Gaikai: Cloud Computing Gameplay That Works? Comments by Richard Leadbetter

This is a really interesting read and I think Gaikai have the right idea. If somebody wants to reach mass adoption the internet is unquestionably the way to do it.

With this technology any device that can run flash in a web browser can play a game. That means anything from your iPhone to your netbook and desktop PC can be streaming content over a reasonable 2 megabit connection.

If the service scales very well theres nothing stopping you from playing a simple game over a 3G connection on your cellphone and then using wi-fi to play some World of Warcraft on your netbook and then finally go home to your faster connection and play a more bandwidth demanding game like NFS: Prostreet.

I think this is the right way to go about the whole cloud computing thing to. It's really quite baffling why OnLive are pushing a solution that they say is cheap but then trying to meet the expectations of another target audience.

These cloud computing services don't have to aim for insane resolutions they just have to be cheap and provide basic graphics on a wide range of platforms accessible anywhere with a internet connection.

Gaikai and OnLive don't even have to bother with HD resolutions they can just use Standard Defintion like Nintendo did with the Wii and as long as their services are accessible and cheap they will sell.

Most people don't give a crap about HD or 5.1 surround they just care about whats cheaper and more convenient. If you ask people to purchase a dedicated piece of hardware when they can get an experience nearly as good through their internet connection for much less money they will ditch the hardware and go for the cheaper option.

We've already seen this with MP3 killing big dedicated hi-fi systems people used to own. Now nobody buys hi-fis as MP3's are just cheaper and they already have a computer. It's not hard to tell consoles will eventually go the same way.

@Kratos1986

Agree, like Mp3 etc, people will accept a drop in quality for accessibility and convenience. Aiming for HD and full surround for a service like this, even if technically possible is overkill right now.

The whole reason PC gamers are adamant that their system is the superior gaming platform is exactly this latency issue. Console games are just SO unresponsive compared to PC games. Those select few console games that let you play with a mouse just highlight this further.

Double buffered vsync to an 8ms monitor produces input latency of up to 8ms. This is unacceptable, so most PC gamers turn it off. Even triple buffering can produce enough lag to be offputting.

A console control system can introduce 100-150ms of input latency. And now you want to add network latency on top of that?

This Gaikai system might work for console games, but PC gamers will not accept these latency figures. People WILL notice, in those games that are designed to take advantage of the PCs responsiveness.

Still not convinced. Streaming flash video is not the ideal way to watch anything, as anybody who tried to watch Wimbledon last night on the BBC iplayer will tell you. Another point is control methods. A lot of these games are designed with console controllers in mind, I don't think that will translate that well to keyboard and mouse.

Add in the inevitable extra lag and I think the quality of the experience will be too low to convince many potential purchasers. Live game streaming will have it's day, it's just not today.

"If you go to Kongregate.com.. there were 61 million plays.. That's 61 million plays that the games industry missed out on."

The implication that flash gaming is some kind of undesirable aberation - not part of the 'games industry' is somewhat telling. It seems like Perry hasn't considered that at least some of those plays might be due to properties of the flash games themselves, rather than the lack of the ability to stream traditional hardcore games down the browser.

For example, the fact that because flash has a much lower technical barrier to entry, it empowers game creators who aren't necessarily technically minded. Writers, artists, kids, dentists... who may have a fresh and interesting take on gaming. Or that its a great prototyping tool which can be used to rapidly explore new gameplay mechanics. Or even that the short form of many flash games makes them ideal for lunchtime gaming sessions.

Oh well. Good luck to him though, like.

this seems actually pretty nice. as a tiral platform to try out games before you buy them, mind you.
since, you know, nowadays demos are pretty scarce.

but as a full fledged gaming experience... eeeeh.

I'd be more interested if they provided also console games, since I lack the proper platform and don't intend to buy it for just a couple of titles.

BTW since all the talk about streaming gameplay. There is an old service like that, yes you need to have your own pc and your own games, but it works just fine, when i had a netbook it was quite cool to play crysis, hl2 and other stuff on it and i was getting around 50-60fps at 640x480 (limit on trial accounts). So you may want to try it out since it's the only aviable working solution now. And latency is pretty good.. but i never got out of our provider's network so connection latency was around 5-6ms. For ingame latency I guess something around 40ms based on how it felt.

So check it out and maybe it will be interesting to analyze or something. And I believe I'll use it when Pandora console will be in my hand =) (As I recall they said thay will make a Linux ARM version some day)

http://streammygame.com/smg/index.php

Well, this service certainly seems a lot more grounded in reality. OnLive's claims are simply ridiculous.

If all PC games came out like this... That would be the end of piracy.

perry seems to make some alid points.
it's not about platform that much, it can be any nintendo game, some shell shaded xbox360 and ps3 games.
it''s a good way to sample games, or to have demo leels
it end piracy, that a huge ting! to bad crysis doesn't work well ;)
but mouse and keyboard input for playing games designed to play with game controllers would be a probblem, it would limit the amount and diversity of games available on the system

I don't do office desk gaming. I need a box with no keyboard and a gamepad on my living room.

