Yeah. They do tout some super secret patented compression technique that allows it to be possible, but as a general rule, the more you comress the more lossy it can get. I really doubt it's lossless compression, so it likely wouldn't really be true HD (though you may also not notice so much on a TV).

Actually at standard def it woll only require 1.5mb/s while 720 will rquire 5mb/s. I haqve to say it has amazing potential and i could see it replace console and pc gaming. The consoles will just have to be inovated to stay ahead of this like the wii. The article on IGN says that this could allow you to play Crysis on a netbook theoreticaly. Definitely cheaper than spneding money on a high end gaming rig. I could see this actually helping pc gaming by giving people an affordable alternative.

Um...

Ok, I have 1.5mb/s.  It's satellite.  There's more to playing a game than bandwidth.

Playing games at high def on that system will eat gigs an hour.  If you have a cap, you're insta fucked.  Even the guys with 100 gig caps would still be hosed for the month very quickly.  Only uncapped connections would work with this.

Then there's lag, whatever latency you have in standard multiplayer, multiply it.  A lot.  Send and recieve is one thing, this is streaming large quantities at the same time.  That latency you have when playing online is from just a few packets being sent back and forth to update your actions.  This will be streaming the real time generated video created by the central server, to your computer.  You have not only the latency of sending the actions to the server, but the added latency of generating and transmitting the video.

A really, really good connection far in excess of their minimum requirement for the service, with an outstanding <15ms ping response time from their servers.  That's what you're looking for to have it even resemble the quality of play you'd get from having the game on your own drive with a sufficient system to run it well.  I wouldn't be surprised if the average cable connection is well over 100ms latency with this service.

I didn't read the cnet article but I have read a few different articles about this since the news came out yesterday. Like Annatar said in the OP aside from the "what if they go out of business" issue, the concept is sound in every way. It is easier said than done though. Obviously the biggest hurdle for this company will be providing lag free gaming to a user with an average bandwidth. In one article I read they have been working on this super secret compression stuff for 7 years now and they are to the point where they claim you won't be able to tell the difference between an onlive game and a game on a console sitting in front of you. Once they are outside of a controlled environment I'm sure a lot of problems will pop up just like with anything new, but if they can get it to work as advertised I think it will change everything.

Even though it may not seem like it, the Internet is relatively still a brand new technology so you can only assume that lossless compression and lag free connections will be a reality at some point. I really think that as Internet access expands and improves "Cloud" is the future of not just gaming but of computing in general. It may not be a great comparison but I am sure that when people first started to move gaming away from arcades and more into homes a lot of people scoffed at the idea and said it wouldn't work. I also think this could huge for PC gaming since hardware is by far the biggest drawback.  I wonder what the guys at Stardock think about this, it obviously has the potential to affect them in a big, maybe unpredictable way.

Bullface, latency is an insurmountable fact of life, barring a breakthrough in physics.  There will always be lag.  Electromagnetic waves move very, very slowly when you're measuring in the milliseconds and the thousands of miles.  If we figure out some way to ignore the laws of physics and send information at hundreds or thousands of times the speed of light, then we can skip the whole lag thing, till then you're shit out of luck.[/quote]

Thanks for the science lesson. Yes, I understand that technically there is and always will be lag in everything. If you step on a piece of glass there is lag in the time it takes your brain to realize you did it. When I say lag free I mean relatively not literally. Since we aren't playing from Mars, at roughly 186000 miles a second, the speed of light is more than fast enough to send and receive something with relatively (again at the point a human can detect it) no lag. While doing that seamlessly might not be possible now, assuming that communication technology has gone has gone as far as it can is just ignorant. Regardless, I don't really want to argue about the future telecommunications. I'm just excited about Onlive and I am excited by the possibilities of this kind of service.

Worth reading: http://www.gamesindustry.biz/newsletter/weekly_20090326140025

...

OnLive is a brave and interesting idea, and its technology will undoubtedly find applications - I can imagine, for instance, that in future we may tuck away the hot, noisy, bulky PCs which hold our graphics cards and hard drives into a closet somewhere, and play games on thin clients like our televisions or netbook laptops, streamed over the wireless network using an OnLive-style system.

However, even overlooking the technological problems (frankly, broadband networks in most places simply aren't up to the task suggested here, and that's even if the video compression works as well as advertised), OnLive's business model simply doesn't add up. Videogames remain cutting edge, technologically, because the cost of new hardware is spread out among consumers. Pooling that entire cost into a gigantic datacentre, an IT engineering feat beyond anything previously attempted, and then trying to pay for it through an ongoing subscription system is an idea which may sound great - but making it work economically is very, very hard.

...

Although this sounds intresting, even if it works like they say it will, I highly doubt it will spell the end for console/pc gaming.

For one thing, there are plenty people who have lower grade internet connections that won't be able to stream games. I can't even stream a 15-sec video clip, but I will somehow be able to stream an entire game? I don't think so.

And what about if you don't have an internet connection at all? If you are on your laptop in the car or something, how are you supposed to play any games? Or maybe you just live somewhere that doesn't have internet (although that is unlikely).

What about multiplayer? Would their system even be able to handle mutiple players on the same game with minimal lag?

And then, of course, you have the people who prefer to have boxed versions of all of their games. If you're streaming them, will you even be able to archive a copy, like you can with Impulse?

I'd love to see that thing crawl along on my 150K connection. 

Its an interesting concept but I can only see it working locally. Like you said, tuck away the noisy hardware into another room and stream it to the room in your house you play the game. Over the internet "as with TV" is probably a pipe dream. Movies already require ridiculous amounts of bandwidth and there is talk of the internet "running out".

Well, to reduce the assumed bandwidth-problem with OnLive, perhaps with another approach the data compression ratio could be much improved, if they would not stream pixels but (2D)Vertices (only the transformed vertices that are actually shown on the screen) with some additional informations. Textures could be cached locally because they are used for a lot of frames in succession, most vertix-texture-coordinates and color don't change over a view frames so only a short move-vector or 2D-transform matrix has to be send for most vertices for each frame and only the lighting-information would have to come in continously, too (if not some cacheable radiosity lighting is used for the current surrounding with specular highlights calculated clientside), but lighting textures are only greyscale and in its structure much easier to compress then a full color image (same for pixel shader effects if they could not easily created clientside by sending only some parameters). So a typical game image build with vertices could be transfered with much less bandwidth and easily reconstructed at any resolution back again by only using some not-so-state-of-the-art graphicscard. This would need some more hardware clientside compared to OnLive, but not that much more... using some additional coding, maybe you could also transfer ascii-text as some overlay, too - so it can rendered by the client - creating the complete virtual desktop-experience inside some kind of OnLive-Service (thinking stardock desktop...). And if you are only using turnbased strategy-games or slow paced realtime games like sins and desktop applications you don't need super small latency.

so, if there would be millions of people using some OnLive only (games, webbrowser, word processing... all the common applications Joe User ever uses), you wouldn't need a sophisticated OS on clientside anymore, so no more viruses and trojan horses (because the only connection a client would do is to get videodata from some OnLive clone and send encrypted user input, but running nothing else and doing no other connections), you would gain perfect parent-control for content used, no software piracy, stable software (because software would no longer target thousands of different build PCs, but only the server hard- and software that changes only from time to time) so, does this sound like a nice vision...? 

This would essentially kill modding community.