innreport (Outgoing Feeds by Volume)

Tue Jan 25 15:23:27 UTC 2000

Elena Samsonova wrote:
> On Tue, 25 Jan 2000, Kenichi Okada wrote:
> 
> > I found a strange part in my innreport.
> > 
> > Outgoing Feeds (innfeed) by Volume:
> >             AcceptVol RejectVol  TotalVol  Volume/sec   Vol/Art  Elapsed
> >               14.7 GB  196.5 MB   14.9 GB   81.7 KB/s   13.6 KB  53:06:42
> >               10.8 GB  167.1 MB   10.9 GB   59.9 KB/s    9.7 KB  53:06:36
> >                4.6 GB    7.5 MB    4.6 GB   25.1 KB/s  176.6 KB  53:06:55
> >                2.1 GB  196.8 MB    2.3 GB   10.0 KB/s  255.5 KB  67:12:21
> >                1.0 GB   30.7 MB    1.1 GB    5.9 KB/s    2.0 KB  53:10:48
> >              855.9 MB    1.6 MB  857.4 MB    4.6 KB/s  156.2 KB  53:07:02
> >              177.1 MB   28.6 KB  177.1 MB    0.9 KB/s   16.6 KB  53:08:08
> >              168.0 MB  472.6 KB  168.4 MB    0.9 KB/s   17.2 KB  53:06:37
> >              158.3 MB    2.6 MB  160.9 MB    0.9 KB/s   37.8 KB  53:24:01
> >              155.2 MB    1.4 MB  156.6 MB    0.8 KB/s   21.6 KB  54:33:32
> >              142.3 MB    0.0 KB  142.3 MB    0.8 KB/s   16.6 KB  53:06:38
> >              115.4 MB  134.2 MB  249.6 MB    1.3 KB/s    3.8 KB  53:10:06
> >               80.7 MB    5.2 MB   85.9 MB    0.5 KB/s   18.1 KB  53:06:38
> >               77.2 MB    3.9 MB   81.1 MB    0.4 KB/s    8.1 KB  53:06:38
> >               51.7 MB  327.2 KB   52.0 MB    0.3 KB/s    2.6 KB  53:06:38
> >               47.8 MB    6.4 MB   54.2 MB    0.3 KB/s    2.1 KB  53:06:35
> >               46.6 MB    2.7 MB   49.2 MB    0.3 KB/s   17.2 KB  53:06:29
> >               23.0 MB   12.0 MB   35.0 MB    0.1 KB/s    3.1 KB  81:17:17
> >               22.4 MB  146.9 MB  169.3 MB    0.9 KB/s   16.4 KB  53:08:09
> >               17.5 MB   36.0 KB   17.5 MB    0.1 KB/s    5.0 KB  53:06:38
> >               11.4 MB    2.5 MB   13.9 MB    0.1 KB/s    3.7 KB  53:07:25
> >               10.7 MB  116.5 KB   10.8 MB    0.1 KB/s    2.9 KB  53:06:35
> >                5.8 MB    7.7 MB   13.5 MB    0.1 KB/s    4.8 KB  53:13:45
> >                5.4 MB    7.4 MB   12.8 MB    0.1 KB/s    4.6 KB  54:11:06
> > TOTAL: 24     35.3 GB  934.0 MB   36.2 GB    8.0 KB/s   12.1 KB 1320:03:19
> >                                              ^^^^^^^^
> > 
> > The last line of "Volume/sec" seems to an average, it should be a *total* ?
> > Is this on purpose?
> 
> A total of a timed item like volume/sec does not make sense.  Would you
> want all numbers to be summed up?  That would give you 196.1 KB/s which is
> probably not what you want.

I think that we need to look at what the volume/sec in the total
line means:

It appears as though 35.3GB were transferred in ~55 hours (since
all the tranfers for the 24 feeds took place in parallel). Working
that out, I get 186KB/s which is quite close to the sum you give
for the column.

However, I suspect that each feed had multiple simultaneous
connections and the figure for volume/sec for each feed line is
the average transfer rate of each of those connections, and the
time is the total time for all those connections.

Looking at it this way, the total volume/sec is probably the average
transfer rate for each individual nntp connection. And there may
have been multiple connections for each feed. In fact, we might
even deduce that there were an average of 5 simultaneous connections
for the feed described on line 1 of the report.

I think that what Kenichi Okada is after is to take the volume of
each feed divided by the time beetween the rotation of his logs
which I assume is 24 hours. In which case the report would look
like (taking AcceptVol/86400 which give line utilisation approximately):

             AcceptVol RejectVol  TotalVol  Volume/sec   Vol/Art  Elapsed 
               14.7 GB  196.5 MB   14.9 GB  178.4 KB/s   13.6 KB  53:06:42
~~~~~
 TOTAL: 24     35.3 GB  934.0 MB   36.2 GB  428.4 KB/s   12.1 KB 1320:03:19

Perhaps this clears things up?

-- 
Ian Freislich
UUNET S.A. (Pty) Ltd