Wrote this article last night. It's really got me thinking about the implications for web apps. If they really have got sub 100ms response times it could revolutionise web app delivery. It solves massive issues surrounding virtual worlds too. All graphics processing is done at the server. Amazing stuff if it really does work.
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.
Hardware seems to get cheaper/better faster than networks do. Also the trend in networking is bigger bands not lower latency. Someday cloud based low latency apps will work, but I doubt this will happen anytime soon.
The real news here is that compression card they have created. The article doesn't go into detail but large scale 1ms compression?!
What it would take to make this work:
A unified, or at least more unified internet topology. Latency occurs at the gateways/modems/MUXes where you are translating from one type of packet to another. Perhaps you could get around enough of that with peering... hmmm Ultimately I just don't see a lot of applications requiring this kind of latency so OnLive will be swimming upstream.
3 comments
[ 3.2 ms ] story [ 16.0 ms ] threadBut 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.
The real news here is that compression card they have created. The article doesn't go into detail but large scale 1ms compression?!
What it would take to make this work: A unified, or at least more unified internet topology. Latency occurs at the gateways/modems/MUXes where you are translating from one type of packet to another. Perhaps you could get around enough of that with peering... hmmm Ultimately I just don't see a lot of applications requiring this kind of latency so OnLive will be swimming upstream.