OnLive Begins Closed Beta

[modeps]

http://evavhost.com/i/news/onlive.png

OnLive was announced (http://evilavatar.com/forums/showthread.php?t=80907) to a very skeptical world many months ago, however today, they have entered into their closed beta test period. Signups for the beta test have been available since the announcement was made (and are still available here (http://www.onlive.com/beta_program.html)) so if you're one of the lucky folks, let us know! Thanks Shacknews (http://www.shacknews.com/onearticle.x/60304).

I still think that the infrastructure for the ole US of A will be a major hurdle.

[Evil Avatar]

*cough* Vaporware *cough*

This thing didn't work correctly under ideal conditions with no server strain and a fiber connection, it will never work in the real world on real connections.

[Isamura]

considering I can't even stream a movie without it stuttering from netflix on my 6mb connection, I'm inclined to agree with you.

[Sarconix]

Well, this isn't teleportation or time travel. We can reasonably expect to see something like OnLive in our lifetimes. Will OnLive itself be the one to do it? Probably not, since the infrastructure just isn't there... yet.

[Deepsleeper]

*cough* Vaporware *cough*


Given that it's going into beta testing I don't think you actually know what that word means.

[Orz]

Given that it's going into beta testing I don't think you actually know what that word means.

Clouds are made of vapour.

[LilAbner]

Oh man! I *must* try the Phantom....er, OnLive....

[92miata]

i think this is a great idea. hopefully cheap also!!!!!

[net7runner]

Actually going into beta? *GASP*

Also, even if this only works in the 20% most densely populated areas of the country, they still have a nice big market to chew on. Doesn't have to work in Nebraska to be successful.

[Anenome]

Well, this isn't teleportation or time travel. We can reasonably expect to see something like OnLive in our lifetimes. Will OnLive itself be the one to do it? Probably not, since the infrastructure just isn't there... yet.
- Look, OnLive is little more than an attempt to make an end-run around piracy and DRM by forcing gamers to rent games rather than buy them, such that the user never gets access to the executable, etc. OnLive is a dumb idea that has absolutely no appeal to the end user, but a lot of appeal to publishing companies for obvious and stupid reasons. Why any gamer would -want- this service to succeed is beyond me, because it is completely anti-gamer, and assumes you're a crook and a pirate.

Games are among the most demanding applications for a computer to run. The thought of streaming a game in real-time... it's not going to work. Latency, hello? And what about memory usage. If the memory space for the game requires 500mb+ to run, then are you going to download 500mb before you play every time? I don't think so. Yeah, let me load up the latest AAA game on OnLive, ah, there's a 4-hour wait while it downloads the game, again? Let's go get dinner.

If streaming a movie ends up choppy, and that is a straight-forward data transfer and decode with no back and forth communication such as in a game, then how do they expect this to work. Even a dozen milliseconds of latency becomes a problem for many game types, especially FPS. We have enough trouble getting an on-site computer to run that smooth!

I don't think anyone who gets in the Beta is lucky if they have to sign something saying they can't talk about the experience.

So then, the only types of games you could do with a system like this are latency-tolerant games, without high demands on the computer of the user and with low memory footprints. We're talking about things like, oh say, Solitaire, Peggle, Plants vs Vegetables, and other sprite based, 2D games. No 3D games! You won't be running Batman: Arkham Asylum on this system, ever!

Which, of course, means that this will not be a system for hardcore gamers at all. I see this as evolving into an education system for schools and children, something like that. Something where you can just hook up a cheap internet portal laptop and run the app in a browser type thing.

As EvilAvatar put it (http://evilavatar.com/forums/showpost.php?p=1733470&postcount=13):
"Low resolution gaming. (480p)
Lag.
Download waits/delays. (Load time)
Lag.
No ownership of the games you pay full price for.
Lag.
Limited selection of mostly 'older' games.
Lag.
Pixelated gaming.

Did I mention LAG?"


Also, even if this only works in the 20% most densely populated areas of the country, they still have a nice big market to chew on. Doesn't have to work in Nebraska to be successful.
- This wouldn't work in Los Angeles even. The kinds of places this could possibly work are places with a very high population density and the consequent massive investment in short-range communications infrastructure, notably, Seoul, South Korea; Tokyo, Japan; and the like.

Services that don't work nationally also don't receive national ad campaigns or national attention--and certainly don't receive development attention from the big producers. I'm going on record to say that Squaresoft will never produce a game for OnLive :P

[gzsfrk]

Beta is beginning? Prepare for mass of incoming fail reports (NDAs notwithstanding).

