It is something to always take into consideration and not forget.

  • spaphy@lemmy.ml
    link
    fedilink
    arrow-up
    4
    arrow-down
    1
    ·
    9 months ago

    Wondering how these magicians measure quality of service then, since they collect no juicy data. I find this hard to believe.

    • JDubbleu@programming.dev
      link
      fedilink
      arrow-up
      7
      ·
      9 months ago

      Quality of service is usually only useful with aggregate data which is worthless for prosecuting an individual.

      • spaphy@lemmy.ml
        link
        fedilink
        arrow-up
        1
        arrow-down
        2
        ·
        9 months ago

        That’s not true. We used to collect client and server data both to detect issues and even if it was only in a subset of customers there is just some customer facing QoS issues you wouldn’t find unless you were collecting data, that wouldn’t be found on server side for example. Like let’s say iPhones make an update and you’re doing video streaming, maybe certain video formats would lag when streaming to the player but not on an android or vice versa.

        • JDubbleu@programming.dev
          link
          fedilink
          arrow-up
          4
          arrow-down
          1
          ·
          9 months ago

          Aggregate data doesn’t mean no client side data. It’s possible they’re collecting aggregate level client data too. They could go further and collect data on individuals that is not identifiable or useful to law enforcement in any way. I can think of a few ways to get anonymous usage data that allows you to improve your service while protecting your users. I don’t know their scheme but they clearly don’t need overly invasive forms of analytics as they have a solid service.

          • spaphy@lemmy.ml
            link
            fedilink
            arrow-up
            1
            ·
            9 months ago

            If your data is being collected then are you really private or anonymous? I can think of a lot you can infer simply from metrics in a client, time window of connection and a few metrics. That’s just removed.

              • spaphy@lemmy.ml
                link
                fedilink
                arrow-up
                1
                ·
                9 months ago

                I worked directly for one of the two biggest log and search systems for big data for years and I can tell you that there is always a way to correlate data lol. And the data you don’t have you can always buy to help put the missing pieces together.