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

[LDM #VNG-594789]: What's with the lag time on my NIMAGE feed?



Hi Gilbert,

re: 
> OK. This is NOT, repeat, NOT a big deal. I am just trying to understand
> why this is happening, for my own purposes.
> 
> http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?NIMAGE+weather2.admin.niu.edu
> 
> My NIMAGE feed is frequently up around 30 seconds.

We took a look at the latencies you are seeing for some NIMAGE products on
Friday, and compared them with the latencies seen on your upstream feed
host, idd.unidata.ucar.edu.  We noticed that there was no additional
latency being added by idd.unidata.ucar.edu, but there was some latency
from the NOAAPORT ingest boxes to idd's accumulator front ends, 
oliver.unidata.ucar.edu
and emo.unidata.ucar.edu.  We have not explanation for this latency at
this time.

> I am on the Internet 2,
> 1 GB pipe, nothing significant to interfere. My machine is a 3 GHZ
> processor, 800 MHZ bus processor with a 250 GB IDE hard drive. I think the
> slkow drive is causing the latency. 

We do not thing so, but I can validate your notion that local system performance
can introduce latency.  It is much more common, however, for local system
performance to affect how fast data can be processed out of the queue
and not how fast data can be put into the queue.  Latencies seen on products
you receive are calculated as the difference in time the products were
received by you (not when they are processed out of the queue) and the time
that products were first injected into the IDD (i.e., when the product first
was inserted into the first LDM's queue).

> College of DuPage gets no latency, but
> they have a much faster machine...a true high-level server, on the
> commodity Internet. They get NO delays, but they also only get a fraction
> of the feed.
> 
> What do you think? Given that your UCAR planes can do 4 GB with no delays,
> I am thinking my machine is slowwww...


I don't think so, but I am not positive enough to assert that you can stop
looking at your own system's performance.

> *******************************************************************************
> Gilbert Sebenste                                                     ********
> (My opinions only!)                                                  ******
> Staff Meteorologist, Northern Illinois University                      ****
> E-mail: address@hidden                                  ***
> web: http://weather.admin.niu.edu                                      **
> *******************************************************************************
> 
> 

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: VNG-594789
Department: Support LDM
Priority: Normal
Status: Closed