My guess is this will be a slightly more polished version of the games I can stream via DirecTVs receiver.

[Sarconix]

And what about memory usage. If the memory space for the game requires 500mb+ to run, then are you going to download 500mb before you play every time? I don't think so. Yeah, let me load up the latest AAA game on OnLive, ah, there's a 4-hour wait while it downloads the game, again? Let's go get dinner.

This is not how the system is supposed to work. It is not streaming the game code or assets to your machine. Your machine is just receiving images rendered on the servers, and you are sending back mouse / keyboard / gamepad input. The images are discarded as soon as they are displayed.

Have you ever used VNC? It is sort of like that. Now that I think about it, the hotel game systems probably work like that as well, only with a very local (video) connection. So memory (or load times) should not be an issue with OnLive. Of course, the servers will have to deal with actually running the games, so they need to handle the loading.

Latency is definitely the problem here. Typical streaming video depends on buffering, but buffering only works because your interaction doesn't affect what is streamed. In OnLive's case your interaction does affect it... many times per second.

[brandonjclark]

- Look, OnLive is little more than an attempt to make an end-run around piracy and DRM by forcing gamers to rent games rather than buy them, such that the user never gets access to the executable, etc. OnLive is a dumb idea that has absolutely no appeal to the end user, but a lot of appeal to publishing companies for obvious and stupid reasons. Why any gamer would -want- this service to succeed is beyond me, because it is completely anti-gamer, and assumes you're a crook and a pirate.

You're right, it's not appealing that you could play a game like Crysis on a shitty, integrated graphics computer. Not at all.


Games are among the most demanding applications for a computer to run. The thought of streaming a game in real-time... it's not going to work. Latency, hello? And what about memory usage. If the memory space for the game requires 500mb+ to run, then are you going to download 500mb before you play every time? I don't think so. Yeah, let me load up the latest AAA game on OnLive, ah, there's a 4-hour wait while it downloads the game, again? Let's go get dinner.

Did it ever occur to you that you won't be running the game locally, as the original service details MONTHS ago stated? The game doesn't run on your computer, it runs on a remote server. You are merely streamed the display output and have a TCP-connected session with the remote server to transfer button presses. I expect OnLive's client hardware or client software installation to include some kick-ass compression technology (possibly existing) to handle the video stream.

If streaming a movie ends up choppy, and that is a straight-forward data transfer and decode with no back and forth communication such as in a game, then how do they expect this to work. Even a dozen milliseconds of latency becomes a problem for many game types, especially FPS. We have enough trouble getting an on-site computer to run that smooth!

Oh gawd, are you still blaming your 12ms ping time for getting wasted? That's sooooooo 1999.


I don't think anyone who gets in the Beta is lucky if they have to sign something saying they can't talk about the experience.

It doesn't seem like you think anyone who gets in the beta is lucky at all.


So then, the only types of games you could do with a system like this are latency-tolerant games, without high demands on the computer of the user and with low memory footprints. We're talking about things like, oh say, Solitaire, Peggle, Plants vs Vegetables, and other sprite based, 2D games. No 3D games! You won't be running Batman: Arkham Asylum on this system, ever!

Pretty stupid statement here, as it shows you never read a thing about the system when they discussed the details. 3D games are the drawing power of this system, please google before opening mouth again.


Which, of course, means that this will not be a system for hardcore gamers at all. I see this as evolving into an education system for schools and children, something like that. Something where you can just hook up a cheap internet portal laptop and run the app in a browser type thing.

It's designed to run 3D applications (i.e., games). What kind of educational gaming do you need that can't be done in a browser on a netbook? Again, for the third time, you are showing you don't understand this product.


As EvilAvatar put it (http://evilavatar.com/forums/showpost.php?p=1733470&postcount=13):
"Low resolution gaming. (480p)
Lag.
Download waits/delays. (Load time)
Lag.
No ownership of the games you pay full price for.
Lag.
Limited selection of mostly 'older' games.
Lag.
Pixelated gaming.

Did I mention LAG?"


Yeah, I get it, you suck at CS:S because of lag, I KNOW.


- This wouldn't work in Los Angeles even. The kinds of places this could possibly work are places with a very high population density and the consequent massive investment in short-range communications infrastructure, notably, Seoul, South Korea; Tokyo, Japan; and the like.