While the idea is quite interesting, I'd like to list two points that came to my mind.

First one, the example cited in the article of 61 million people playing flash games. I do understand that it's directed at possible investors reading about the service, but at the same time, as was already noted above in this thread, those 61 million people are not necessarily going to run into open arms of big game publishers. Since to the best of my knowledge, no one is forcing anyone to play flash games, the logical conclusion is that those 61 million people are playing flash games by choice, and not due to lack of means of play something else.
On related note, as was also mentioned above, the controls for console games translate rather poorly into mouse and keyboard setup, as any person emulating console games on pc can tell.

Second issue concerns the payment for playing games using this service. While it sounds well and good in the interview, does anyone think that all, or even most big publishers will suddenly let you play for a free for a bit (10 minutes perhaps? 20?) using this service? And if not, how will it work? The answer is special plugins distributed by specific publishers, like the infamous EA downloader, that will validate if you bought the game before you will be granted access to it. So basically what it really comes down to is access to the games you already bought, but outside your house. Now tell me, how many people really can afford to play 'big' games at work (or while driving!)? They tend to require large amounts of time to get tangible progress (30-45 minutes for 'roleplaying' games, a bit less for driving ones), and that amount of time is just not feasible at any setting other than home or holiday (but if you go for holiday just to play computer games, you can as well cut the middle man and stay at home..).
So the whole 'play your games at work' idea is, to me, a good way to get sacked, and not bringing about the future to you.

Silverveing, kommanderklob:
I think some commenters missed the point about the flash games. The good thing about flash games is that you just have to click and play. Most other games you have to spend alot of money, but or download, install, configure and then play. Making it easy to access games is the point, not that flash games are good or bad... I didn't see a statement on how flash games are bad, just on how people who play flash games miss out on these other games because of the hassle there is in getting them.

Also, Silvervein, what's wrong with a simple login to see if you've payed for a game? An account where you can add your legit CD-KEYS? Much better for the game-companies to weed out keys published online or keygened. There could be simple separation: games you've bought before are ignored, but you can pay to play on the system other games you don't have. It's up to the publisher.
Also, why can't other people who aren't gamers be enticed by a system like this to play at home?

"no plugins" is obviously false since you need flash.

Unfortunately, flash is terrible on anything but Windows (slow, CPU hungry). So once again, you can play these Windows games... if you're on Windows.

@Palmada

I don't think that they tried to come across as provider of easy to install and use games. In case of consoles, you put in a disk and play. In case of pc, you put in a disk and play. In case of the system discussed here, you open a window and play. What's the difference?

As far as I understood, their main point is that they allow people to play games on whatever system they have, wherever they are. So on other words, give publishers additional revenue by providing their games to people who don't play them now (they used flash games as an example). Good luck to them, but I find a small logical error in the assumption they made about vast majority of people who play flash games starting to play other game types, given a chance.

As for the simple login, there is nothing wrong with it of course. The problem is that so far, no company using online cd key validation method went for 'simple login', instead forcing me to install very hard to remove software that in some cases allows them access to some of the data on my computer without my consent. And I do have a problem with that.
Which is why I don't think that if this service goes live 'simple login' will be the way taken by publishers. And if you think about the number of publishers present nowadays, and think that each one will have separate application checking validity of their games... well.

And people who aren't gamers can enjoy games of course. The problem is that, to my knowledge (based on all the people I know or play with online) they are not playing games because they are not into games and prefer other kinds of entertainment. Giving them one more way to use entertainment product they don't want won't necessarily make them change their mind about games.

@dominalien (and the article) Flash for video is no longer required, as long as you have an up to date browser it should support HTML5, which with the new tag means that this could stream the raw video to you without even having to encode it to flv on the way out. So this would negate the need for a plug-in at all. Though he may imply no plug-in on the fact most folks tend ot have Flash anyway.

A console control system can introduce 100-150ms of input latency. And now you want to add network latency on top of that?

I think the point is that the network latency would replace console latency, so you wouldn't see too much difference. These guys sound much less stupid than OnLive; I'd give it a go.

doesn't flash use tcp to stream? what ahppens when packets need to be retransmitted. the lag will suck big time. i predict this will fail.

Let's see:
1. Perry posts a single video with low-res graphics and low frame rate.
2. Perry doesn't allow anyone else who has tried Gaikai to say anything on the record.
3. Perry says he's got publishers and investors all over him, yet not a single one has said they have any involvement.
4. Perry has admitted he has just started testing testing the system in homes at a distance from the server
5. Perry was supposed to give the whole world a chance to try Gaikai at E3, then pulled out at the last minute because of "patent problems", and instead gave demos that no one was allowed to report on except Perry.
6. Perry has not explained how Gaikai's codecs will be distributed with Flash. There is no relationship with Adobe. If Gaikai's codec's a proprietary browser download, how is this different than OnLive's browser download? If there is a backdoor into Flash allowed by Adobe, why can't OnLive use it, too?
7. It seems like every other comment from Perry is dissing OnLive. The guy's hyper-defensive. Hardly what I'd expect from someone confident about his own tech/business.

So, what part of this isn't amateur-hour?

