The virtual box is going to suffer a round-trip lag from keyboard through the network to the server then through compression and decompression plus another network trip back again. I can’t see the whole round trip being less than 50 milliseconds.

Lag is OK providing that everyone playing the game has close to equal lag (so it is fair). If you mix players running on the virtual gaming box with people running on real gaming boxes then players are going to get grumpy about the unfairness.

That being said, most hard core gamers buy a very expensive box but probably only use it on average for a few hours a day and it becomes worthless in a couple of years when technology moves forward. Virtual game machines in a “cloud” environment can spread the usage across timezones.

One can envision a future when even the processing is done in the cloud. What do you have at home? Input devices and a screen. That’s it. Your work (or play) is just the same as today except that the computer power is somewhere else. Is this the better solution? Well, it does diminish the current waste through rapid hardware obsolescence. No BSDs. No reboots. No upgrades. Less power consumption. You have better data protection. You have unlimited storage and it doesn’t use multiple drives (Example: if I buy a film, I don’t need the external drive to store it on and the server doesn’t need as many copies as there are clients as one copy is enough.) I see companies (IBM, Apple, Microsoft…) competing on the software and hardware end, selling the fastest, most reliable experience for a basic monthly fee. Third-party software is a module you buy access to. (The end of constant upgrades in pro software would be welcome!) With ultra fast data transmission (fiber optics to the home) it might be something foreseeable. I hope.

Great post, great service. I look forward to see it on the market. One dimension which is seems forgotten when it comes to cloud computing is piracy. Piracy is a a quite big problem for gaming and software, and by putting the software on the cloud you have created a new control mechanism. I would like to see some more technical discussion regarding that.

Great blog! Keep the good work up.

Mathias

Cool, some very thoughtful comments…

Tel, I agree, 50 milleseconds might be as good as we could realistically expect (for today), even assuming almost no loss in hardware before it hits the network. Games might still be very playable at that speed though, but as you say any differences in lag would create an unfair advantage.

Twowan, great examples of advantages of cloud processing. I also like the idea that there isn’t expensive hardware sitting on someone’s desktop that doesn’t get used for most of the day (or night). Great way to put that processing power to use in a cloud setting to optimize it full-time.

Mathias, you’re right, piracy is probably one of the biggest reasons to move over to this model. MMORPGs really benefit from this, so now maybe single player will too. …then you just have to worry about shared user accounts and/or hacked accounts, or accounts that are bought/sold. But those seem much easier to contend with. Wonder if there are also new risks with similar technology applied in a P2P setting though – end users serving up the games they host on their machines to allow others to play them for free – a bit like webex desktop sharing. People are most certainly going to try to hack the system right away.

I was recently introduced to http://quakelive.com which allows you to play quake inside the browser (you have to download a plugin, just like you would download an ActiveX plugin). It’s an addictive game, and the idea behind it is pretty cool as well. It’s currently in Beta if you want to have a try.

David

David, great example. I’m definitely going to give it a shot!