Services that don't work nationally also don't receive national ad campaigns or national attention--and certainly don't receive development attention from the big producers. I'm going on record to say that Squaresoft will never produce a game for OnLive :P

Well, that's what this beta is all about. It's about tuning this system for different network types. Christ, talk about jumping the gun.

[BGSS]

brandonjclark, I don't think we can safely rule on the lag issue until we've tried it ourselves. A typical online game like CS isn't bouncing nearly as much data around as a streaming video + everything it normally does. A similar technology exists on the PS3/PSP, allowing your PSP to stream the video feed from your PS3. I attempted this technology using a PSX game, Symphony of the Night. While the game was semi-playable, there is undoubtedly a noticeable lag that affected gameplay. While I'm not saying OnLive is automatically just as bad as the PS3 video streaming, I am saying that isn't obviously not as simple to get perfect as it looks. Remote desktop software is the same way. Every remote desktop software I've ever used has a hint of lag. Whether or not OnLive has addressed this issue and implemented some technology to make it actually viable remains to be seen until we play it ourselves. I'm not saying you're wrong, but I don't think you can safely say you're right either.

Also, it seems to me that there are only two major advantages to this service (with one of them possibly being a disadvantage). Being able to play graphically intensive games, and playing new games for a cheaper cost. If the cost isn't cheaper, or its slightly cheaper but you don't get to keep the game if you unsubscribe, then it's debatable if the value is there. As far as being able to play high horsepower games, you can build a PC that will run Crysis for $600 or less these days. Wouldn't your OnLive subscription end up costing more than that as time goes on? Why not just build a new computer?

I'm not immediately ruling out the usefulness of OnLive, I want to try it first as well as study the pricing structure. However I would say I am more doubtful than positive at this point.

[FreezaSama]

*cough* Vaporware *cough*

This thing didn't work correctly under ideal conditions with no server strain and a fiber connection, it will never work in the real world on real connections.

oh, I agree. But it'll certainly prove to be much more tangible than the Phantom ever was.

[brandonjclark]

brandonjclark, I don't think we can safely rule on the lag issue until we've tried it ourselves. A typical online game like CS isn't bouncing nearly as much data around as a streaming video + everything it normally does. A similar technology exists on the PS3/PSP, allowing your PSP to stream the video feed from your PS3. I attempted this technology using a PSX game, Symphony of the Night. While the game was semi-playable, there is undoubtedly a noticeable lag that affected gameplay. While I'm not saying OnLive is automatically just as bad as the PS3 video streaming, I am saying that isn't obviously not as simple to get perfect as it looks. Remote desktop software is the same way. Every remote desktop software I've ever used has a hint of lag. Whether or not OnLive has addressed this issue and implemented some technology to make it actually viable remains to be seen until we play it ourselves. I'm not saying you're wrong, but I don't think you can safely say you're right either.

Also, it seems to me that there are only two major advantages to this service (with one of them possibly being a disadvantage). Being able to play graphically intensive games, and playing new games for a cheaper cost. If the cost isn't cheaper, or its slightly cheaper but you don't get to keep the game if you unsubscribe, then it's debatable if the value is there. As far as being able to play high horsepower games, you can build a PC that will run Crysis for $600 or less these days. Wouldn't your OnLive subscription end up costing more than that as time goes on? Why not just build a new computer?

I'm not immediately ruling out the usefulness of OnLive, I want to try it first as well as study the pricing structure. However I would say I am more doubtful than positive at this point.

A sensible argument, how refreshing.

I'm not saying I'm "right" or that I think evil lag will detroy this systems hopes. I'm just saying, "Give it a chance to go through beta testing before you actually stamp it suxorz.

[Mr.Green]

I'm sorry but this has fail written all over it. Phantom style.

[Anenome]

Brandon, your post is just full of unreasonable optimism. The claim that a streaming system can be anywhere comparable to local hardware and local running software off local memory, is ludicrous. It's like saying that a thousand PS3's lashed together is the same as a super-computer--that's an ignorant statement. What makes a supercomputer awesome is that all the chips can talk to each other very quickly and easily. In the same way, client-server is not equal to running off local hardware.

Similarly, if you want to run bleeding-edge games, you mentioned Crysis, you need to run it locally. My assumption was that they would try to mitigate this problem by running code locally in memory, you took exception to that, fine. No, I haven't researched it in depth, it's not worthy of more than five minutes thought because it's akin to saying there's gold at the end of the rainbow, when two seconds of thought -at most- tells you that's ridiculous.

