Announcement

Collapse
No announcement yet.

It's 2021 & The FSF Is Still Endorsing 802.11n WiFi Hardware

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #71
    Originally posted by storkus View Post
    I usually lurk and just quick browse, but this has really irked me to the point I had to log in & comment:

    While the vast majority of you including Larabel himself are concentrating on the RF parts, I want to point you to the firmware parts, specifically there are no blobs. With all the router exploits running in the wild, your initial firewall is your first line of defense against the outside world. (I mean, besides running winshit, of course.) Being able to run 100% open software (ideally on open hardware, but hopefully that'll be in the next decade or less) is a huge deal in this application. I realize it may not be the fastest, but as long as it's the most secure, that's what matters. All you need is one blob with negative ring (above hypervisor) access with an externally-accessable vulnerability and you're done!

    PLEASE think about this WRT RYF.
    The question then is, how do we balance legal requirement that stuck Wifi-5's or Wifi-6's software defined radio in closed source (security to governments) and the goal of open source hardware (security to users)? Will there one day a specially designed wireless router / AP comes out, creating a clean separation and sandbox the part government care into a less-privileged module? That module shall only do the bare minimum to satisfy legal need, and leave the rest in auditable open source code. Or is that the software defined radio cannot be sandboxed and I am too naive?

    Comment


    • #72
      Originally posted by billyswong View Post
      The question then is, how do we balance legal requirement that stuck Wifi-5's or Wifi-6's software defined radio in closed source (security to governments) and the goal of open source hardware (security to users)?
      there is no such legal requirement for most of the world, including the US.

      Comment


      • #73
        Originally posted by hotaru View Post
        theoretically, yes. but in practice, you'll likely only see around 60 Mbps throughput on a device like this.
        you can get low speed on router with slow cpu which can't keep up with decoding of upstream connection, i'm not aware of any issues caused by fast ethernet standard itself

        Comment


        • #74
          Originally posted by pal666 View Post
          you can get low speed on router with slow cpu which can't keep up with decoding of upstream connection, i'm not aware of any issues caused by fast ethernet standard itself
          I said "on a device like this."

          Qualcomm QCA9531 SoC, 650MHz CPU

          Comment


          • #75
            Originally posted by hotaru View Post

            I said "on a device like this."
            this thread of discussion started from complaint about using fast ethernet ports instead of gigabit ethernet ports. cpu wasn't mentioned and on slow cpu you don't need gigabit ethernet

            Comment


            • #76
              Originally posted by fransdb View Post

              Yes, for LAN's a higher speed can be practical. However, I have 11n in my home and have xfer speeds of up to 300 Mb/s which is fast enough for even the most demanding data transfers.
              At any time there are about 8 devices concurrently connected of which six have a bearing on the available bandwidth. They still have uninterrupted video, data transfers. However, the only limiting factor is the maximum of 110 Mb/s WAN connection (VDSL) I have.

              As about interference, the internal WiFi router and repeaters have been configured to use auto configuration regarding the channels to use. And looking at the screen and seeing around 15 different WiFi signals around the house, that is/was the only option to guarantee some useful throughput (60-85 Mb/s) on the WAN side.
              Run a long speed test on three or four devices at once, you'll se what I mean. Yes for very asymmetric traffic you can get away with it. but if all the devices have a lot to broadcast, the chance of packet collision goes up greatly. For the basic home network, N is still fine, but more advance setups run into issues, even say multiple consoles, or a home office setup for which you'd need to hard wire or have more advance wireless.

              Re: Channels. - For the U.S In the 2.4 GHz range N really only has 2 channels effectively at full bandwidth, 4 at at a 1/2 channel speed. 5G is less congested, but still only 5 channels for N (a further 8 are available but under restricted conditions). Hence why MIMO is so vital on the base station, it allow spatial separation as well as channel separation.

              Comment

              Working...
              X