thanks "KwikOne" very nicely explained!!

On Thu, Mar 4, 2010 at 5:50 PM, KwikOne <[email protected]> wrote:

> Coalwater,
> The 'granularity' is the 'slice' of time used for the calculations
> (this could be
> either actual time slots or 'elapsed time'). The smaller the
> granularity the
> larger the variations will appear in a graph. Using the previous
> example if
> the calculation was done with a once a second granularity (I will
> presume
> that a 1400 Byte chunk arrives each 100 ms.) which would mean that
> there
> would have been 1400 + 1400 * 10 chunks = 15,400 Bytes in the first
> second = 15,400 Bytes per second as the first point; but the second
> point
> would only have 1400 * 10 = 14,000 Bytes/second, a much lower
> difference.
>
> On Feb 28, 6:53 pm, Mohammad AbuShady <[email protected]> wrote:
> > I see, thanks a lot, and for your question about where the speed is being
> > show, it's both in the download window (firefox download window) and also
> a
> > network speed monitor over the whole system, of course they don't show
> the
> > same numbers but they both show the small increase in the start, and I
> don't
> > really know what 'granularity' means.
> >
> > ~Coalwater~
> >
> > On Mon, Mar 1, 2010 at 1:33 AM, KwikOne <[email protected]> wrote:
> > > You did not indicate "where" your download speed is being shown/
> > > calculated, nor the time
> > > granularity used. Quite often the speed spike are not really spikes
> > > (especially when first
> > > starting the download) because of the way the download speed is being
> > > calculated and
> > > as time goes on the download speed stays constant, or it will drop off
> > > then stay constant.
> > > Think of it this way (note the figures are just approximates as
> > > illustration only with rounding)...
> > > 1) download requested from server
> > > 2) first bit of data arrives; calculated speed is roughly amount of
> > > data / 1 ms. (smallest granularity)
> > >    = very high speed (1000 times amount of data)
> > >    (example - first chunk is 1400 Byte, this would give approx
> > > 1,400,000 Byte/sec rate)
> > > 3) next bit of data arrives; calculated speed is roughly total amount
> > > of data / (time from first bit
> > >    to second bit) = lower speed (unless the server is close enough
> > > that you have < 1ms time)
> > >    (example - second chunk 1400 Byte arrives 100 ms later which would
> > > give approx
> > >     2800 Byte for 100 ms. = 28,000 Byte/sec).
> > > 4) next bit of data arrives (example 1400 Byte 100 ms later which
> > > calculates out to
> > >    4200 Byte in 200 ms. = 21,000 Byte/sec).
> > > and so on... get the picture (it all depends upon how the calculations
> > > are done)?
> > > Notice how the rate is dropping off when in actual fact in this
> > > example it is actually 14,000 Byte/sec.
> >
> > > On Feb 22, 7:37 pm, Coalwater <[email protected]> wrote:
> > > > Well i noticed that i when ever i start a download, the download
> speed
> > > > starts with a very high speed spike in the beginning that could reach
> > > > up to 3x of my bandwidth, and last for like 3 to 10 secs then starts
> > > > to gradually drop till it reaches my usual expected download speed,
> > > > this isn't really a problem but i would really like to understand it
> > > > from the networks point of view, how does it get past the ISP limit
> > > > even if for a very short period of time, because i might think of
> > > > using it to my advantage somehow if i understand it, Hope some
> > > > networks guy around here could explain to me :D
> > > > oh and it doesn't depend on the connection type nor the operating
> > > > system, i saw this happen on cable connection and mobile broadband,
> > > > windows and ubuntu so it doesn't have any thing to do with those
> > > > differences.. oh and it works also with both normal single connection
> > > > download and download accelerators that use multipart downloading
> >
> > > --
> > > You received this message because you are subscribed to the Google
> Groups
> > > "Computer Tech Support" group.
> > > To post to this group, send email to
> > > [email protected].
> > > To unsubscribe from this group, send email to
> > > [email protected]<computer-tech-support%[email protected]>
> <computer-tech-support%[email protected]<computer-tech-support%[email protected]>
> >
> > > .
> > > For more options, visit this group at
> > >http://groups.google.com/group/computer-tech-support?hl=en.
>
> --
> You received this message because you are subscribed to the Google Groups
> "Computer Tech Support" group.
> To post to this group, send email to
> [email protected].
> To unsubscribe from this group, send email to
> [email protected]<computer-tech-support%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/computer-tech-support?hl=en.
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Computer Tech Support" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/computer-tech-support?hl=en.

Reply via email to