A system like this might, might be possible when internet-2 is up and running and available to the masses.

Let's talk input: Think about something like the Wiimote, or say a mouse. The Wiimote captures position data a couple hundred times a second. You think that's gonna work in software being streamed to a server and translated back to you in under 10ms? I think you're fooling yourself because you don't fully understand the limitations of the systems involved.

My position is extremely simple: I'll believe it when I see it. And you shouldn't be happy about a press conference where they admitted the demo was essentially faked.

I'll ask again, in what way does this benefit users? Someone mentioned the egalitarian dream of running bleeding egde games on cheap portal machines. It's a pipe dream. This system only benefits publishers. You can bet the games will cost a lot more in exchange for not needing a powerful system. There's no free lunch, and you've lost a lot of freedom in the meantime.

[Strider]

Hahaha, this is going to be fun, but pretty funny thread here already.

Meanwhile, the tech pro's sit back, grab some popcorn and enjoy!

[Deunnero]

Thank you brandonjclark for adding some common sense to the thread. :D

[brandonjclark]

Brandon, your post is just full of unreasonable optimism.

I'm obviously dealing with a half-empty guy here. Not a good start.

The claim that a streaming system can be anywhere comparable to local hardware and local running software off local memory, is ludicrous.


In what respect? Graphics? Control? Please better define your statement.

It's like saying that a thousand PS3's lashed together is the same as a super-computer--that's an ignorant statement.

You're right, that is an ignorant statement, so why did YOU just make it?

What makes a supercomputer awesome is that all the chips can talk to each other very quickly and easily. In the same way, client-server is not equal to running off local hardware.

Hmmm, they all talk to eachother? I wonder what the topics tend to be? Oh, I know, bits and pieces.... har har



Similarly, if you want to run bleeding-edge games, you mentioned Crysis, you need to run it locally.

You mean like your monitor is local to your CPU? Imagine stretching the VGA/DVI/HDMI cable... ...over a network. There, now you have it.


My assumption was that they would try to mitigate this problem by running code locally in memory, you took exception to that, fine.

At least you're admitting you had no idea and completely assumed. That's step 1 to becoming a half-full kinda guy;)


No, I haven't researched it in depth, it's not worthy of more than five minutes thought because it's akin to saying there's gold at the end of the rainbow, when two seconds of thought -at most- tells you that's ridiculous.

You're right! You know what else is ridiculous? Microwaves, combustion engines, and powered flight, that shit will never work! Need a more modern example? How about KVM.net, user-content created websites or mass broadband deployment.


A system like this might, might be possible when internet-2 is up and running and available to the masses.

Uh-oh, someone (hint* you) is starting to see a glass half full, I think!


Let's talk input: Think about something like the Wiimote, or say a mouse. The Wiimote captures position data a couple hundred times a second. You think that's gonna work in software being streamed to a server and translated back to you in under 10ms? I think you're fooling yourself because you don't fully understand the limitations of the systems involved.


First things first. A couple of hundred times a second, right? Well then, let's do your math since you obviously didn't.

Wii 1000/200 = 5 ms response time needed

OnLive 10ms = 100 input changes/position data per second. Do you think this will be enough? I do!


Secondly, I've ran a datacenter for the last five years straight, and before that I've designed circuit boards and written programs professionally. Don't even get me started on my years of service in the USAF working directly with communications devices between fighter jets nor the fact that I'm certified with Microsoft for Active Directory and have earned two degrees:one in computer science and the other in computer networking. So yes, Anenome, I know exactly what the fuck we're talking about here.


My position is extremely simple: I'll believe it when I see it. And you shouldn't be happy about a press conference where they admitted the demo was essentially faked.


Ugh, sorry buddy. But that's my position. You're position was (just a few paragraphs ago) "it's not worthy of more than five minutes thought because it's akin to saying there's gold at the end of the rainbow". I've been saying just wait until the beta is over, while you've been judging it without actually seeing it happen. Who's ridiculous now?


I'll ask again, in what way does this benefit users? Someone mentioned the egalitarian dream of running bleeding egde games on cheap portal machines. It's a pipe dream. This system only benefits publishers. You can bet the games will cost a lot more in exchange for not needing a powerful system.

I'm not going to repeat myself. Please read my previous posts if you want the answers to those questions.


