20060609: CONDUIT Latency (cont.)

>From:  Jeff Weber <address@hidden>
>Organization:  UCAR/Unidata
>Keywords:  200606092148.k59LmVjs003468

Hi Waldenio,

I just wanted to send you a quick note to encourage CPTEC to modify
CONDUIT request lines on moingobe.  If you are not the appropriate
contact, please let us know who we should be talking to (Alex?).

I hope that things are going well for you!  It is too bad that you
were unable to attend the AGU Spring Joint Assembly.  I did, however,
give your paper with as much enthusiasm as I could, and I think
that it was appreciated.

Cheers,

Tom

Jef Weber wrote:
>Lately we have seen high latencies for the CONDUIT feed on:
>
>moingobe.cptec.inpe.br
>
>http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CONDUIT+moingobe.cptec
> .inpe.br
>
>We were also seeing this same pattern with the data going to Portugal:
>
>http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CONDUIT+atm77.fis.ua.p
> t
>
>As you can see, the latencies to Portugal dropped dramatically around 0Z
>on June 9th.
>
>
>What we did to improve the ingest of the CONDUIT data was to "split" the
>feed request into 5 different request lines. This is effective if it is
>not the absolute capacity of the network that is causing the latency, but
>rather the volume and timing of the ingest of the products (ingestion
>rate)..
>
>Here is an example of how a single request can be divided into 5 requests:
>
>[since you are running LDM 6.4.* the "PRIMARY" designation is not needed]
>
>request CONDUIT "MT.gfs_CY.00*|MT.gfs_CY.12*" idd.cise-nsf.gov PRIMARY
>
>(or whatever it is currently) and split it into 5 request lines:
>
>request CONDUIT "MT.gfs_CY.(00|12).*[05]$" idd.cise-nsf.gov PRIMARY
>request CONDUIT "MT.gfs_CY.(00|12).*[16]$" idd.cise-nsf.gov PRIMARY
>request CONDUIT "MT.gfs_CY.(00|12).*[27]$" idd.cise-nsf.gov PRIMARY
>request CONDUIT "MT.gfs_CY.(00|12).*[38]$" idd.cise-nsf.gov PRIMARY
>request CONDUIT "MT.gfs_CY.(00|12).*[49]$" idd.cise-nsf.gov PRIMARY
>
>
>I also believe you are ingesting the 0,6,12, and 18Z runs rather than the
>0 Z and 12 Z runs in the example.
>
>(00|06|12|18) or a * would get all the times.
>
>Please let us know if you have any questions, and hopefully the data will
>arrive more timely.
>
>Cheers,
>
>
>Jeff
>---------------------------------------------------------------------
>Jeff Weber                                    address@hidden        :
>Unidata Program Center                        PH:303-497-8676        :
>University Corp for Atmospheric Research      3300 Mitchell Ln       :
>http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000  :
>---------------------------------------------------------------------

--
+----------------------------------------------------------------------------+
* Tom Yoksas                                            UCAR Unidata Program *
* (303) 497-8642 (last resort)                                 P.O. Box 3000 *
* address@hidden                                  Boulder, CO 80307 *
* Unidata WWW Service                            http://www.unidata.ucar.edu/*
+----------------------------------------------------------------------------+