[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[IDD #VYA-733962]: LDM latency info and McIDAS access



Hi Gilherme,

re: it is likely that the long term rtstats files we have does not have
usable latency information

> Ok! I I do believe, however, that I can convince them (network admins)
> that our problems are of internal origin mainly by comparing our
> current latency data with UIB (mm5opteron.uib.es - the nearest LDM
> server around).

OK, this would be good.

> Nevertheless, the summary records could give an
> insight of the 'variability' of total transferred data and, since the
> feed itself doesn't change that much, could be linked to a change in
> the latency...

You can review the summary data online at:

Unidata HomePage
http://www.unidata.ucar.edu
  Unidata IDD
  http://www.unidata.ucar.edu/software/idd
    Real Time IDD Statistics
    http://www.unidata.ucar.edu/software/idd/rtstats/
      Summary Statistics by Host
      http://www.unidata.ucar.edu/cgi-bin/rtstats/sitesummaryindex

Look through the list presented in the last page for the U Aveiro machines.
I currently see entries for atm77-fis.clients.ua.pt and atm78.fis.ua.pt.
atm77.fis.ua.pt does not show up since it is now reporting statistics
as atm77-fis.clients.ua.pt.

When you review the summary CONDUIT statistics for atm78.fis.ua.pt, you
will see that there was a marked increase in volume starting sometime
in January.  This leads me to ask what may be a key question:

Question:

- was the CONDUIT request at UA changed from a subset to the full datastream
  sometime near the beginning of the year?

I distinctly recall that Yoshihiro had limited the CONDUIT request to a
subset of what was available.  I _think_ that the subset was the global
GFS data, but I do not recall if this was the 1 degree of 0.5 degree data.
 
re: subnet

> About three weeks ago the network admins 'created' a new subnet and
> moved our servers to this new nomenclature... we initially hoped that
> this change could improve our latency, but it stayed just about the
> same as before.

OK.  The question remains: was the request for CONDUIT changed?  If it
was, we may be seeing the effects of receiving more data than the
connection can handle.  An increase in latency in this case would be
expected.

> Thanks for your help!

No worries.

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: VYA-733962
Department: Support IDD
Priority: Normal
Status: Closed