There's no free lunch, and you've lost a lot of freedom in the meantime.

WTF are you talking about?

[blackzc]

I'm obviously dealing with a half-empty guy here. Not a good start.




In what respect? Graphics? Control? Please better define your statement.



You're right, that is an ignorant statement, so why did YOU just make it?



Hmmm, they all talk to eachother? I wonder what the topics tend to be? Oh, I know, bits and pieces.... har har




You mean like your monitor is local to your CPU? Imagine stretching the VGA/DVI/HDMI cable... ...over a network. There, now you have it.



At least you're admitting you had no idea and completely assumed. That's step 1 to becoming a half-full kinda guy;)



You're right! You know what else is ridiculous? Microwaves, combustion engines, and powered flight, that shit will never work! Need a more modern example? How about KVM.net, user-content created websites or mass broadband deployment.



Uh-oh, someone (hint* you) is starting to see a glass half full, I think!



First things first. A couple of hundred times a second, right? Well then, let's do your math since you obviously didn't.

Wii 1000/200 = 5 ms response time needed

OnLive 10ms = 100 input changes/position data per second. Do you think this will be enough? I do!


Secondly, I've ran a datacenter for the last five years straight, and before that I've designed circuit boards and written programs professionally. Don't even get me started on my years of service in the USAF working directly with communications devices between fighter jets nor the fact that I'm certified with Microsoft for Active Directory and have earned two degrees:one in computer science and the other in computer networking. So yes, Anenome, I know exactly what the fuck we're talking about here.



Ugh, sorry buddy. But that's my position. You're position was (just a few paragraphs ago) "it's not worthy of more than five minutes thought because it's akin to saying there's gold at the end of the rainbow". I've been saying just wait until the beta is over, while you've been judging it without actually seeing it happen. Who's ridiculous now?



I'm not going to repeat myself. Please read my previous posts if you want the answers to those questions.



WTF are you talking about?


OMFG dude, go outside. It must have took you a good 20 mins to do this post.

[Evil Avatar]

Several developers pointed out -- if these people had actually figured out how to make this work, they could make way more money streaming HD video to people than offering a cloud gaming service where they are going to essentially tie up a single machine per user for any game from this decade.

Complete and total waste of time.

[brandonjclark]

OMFG dude, go outside. It must have took you a good 20 mins to do this post.


Oh, at least.... and I enjoyed every minute of it.

[Anenome]

Everything has tradeoffs. Your position so far has been that there will be no tradeoffs. That's ridiculous on the face, thus my 'end of the rainbow' comment. What's worse, none of those tradeoffs are positive for gamers.

All I'm doing is pointing out trade-offs, and you're saying, "Just wait! I'm an expert, just wait!" Well, they admitted they had to fake the demo, so what are we waiting for. They've got bullshots running on their press-demo. I think that's all we need to know.

So I have to ask, do you have a personal stake in the outcome of the beta, are you involved with OnLive in some way? Are you in the beta? Is that why you're sitting here citing expertise and offering no technical explanations, only hope and optimism? I doubt it, but we'll give you a chance to set the record straight.


There's no free lunch, and you've lost a lot of freedom in the meantime.
WTF are you talking about?

You so easily dismiss the main reason why gamers should be against this. Let's say it is all technically feasible and possible, which is clearly your claim, you're still not better off as a gamer. Gaming rigs are only going to get cheaper as time goes on, making the aim of this service less and less viable. Moore's law will see to that.

The main problem is that this service is designed to force you to rent games. It assumes users are pirates. No longer will we own our games. That has many implications, but it is a net loss for gamers. On top of that, the games are not likely to be cheaper while being rented from OnLive. Nor will they likely play as smoothly as if locally run, especially if they don't DL anything to the system as I charitably offered it might do previously.

There are trade-offs for everything. Unless they charge something that seems ridiculously cheap in comparison to owning games and buying a gaming rig, the trade-offs will not balance out. Are they likely to make the games cheap? I doubt it.

I've been saying just wait until the beta is over, while you've been judging it without actually seeing it happen.
- I don't have to see it, they already admitted the demonstration was running off videotape. In short, they admit the system doesn't work right now. That doesn't help your case, chump.

Everyone seems to 'get it' except you, which is why I asked if you have a conflict of interest here. Listen to EvilAvatar on this one. Complete and total waste of time.

When even developers are poo-pooing this idea, where does your unreasonable optimism spring from (if not personal involvement).