Maybe cloud gaming will work and maybe it won't. I, for one, hope it does work. But if it's going to happen, be real: it's going to take a company with major support and it's going to take tons of testing and debugging in lots of homes. It's not coming from some company that's proud of the fact it hasn't spent years doing in-home testing.

With OnLive, I see real investors, real publisher partnerships and tons of on-the-record articles where press, analysts, publishers and bloggers are playing with OnLive on different kinds of connections, like the Joystiq E3 test on a cable modem.

When Gaikai, I see one video of Dave Perry playing with himself while he fantasizes about investors and publisher partners and reports that it feels good.

"doesn't flash use tcp to stream? what ahppens when packets need to be retransmitted. the lag will suck big time. i predict this will fail."

This is why I'm confused with the term "streaming". I thought streaming was immediately consuming video data straight from the line, whereas Flash is actually downloading the video to memory (although from beginning to end so you can start watching), hence that grey bit in youtube videos which shows the downloaded part you can seek to. Thus flash needs to download the video in its absolute entirety, so a guaranteed data-complete service like TCP is needed. So "flash streaming" is a bit of an oxymoron?

Or has "streaming video" term carried over from the dark ages of Realplayer/WMP plugin streams?

And can anyone shed some light on whether flash or that HTML5 thing can use UDP for video?

As most people will have realised. The idea of streaming a modern game down the line for remote play doesn't actually work in practice. These services simply do not scale given current internet technology. The fact that they seem to be trying to mislead people by comparing their services to flash video streaming is worrying to say the least. Video streaming is a one way service. You are simply watching something that has already been recorded. When playing a game the traffic immediately becomes 2 way. If you check your broadband systems, you'll immediately realise that your 2mbit download speed is significantly fater than your upload speed. In other words you may be able to view a video at 2mbits but, you'll be uploading your control commands at a fraction of that speed. Even then you'll also notice that the video will need to be pre-cached for a few seconds to compensate for the inevitable lag associated with data moving through the various hubs throughout the internet.

These streaming game services are a con pure and simple. Whoever is investing in them needs to go get some impartial technical advice pronto because they are being made fools of!

There's no getting away from the fact that their own demo proves it doesn't work. The only 2 games he plays for more than 2 seconds that require any reactions at all (i.e - the 2 driving games) he's shit at because he's not turning in time and he's not turning in time because his visuals are too far behind the real game and the response to his inputs too slow.

For something non-action, super. For action it won't work and Pro Street is hardly the most twitch driving game.

And I'm on a 20mbit connection and the video of the demo paused twice. Nothing big for watching a demo, death if I was trying to play a game.

I totally agreed with you about OnLive, but the way you've structured this argument is ridiculous. You seem to suddenly believe it is all possible and that Gaikai have all the answers, just because... well I don't know. Because you love David Perry? (He is a legend, but still!)

It is particularly grinding when you point out that it uses the same codec as the Eurogamer Flash video player, but then pretending to not understand how it could possibly work without a plug in. I thought that if they're decoding the same h264 video stream as the Eurogamer player, then practically the same player (made in Flash) would work!

Anyway, you go on to make a clear difference between OnLive and Gaikai's business model and target markets, so I'll stop complaining. Gaikai certainly sound more reserved than OnLive.

These cloud computing articles are badly argued, theres little basis for singing Gaikai's praises and naysaying Onlive.

In the first article you assume OnLive can't do any better than h264, a codec designed with entirely different assumptions in mind, then declare it perfect for Gaikai. How could Onlive possbily be worse than Gaikai at SD for fast motion after 7 years development? Why's the notoriously demanding Flash plugin a better solution? Even if their HD option isnt up to snuff, theres no advantage for Gaikai here.

How are publishers using the Gaikai service better able to anticipate demand? Onlive know how many subscribers they have and the growth rate and user habits are predictable.

What's the issue with Onlive designing its own hardware? It's nowhere near as ambitious as releasing another console platform. It has little impact on the difficulty of porting a game to each, both simply need a few weeks of work to remove irrelevant settings etc.

How's Gaikai "constantly adding datacenters" a good thing when the authors worried about each service's ambition?

I agree that taking on 60fps HD gaming with the broadband problems we all have is really the suicidal route. Giving people another way to play the games they already know at a slight drop in quality is a niche that is about to be neatly plugged by this. Xbox 360 and PS3 have now shown that 60fps HD gaming is at its peak, and anything else that tries to muscle in is going to have to be something extreme indeed. Rather than trying to improve on this, things just need to go sideways.

Another thing is that people (well, most serious players anyway) don't want pretty pictures over gameplay. I'd rather play FFVII these days than the latest flashy NFS game on my Xbox 360. Very rarely do we get the gameplay and the visuals, and when we do, we get fleeced. Publishers make out that their latest offering has 60fps HD whizz bang fluid graphics and when asked is the story any good, they answer - 'but we have 60fps, man!'

I'll be up for this, as I'm just the type that has god knows how many usernames and passwords for trials and games on web browsers, as I get to play and mess around whilst she's watching Eastenders and the like...

Good luck Gaikai, I say...