if OnLive fails.
Games over VNC are quite doable, but you likely want to play RTSes, or just straight strategy games, not FPSes or racing games. Which, incidentally, require a totally different approach to compression as you want nice crisp clear text in stationary parts of the screen.
OTOH, I could be proven wrong. I wasn’t impressed by what I saw at GDC, but I know the consumer cares little for quality if you can get the convenience sufficiently high. I’ve also seriously considered hosting games via VNC, so it is an idea that is within grasp. But I certainly was *not* thinking of twitch games. Or expanding past 100 users.
The instinct might be to compete with these guys, but the smart money is on partnering with them… or acquiring them.
Funny story… about ten years ago I was hired as an actor for a mystery party with a small group of Microsoft employees. I was free to create my own backstory, so I spun a yarn about being recruited for my revolutionary data compression algorithm, “borrrowed” from Navy cryptographers and utilizing trapdoor functions, that I had incorporated into a multiplayer game (Bloodrace 2020, I think I called it).
I wonder if I released a wild meme?
You said Apple knows the terrain better. That’s wrong. Perlman and his team know the terrain better, and Apple will go to him if they want to play on the same field.
Relative to other costs of doing business, bandwidth is cheap. It’s a nonissue really.
@Morgan
“Perlman and his team know the terrain better…”
I am old enough to recall the same argument put forth
during in a famous quarterly report in the in early
2002. Only then it was Sony claiming to have more
experience in the music industry, and more know how
in consumer electronics. Listen, maybe these guys are
as good as you say, but that doesn’t change reality.
The central point of the original post was that Apple
controls the mobile and handheld gaming hardware.
How does Perlman get his game cloud supported on iPhone
and Touch. Without Apple’s approval . . . he doesn’t.
Simple as that. With Apple’s approval . . . he has less
money to pay all of the other bills, because Apple will
charge a pretty penny.
I don’t see any of these guys beating Apple unless some
competitive mobile and handheld alternative hardware is
forthcoming.
Since when? They’re dominant in certain segments, but they’re not in control. They’re not the only game out there. I know Apple fanatics don’t like to hear that, “but that doesn’t change reality.”
Apple isn’t in the games business. They manufacture platforms. There’s no reason to “beat” Apple. As I already mentioned, the right question is “who are viable partners?”
@Morgan
“Apple isn’t in the games business. They manufacture platforms…”
Gaikai isn’t a platform?
OnLive isn’t a platform?
When did Apple last allow a competitor’s platform onto their mobile
and handheld products? Apple’s attitude seems to be “if you want to
build for our platform, then build for our platform. Not Flash.
Not Java. Not .
And Oh yeah . . . use a Mac.”
Also you said:
“There’s no reason to “beat” Apple”
Even if Apple chooses to buy or partner their way into this
market . . . a big if . . . Apple will only buy or partner with
one company. What is the plan for the other company(ies). The
“Apple didn’t choose us…” plan had better entail a plausible
outline of how they are going to compete with Apple. Wouldn’t
you say?
PS – I also disagree with your assessment of the handheld gaming
hardware market. I believe the iPod Touch in concert with
the iTunes App Store offers a compelling package to the consumer.
One that will only grow marketshare in the future. I have not
seen another handheld platform out there that matches its value.
Even in the Asian markets.
One of the things I think all these companies open up is the notion that today’s 10GB – 12GB MMO installs are quickly becoming a thing of the past. The bandwidth is there to support complete remote gameplay and we already know that we can build a MMO client and distribute a fat install so why not marry the two.
A custom client with a cache of some sort that only downloads the assets it needs as it needs them. I believe Unity 3D is taking this approach with the ability to stream assets to their engine from a web server. That’s pretty powerful on many levels. A hybrid model is what I expect to see come out of all this real time streaming talk in the short term.
How about an MMO where the assets couldn’t be parsed offline until the player actually visited the area? Limit the cache size and the player only ever has an up to date snapshot of some subset of the total world geometry.(Incidentally that’s how the web works) There’s lots you can do there to make your game more “Explorer” oriented. There’s lot’s that can be done from the standpoint of players terraforming/modifying the land in some way when the latencies of replicating that change are smaller as would be the case in a cached as you go world. (trench/moat digging, farm planting, road constructions, adding trees, … all under some level of player control.)
@Cleo, we’re a bit off topic here with handhelds but please tell me that you do not believe that today’s dominance of the iPod will translate into continued dominance in the future? WM7, Silverlight, Android and Flash are all coming to handhelds nearest you this year. All with the ease of use you see in the iTunes store. All with more flexibility than the rigidity of the Apple development platform. Apple has repeatedly stated that Flash and Silverlight would never see the light of day on the iPhone. They’re afraid of roll your own App Stores. That stance will have to change for them to survive the coming wave of new devices that took a beating from the iPhone and are intent on not letting it happen again.
Don’t make the junior mistake of assuming today’s success means continued success tomorrow. IE6 made that mistake and look at the market share of IE now. If Apple fails to recognize the coming competition, it too will lose market share in a handheld gaming market that is very much still up for grabs.
No, from my reading of these services, Gaikai and OnLive are mediating technologies that rest between platforms and software. They’re more like routers or hubs, allowing data to be accessed from any location.
The Nintendo DS is sitting with 100 million. Best I can tell with iPhone is a question if it has reached 10 million. I can’t find anything except speculation on slowness re: iPod Touch. So if there is huge value in this as a gaming platform, the players aren’t seeing it.
In related news, this report,
http://www.businessweek.com/technology/content/mar2009/tc20090331_726397.htm
suggests a lot of americans may suddenly find themselves paying $59 for a 40 GB/month cap with a $1/Gb premium above that. 1.8 Gb per hour suddenly means the consumer pays $1.8/hour *before* OnLive sees a dime.
No, it means customers migrate en masse to another broadband option. And in areas where there are no other options, there’ll be a sudden upswing in the incentive to provide them.
The cable companies are uniquely positioned to deliver (and profit from) a video-streamed gaming experience like Gaikai and OnLine promise. Those among them who wake up to this fact are going to make a whole hell of a lot of money.
Speaking of cable companies…
IT tends to make awful business decisions. I know of a defense contractor whose IT division shut off access to all .mil websites for “bandwidth reasons.”
On a related note, check out Glenn Britt’s total compensation for 2007. I guess he can afford to exceed his new bandwidth caps.
Not that easy. Most cable customers are locked into specific provider. I’d have to relocate to switch from Cox. The cable business is flooded with monopolies.
I think Time Warner Cable is about to have a big problem with it’s customers, if those numbers is remotely accurate. But I hope it’s using the relatively low-res compressed values that TV shows actually stream in and not the 8gb high res movie numbers.
The biggest problem here is that there usually isn’t much in the way of broadband competition. Often times there isn’t adequate coverage from more than one provide, which kinda sucks. But the more people try to do things like the above, the better the opportunities for competitors to poach their users, so it’ll probably sort itself in the end.
[…] talk at GDC about OnLive and other services like OnLive got me thinking about our plans for Ages of Athiria. Our goal with AoA is to create a single […]