I'll say it again. This service is designed to be the ultimate DRM = anti-gamer. This service is designed to be the ultimate money maker for publishers by making resale impossible = anti-gamer. This service will degrade the quality of gameplay for the user = anti-gamer.

The only place to makeup on those scores is for the price of the games to be extremely low for the end-user. Is that likely to happen? No. = anti-gamer.

[Anenome]

Let's talk about what the owners of the company are saying about their own system:

"According to Perlman, OnLive's technology will make it possible to stream the games in such a manner--high quality, no matter what kind of system the user has--by virtue of a series of patented and patent-pending compression technologies.
- Mmm, compression. I forgot to even mention resolution and compression issues previously. That's gonna be just awesome. I love me some artifacting in the morning. That always looks exactly like locally generated frames off my own graphics card, oh yeah. Gonna be fun when they upconvert it to match the local (various) resolutions.

An intended benefit of this infrastructure, Perlman and McGarvey explained, is that users will be able to play streamed games via OnLive with no lag, so long as their Internet connections meet minimum thresholds. For standard-definition play, that would mean a minimum 1.5 Mbps connection, and for high-def, 5 Mbps.
- So, if you're too poor to afford a gaming rig, but you pay out the ass for internet service, you can run this thing at a decent resolution. Guess that destroys the 'this is for poor people who will be able to now run Crysis on a crap system' argument. Better hope sis isn't running any torrents tonight. Gonna suck when you're facing that boss right when dad decides to hop on a funny video site and lag-city results.

That's obviously an essential feature, as it's hard to imagine anyone paying for a service like OnLive, no matter what games are on offer, if the user experience is inadequate. But the company promises that as long as users have the requisite minimum hardware, operating systems, and Internet connections, they should be able to have seamless play.
- Under ideal conditions. Which include forcing your familial relations to not be internet hogs themselves. Good luck with that. Perhaps this service is aimed at loners though, who knows.

And for the nine--to date--third-party publishers who have committed to being involved, McGarvey said, OnLive presents a much more efficient and profitable distribution model than the standard retail structure. That's because the system is all digital, cutting down on physical distribution costs, and because it is designed to eradicate piracy and second-hand sales, both of which are banes of the publishers' existence.
- Mmm, someone mentioned that fact, as I recall.

How do they deal with the cost of providing hours of video? One of the comments on that page:
So many gullible people ...

Rather than being cheaper, delivering gameplay via internet streaming is far more costly to the publisher than a DVD or Blu-Ray. For a game publisher to provide 30 hours of High Definition gaming to the consumer

DVD or Blu-Ray: $1 (print and ship to retail)
Streaming game: $30 (conservatively based on current hosting costs and QoS requirement)

This is an investor scam at best. There's a sucker born every day.


Ah, finally some resolutions being cited:
Another problem is the resolution offered. They are talking about 760p, but most PC games support 1080p and beyond. So, until they can match that, I think that their graphics quality is a limitation.
- I agree. 760p sucks. Yet another tradeoff. I really don't see why Brandon is intent on putting our gaming experience into someone else's hands like this. You lose so many options.

Here's a good challenge for you, B:
agreed. well, my main question is, how can they say there will be " no noticeable lag". you can't even ping most websites and get any less than 100 ms. Add a video stream to that, and it could double. most people don't have an internet connection fast enough to sustain the DL speeds required to stream hd video, even if it is only crappy 720p. I can only see this working for those select few who have fat, uncapped pipes for an internet connection. but, chances are if they have that, they'll have a beast of a PC that can play any game thrown at it without breaking a sweat. Now, my question is if they have developed a video stream that supports 720p and can STREAM though a broadband (1.5 Mbps) how come they haven't licensed it out for other uses? It just seems so unreal of a concept I don't even know how they can claim it will work. They would have to have thousands and thousands of servers to even support under 100,000 clients, not to mention you would have to constantly be upgrading the servers and adding new ones. I'm sure most of you guys have seen the demo video, but there is one thing they didn't explain in that. In the demo, the server running everything was probably in the room behind them, and even if he wasn't playing directly off the server itself, the client PC was likely connected to the server on a gigabit Ethernet network, which equals virtually zero lag.

The only way to have video that doesn't lose quality wh3en compressed is to not compress it. thus, the games played on their service wont look anywhere near what they would if run on the client side. If I spent 500 bucks, I could easily upgrade my PC to be able to play games at resolutions HIGHER than 1080p, without the quality loss of compressed video.

