Economic implications of OnLive

Hey there,

I warn you that this thread will appeal to gamers more than anyone else :stuck_out_tongue:

I stumbled across this today:

GDC Press Conference

More techy and in-depth presentation

If you can’t be bothered to watch it, basically this company is trying to apply cloud computing to gaming. This means that games are played on a server somewhere in a data centre within a radius of 1000 miles. Your computer only needs to have an internet connection, input devices, and a way to display video. You can play Crysis on a low-end mac with such a system. They think they can manage to get the latency below 80 ms, the threshold of human perception, so that there is no perceptible lag between a player selecting an action, and there being a response on the screen. Video games are now to be stored on the servers too, so one needs not acquire physical media or to download it (as with Steam) in order to play. There are no buffer times, etc.

I’m not going to argue about whether this will work or not, as I am not competent enough to know, but assuming it will, this may be a very important precedent in the development of realistic cloud computing for intense applications. Here’s what I see as immediate and eventual consequences: (it’s worth noting this will probably compete with traditional methods for a time, and may never becomes universally appealing)

  • Upgrade costs of computer hardware for PC gamers goes down to zero. All upgrades are done in the data centre. These costs will be much lower than the aggregate of individuals buying their own gaming PCs because of economies of scale and time, e.g. the average gamer underutilizes their gaming platform. They buy a PC that is idle for most of the day, or at least not playing games. With a data centre system, more minutes are being used. When you log off, someone else takes your place. This works because of time zone differences, although it still won’t be perfect. Moreover, the issue of hardware wearing out due to increased use isn’t really there as gaming hardware often becomes obsolete before it breaks.
  • Consoles manufactured by Sony, Microsoft and Nintendo will become obsolete. Not only will platform differences disappear that cause some games to be console exclusive; but the microconsole they advertise will cost a tiny fraction of buying today’s consoles. Again, consumers save money.
  • The retail structure for distributing games largely disappears. Costs are again cut, resources saved, publishers get more of their sales as a percentage and can then invest more of that money into developing new games and technologies.
  • Games will be more marketable that would not even feasibly run on home hardware due to cost constraints of the typical family.
  • More consequences of cloud computing unrelated to games would include the possibility of replacing expensive laptops and PCs with cheap thin clients, thus saving people money.
  • Privacy however may become an issue when we think outside of the scope of games. Nobody needs to hack your computer to access files, they’re all stored elsewhere. The government would probably have priority access to this.
  • You won’t own the games anymore, so there is risk involved. Then again, this is already the case with services such as Steam, and that hasn’t deterred lots of people so far. Plus, I think they intend to allow you to rent games.

So basically: efficiency, lower costs, potential loss of privacy.

Anybody else think of any consequences if cloud computing were to become a reality? Anybody think this is realistic or is interested by the prospect?

Best wishes,

Fred

Here is my take on cloud and thin client computing.

Thin client computing has been promised for ages. And yet hardware requirements, particularly ram and processor multi-threading continue to become greater and greater. Thin client computing is dead, because so much of interactive content is run client side, not server side. For example, Flash, javascript (ajax) is all run in the browser, not rendered on the server side website. And flash, and to a greater degree ajax, are become more and more common, and more and more demanding.

Cloud computing is a way to distribute server costs by only paying for resources on demand. It’s like renting ram and processor power by the minute, metered in some cases. It’s not a way to take computing away from the client side. It’s strictly to make the allocation of resources more flexible and agile on the server side, by clouding many servers together and pooling resources which are then shared on demand.

IMO, this is a pipe dream given current trends. Doesn’t mean someone won’t figure it out, but the promise of thin client computing (which is a Google fantasy too) still hasn’t materialized, simply because the costs server side are too great to deliver media without leveraging the power in the client.

I’m currently on a 390 kbit/s connection (that’s down, not up). If cloud computing ever starts off, I fear that people with low internet bandwidth, such as most people in my rural area, will be cut off entirely from new technological developments.

It’s a nice idea for people on a fast connection, though. Would make a lot of hardware obsolete I guess.