kilobit vs kilobyte

The abbreviation kbit/s, does this stand for kilobit or kilobyte? Personally I'm thinking it stands for kilobit.

 

correct, kbit/s is the abbreviation for "kilobits per second "

also: kb/s = kilobits per second; but kB/s = kilobytes per second

...but the use of 'b' for bit and 'B' for byte is subject to so much misuse that it's probably better to stick with kbit and kByte if you want to be unambiguous

==

wearing my "pedantic" hat, I have to chuckle when folks write things like "512mb RAM" (which should of course be written 512MB RAM if it's to mean 512 megabytes)...

the prefix 'm' stands for milli (ie one thousandth) as opposed to 'M' which stands for Mega (ie one million); so 512mb actually means slightly over half a bit of memory (curious concept)

very best, HJ.

 

I was trying to convert 1.5 megabytes to either kilobits or kilobytes. I was doing a Tweak Tester at http://www.dslreports.com/tweaks. Before the results are displayed they tell you to list your download speed, as advertised by you ISP, in kbit/s. I'm thinking that is mistake.

When I convert it to kilobytes everything comes out just fine, but when I use kilobits it tells me to use an extremely high RWIN value for which I would have to turn on windows scaling. It just doesn't seem right.

 

Thanks a million TonyT. I'm fairly familiar with MTU=1500 and "packet loss" thus requiring resending that information. I've also noticed what you said that quite a few sites use the B (Bytes), b (bits), M (Megabyte) and m (megabits) interchangeably. These erros cause people to get incorrect information when running certain test.

P.S. Although the bandwith test site I previously posted asks for kbits per second (which gives me a rediculously high number), when I use kbytes I get a much more sensible result with an "efficiency" rating of 100% with no packet loss.

 

Tony,

I think you may perhaps have been a bit too generous with the dialup!

Say I start with a byte in my machine, which indeed is 8 bits; if I then want to send that byte to you via my modem: first I have to send a start bit, then there's the data, then at the end there's a stop bit (and ?maybe a parity bit): I've actually had to put something like 10 bits onto the line, in order to send you the 8 bits which become a byte in the computer at your end.

If the measurement of line speed means just what the words say (ie the number of bits transmitted down the line per second), surely it's going to be including this "overhead" as well as the actual data bits themselves - so (forgetting about parity for the moment) we'd get a more realistic idea of the number of useful bytes of data per second by dividing the line speed by 10; and, the numbers this gives would then be more in line with what I see - I'm "pegging the line" on the dark blue graph and the status bar

Now, I wonder, does a similar situation hold for broadband? (particularly as I'm about to change over to broadband in the next couple of days if things go well)...

...asynchronous communications channels generally use a start bit, they need something to give the "wake up call" to the receiver. True for a COM port, for MIDI, for a dialup, and (I gather*) also for ADSL. But what about other kinds of broadband?

*please someone jump on me if this is wrong, was trying to look it up yesterday and found all sorts of confusing stuff but little hard fact to go on

==

Another complicating factor in the relationship between bytes into a computer, and bits on the lines connecting the computer, might be data compression. Which (for my dialup) seems to be enabled, but I don't think that it's actually being used in my case.

In general though, whether dialup or broadband, if there's "on the fly" compansion going on, the actual number of bytes successfully into a machine would be greater than that predicted from calculation based upon the bit rate going down the line into that machine...

sorry to be confusing

best wishes, HJ