• RamRabbit@lemmy.world
    link
    fedilink
    English
    arrow-up
    19
    ·
    edit-2
    14 hours ago
    • This assumes latency between one’s current location and the remote location is almost non-existent. It isn’t.
    • This assumes we have fast and available internet all the time. I sure don’t.
    • This assumes we can use the remote computer in every way we use our current computers. No way.
    • This assumes, as you point out, they won’t be greedy once they control everyone’s machines. They will be.
    • This assumes they won’t censor ‘dangerous’ sites on these machines. They will.

    I will happily pay more for freedom from the corporation.

    • Brkdncr@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      11 hours ago

      Modern desktop streaming is quite impressive. 100ms, 5% loss is no problem for most tasks. You don’t even notice it, and as a result your experience can sometimes be better.

      Additionally you can offload some tasks to the local machine where appropriate.

      You dont need to fit every users needs into a thin client setup, but you could fit probably 50% of all users onto one and they wouldn’t know any different. Think of the energy savings. Think of all that plastic that goes into a desktop or laptop that isn’t needed in a virtualized blade chassis. Think of the rolling performance upgrades. Think of never having your hardware go End of Support. Think of the old equipment that ends up properly e-wasted instead of shoved into a dump. Think of the batteries that no longer need to get produced.

      I might play around with this idea and host my own non-profit Desktop as a Service.

    • tal@lemmy.today
      link
      fedilink
      English
      arrow-up
      3
      ·
      edit-2
      12 hours ago

      I will say that, realistically, in terms purely of physical distance, a lot of the world’s population is in a city and probably isn’t too far from a datacenter.

      https://calculatorshub.net/computing/fiber-latency-calculator/

      It’s about five microseconds of latency per kilometer down fiber optics. Ten microseconds for a round-trip.

      I think a larger issue might be bandwidth for some applications. Like, if you want to unicast uncompressed video to every computer user, say, you’re going to need an ungodly amount of bandwidth.

      DisplayPort looks like it’s currently up to 80Gb/sec. Okay, not everyone is currently saturating that, but if you want comparable capability, that’s what you’re going to have to be moving from a datacenter to every user. For video alone. And that’s assuming that they don’t have multiple monitors or something.

      I can believe that it is cheaper to have many computers in a datacenter. I am not sold that any gains will more than offset the cost of the staggering fiber rollout that this would require.

      EDIT: There are situations where it is completely reasonable to use (relatively) thin clients. That’s, well, what a lot of the Web is — browser thin clients accessing software running on remote computers. I’m typing this comment into Eternity before it gets sent to a Lemmy instance on a server in Oregon, much further away than the closest datacenter to me. That works fine.

      But “do a lot of stuff in a browser” isn’t the same thing as “eliminate the PC entirely”.