AFOS discussion (fwd)

===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================

---------- Forwarded message ----------
Date: Tue, 9 Mar 1999 10:38:39 -0700
From: "Glenn P. Davis" <address@hidden>
To: David Copley <address@hidden>
Subject: AFOS discussion

David:

Thanks for responding to Robb's note.


>      Robb,
>
>      I work for the National Weather Service.  We do not use the NOAAport
>      feed as that has been assigned for our private customers, and we have
>      AFOS in all of our offices.  Also, if all NWS offices pulled NOAAport
>      data we would swamp the server.
>
>      Another consideration is that until AWIPS is commissioned NWS doctrine
>      states that AFOS will remain the primary data source in the NWS.  So
>      we are obligated to maintain AFOS and its data stream for at least 6
>      months to a year (maybe longer).
>
>      To bring AFOS data into the HP workstations or LINUX workstations
>      usually is accomplished through the use of an APT.  The APT is a PC
>      configured with a synchronous to asynchronous modem and a software
>      package that basically passes all AFOS products through to the
>      workstation.  We then assign the APT a synchronous node ID on the AFOS
>      system and AFOS then transmits all received products on to this node.
>      The software/hardware in the APT grabs the data stream and passes it
>      along to the workstation where the LDM program reads and stores the
>      data in a directory structure compatible with the GEMPAK display
>      software.
>
>      That's probably more than you really want to know, but if you have any
>      question feel free to ask.
>
>      David

So, many NWS sites have put together a bridge system
which uses the APT, a *NIX workstation, the LDM and GEMPAK.

What we are suggesting is that you consider migrating the input of the
*NIX workstation LDM off of the APT to an LDM system somewhere providing
data from NOAAPort. You would only request the data you find useful for
your GEMPAK display's, not the full NOAAport feed, so you are not going
to "swamp the server". You would consume more IP bandwidth as opposed to (or
in addition to) your X.25 bandwidth. I don't know if this is in short supply
or not. If one only requested the text products (no model outputs) from the
NOAAport NWSTG channel, one would be within the bounds of asynch modems.

The advantage of this is that you are moving ahead.
When you get a NOAAport downlink at your site, you can just cut over.
When compared to the installed base of LDM and GEMPAK sites, this is
a more common case, so there is better support and you would have
more "buddies" to provide informal tips and so on.
There may also be certain operational advantages, like less hardware to
deal with (APT is out of the picture) and less wiring constraints.

-glenn