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

Re: 20011008: IDD latencies at PSU (cont.)

"Arthur A. Person" wrote:
> 
> 
> I've been watching a notifyme for NMC2 from motherlode and maybe
> understand this a little better.  It appears that large amounts of NMC2
> data are sent within matters of seconds or a few minutes based on the
> transmission times.  If they were received and stored on motherlode in
> like fashion, that means a huge amount of NMC2 data essentially gets
> inserted in the middle of the DDPLUS|IDS|HDS stream if the streams are
> merged when sent to downstream sites.  Thus a delay in the delivery of
> DDPLUS etc... would occur while sending this large amount of NMC2 data
> because it was inserted in a very short amount of time.
> 
>                                     Art.

Hi Art,

I don't think it really works this way.  Although the NMC2 products are
initially dumped into the system in large clumps, they get serialized as
they make their way into the pipe.  By the time the get to motherlode, I
would think they would be intermixed with the other products that
motherlode is getting from our local network.

So, in my own mind, I can't explain it.

Btw, it looks like NCEP was sending some 0 length files yesterday - that
should've improved your latencies for a while!  Not the way we would've
liked, however...

Anne 
-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************