Plus, I bet all those **** ISPs will love the sudden explosion of traffic on their networks.
- So, in other words, if they have invented some revolutionary compression algorithm, they could make millions of dollars developing a movie streaming service with none of the other headaches being worried about here, and they aren't doing it. So, either they're morons, or it's not true. I'm gonna go with option B. But we have these cloud-computing evangelists who want to turn the world on its head but haven't ponied up the goods yet.

By the way, what sort of fps do you imagine we can get? Surely not anything above 60fps. Not sure the shooter fans are gonna like that much.

Here, there's about a hundred similar objections in the comments: http://news.cnet.com/onlive-could-threaten-xbox-ps3-and-wii/
And most of them of a technical nature.

Again, where does your unreasonable optimism spring from?

[Anenome]

Connection Requirements:
Thus, Engadget states that "Broadband connections of 1.5 Mbps dials the image quality down to Wii levels while 4-5 Mbps pipes are required for HD resolution." The average broadband connection speed in the US at the end of 2008 was 3.9 Mbps, while 25% of US broadband connections were rated faster than 5 Mbps.

Streaming older games:
The hardware used is a custom set up consisting of OnLive's proprietary video compression chip as well as standard PC CPU and GPU chips. For older, or lower performance, games such as LEGO Batman, multiple instances can be played on each server using virtualisation technology. However, high-end games such as Crysis Warhead will require one GPU per game. Two video streams are created for each game. One (the live stream) is optimised for gameplay and real-world internet conditions, while the other (the media stream) is a full HD stream that is stored server-side and used for spectators or for gamers to record Brag Clips of their games.

Skepticism:
Soon after its announcement, many game journalists expressed skepticism concerning how the OnLive service would work. These mainly related to the quality of the service in real-world conditions, both in terms of the hardware required in OnLive server centers to render and compress the video, as well as the impact of commercial internet broadband connections on its delivery. During GDC 2009, which was held in San Francisco, the OnLive service was only 50 miles (80 km) from its Santa Clara data center. The closed beta has also only seen "hundreds of users on the system".[17] Near the E3 Conference, which is approximately 350 miles (560 km) away from their data center, OnLive tested their service in a real-world environment with a regular cable modem. Reports say that OnLive performed well.[18][19]

Matt Peckham from PC World stated that it might be technically difficult to transfer the amount of data that a high definition game would require; later on he mentioned that OnLive would need a "deterministic broadband". That is "a guaranteed, non-shared, uninterruptible speed", but "broadband isn't there yet, nor are ISPs willing to offer performance guarantees". He also mentioned concerns about the "mod community" being unable to create and offer mods since all the game data will be stored on the OnLive servers; as well as the fact that any games bought on OnLive are not actually owned by the user. If OnLive were to go under, all the user's games would also disappear with it. Currently, no widespread trial has been made to test the service so it is unclear whether it would work once live.[20][21]

Cevat Yerli, the CEO of Crytek, had researched a method for streaming games but concluded that Crytek's approach would not be viable until 2013 "at earliest".

The most likely reason for something like this to fail or become a niche product only is because consumers simply do not buy it. There are so many negatives, so many trade-offs, that I won't be at all surprised if people prefer the console and ownership model to a streaming model that has so many downsides.

Brandjclark has already shown us that real-time positional input in its most basic form on the Wii is not doable via OnLive. Because it requires a 5ms ping, and that's not gonna happen.

OnLive has been in development for several years now. Truth is, OnLive was broadsided by the Wii, it took gaming in a direction that OnLive is not setup to take advantage of. And, with Sony and MS creating their own positional input solutions for their next consoles, and Nintendo set to vastly improve their positional input--and all three no doubt at the very least thinking about including positional head-tracking, OnLive has a very real problem. They're not going to be able to duplicate that via their system.

200 hertz is just the beginning, we may see 300, 400, even 500 hertz soon. Future positional tracking systems will take that number higher. There are even 'gaming mice' that sample at 1,000 hertz. This is the easiest way for console makers to fight back.

In the long run, the console and the computer will merge. But cloud computing is a step backwards for many applications.

You simply will not be able to play a game being produced remotely as if it were running locally and not notice a difference. Not possible, I don't care what Brandon says or thinks he knows. There will always be tradeoffs; don't let him bullshit you because his brother owns the company or some crap like that.