3 ms·

I have around 8mbps (1 mbyte/sec) and downloading everything big is painful. Nvidia driver - 10min. Vbox update - 3min. Linux netinst - feels like forever. Big

by wruza 4y ago

I have around 8mbps (1 mbyte/sec) and downloading everything big is painful. Nvidia driver - 10min. Vbox update - 3min. Linux netinst - feels like forever. Big npm/docker/etc updates - few minutes of waiting. Witcher 3… ohh. Also you can’t watch 1080p without stutters while waiting. You can’t watch 1440p60 or 4k in any case. Why do we need 25 MBit on all the farms and rural lands? (quoted from another subthread) Because otherwise these areas will get stuck with <25mbps forever, and there’s already no reserve in 8-10mbps.

NoZebra120vClip 4y ago
Please let's not confuse more units here. 8Mbps (please capitalize the "M" for "mega" and lowercase the "b" for bits) is 8,000,000 bits per second. That translates to about 0.954 MiB/sec (Mebibytes per second, see the "i" and the capital "B"?) which is close enough to 1. We must keep in mind that there are two disparate things being measured in this thread and by FCC. The FCC is ostensibly measuring advertised signaling rates. Your Gigabit Ethernet signals at 1,000,000,000 bits per second, but can't transfer data that fast. Likewise, my cable ISP signals at 100Mbps down and 5Mbps up (yeah, it's criminal) but download speeds are a different thing. Download speeds can, and should, be measured in computer-oriented mebibytes per second, rather than bits per second, because you are, after all, transferring files. (And yeah, disk space is often measured in powers of ten...) Your ISP's data cap is undoubtedly measured in gibibytes or tebibytes (even though they call them gigabytes or terabytes, that's what the units mean.) So when you run speedtest.net or Ookla or whatever, you're measuring the actual throughput that the computer's network interface can squeeze through the narrowest straw in your link to the server. That is necessarily a touch lower than the lowest signaling rate of whatever equipment is in-between. Internet connections are sold by trumpeting signaling rates, but those are NOT download nor upload speeds. Never confuse them, because they are overly-optimistic estimates of your maximum throughput (which is infeasible given most PHY and link-layer frame designs.)
- wruza 4y ago
  I don’t get why my comment caused this. No one would think of millibytes per second in that context, and the 4% difference between mebi and mega plays no role here either. I don’t even know if my meters are binary or decimal and am curious where this assumption comes from.

by wruza 4y ago

NoZebra120vClip 4y ago

Please let's not confuse more units here. 8Mbps (please capitalize the "M" for "mega" and lowercase the "b" for bits) is 8,000,000 bits per second. That translates to about 0.954 MiB/sec (Mebibytes per second, see the "i" and the capital "B"?) which is close enough to 1. We must keep in mind that there are two disparate things being measured in this thread and by FCC. The FCC is ostensibly measuring advertised signaling rates. Your Gigabit Ethernet signals at 1,000,000,000 bits per second, but can't transfer data that fast. Likewise, my cable ISP signals at 100Mbps down and 5Mbps up (yeah, it's criminal) but download speeds are a different thing. Download speeds can, and should, be measured in computer-oriented mebibytes per second, rather than bits per second, because you are, after all, transferring files. (And yeah, disk space is often measured in powers of ten...) Your ISP's data cap is undoubtedly measured in gibibytes or tebibytes (even though they call them gigabytes or terabytes, that's what the units mean.) So when you run speedtest.net or Ookla or whatever, you're measuring the actual throughput that the computer's network interface can squeeze through the narrowest straw in your link to the server. That is necessarily a touch lower than the lowest signaling rate of whatever equipment is in-between. Internet connections are sold by trumpeting signaling rates, but those are NOT download nor upload speeds. Never confuse them, because they are overly-optimistic estimates of your maximum throughput (which is infeasible given most PHY and link-layer frame designs.)

wruza 4y ago
I don’t get why my comment caused this. No one would think of millibytes per second in that context, and the 4% difference between mebi and mega plays no role here either. I don’t even know if my meters are binary or decimal and am curious where this assumption comes from.