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

20000306: Plymouth NOAAPORT receivers and GINI imagery (cont.)

>From: Unidata User Support <address@hidden>
>Organization: Unidata Program Center
>Keywords: Plymouth NOAAPORT GINI ADDE

Jim,

re: which NOAAPORT channels do you have
>We have all 4 channels.

OK, so I wasn't dreaming :-)

re: access to the GINI imagery on your FreeBSD boxes
>Currently, the data are being ingested and stored on the four Pentium
>computers that were provided by Unisys as part of the NOAAPORT package.
>These systems are all running QNX for the OS.  WXP ingest is being used
>as the ingest program.

Hmm... Are any changes made to the GINI images, or are they being stored
exactly as they are sent out?  Hopefully, Dan's code is not munging
anything in the images.

>I can use a pqing request on one of my FreeBSD
>machines running LDM to pump data into LDM and potentially the IDD. I
>tried this initially to a local machine, but got deluged with data.

It is daunting to say the least!

>One
>would really need to tailor pqact.conf and reconfigure scour routines
>significantly to handle the volume.

We are getting the data off of the COMET 4-port ingest setup.  Chiz wrote
code that PNG compresses the GINI imagery before it is sent to the UPC
and then uncompresses it after it gets here.  The savings on the 26 MB
VIS images is on the order of 55% (i.e., the compressed image is 45% of the
uncompressed one).  The size savings is a function of the image, so this
number can be as good as 35% or as bad as 55%.

>Since I haven't had time to do
>this, I simply use NFS to mount the filesystems from the ingestor
>machines onto my BSD boxes and access the data in this fashion.

OK, this sounds like a good setup if your NFS is fast.

>Since
>Dan set the system up to be in the typical yymmddhhMM format, all works
>well with my WXP stuff.

My ADDE servers don't care what the files are named, but they have to
be unique (the software has to be able to find all of the data of 
a type using a regular expression).

>BTW, our NFS really works great, since all of
>our internal network is either at 10MB/s (different room) or 100 MB/s
>(same room) and we're bridged off from other campus systems, so we don't
>have many data collisions except from our own traffic.

Sounds good!

re: sharing your wealth
>We definitely would be willing to share,

Excellent.  This is very generous of you!

>but I may need your help to get
>things working.

I am willing to do ALL of the setup work.  All I need is:

o login as 'mcidas' on the machine that you want to allow sites to
  contact using ADDE
o capability of doing things as 'root' (sudo or root login) on that
  machine
o McIDAS to be installed

>We've been so busy lately that we haven't had time to
>even look at McIDAS, since I talked with you in Long Beach.  We're really
>short-handed since our synoptic person left and Joe and I have to cover
>many of that person's duties, conduct a search for a replacement, manage
>four different grants, etc., plus manage a large number of systems.

Phew!  Sounds terrible!!

>I'd be willing to provide you access and show you the NFS paths I may
>have to add some paths for Channel 4, since I haven't really looked that
>much at the data coming over that port.

Wonderful.  As soon as you can give me access (see above), I will login
and setup ADDE access to your GINI imagery data.  What we do from there
(like announce access to the Unidata community) is in question.  I think
that we will need to take some initial steps to insure that your system
doesn't get inundated with data requests.  I am hoping for at least
one more full NOAAPORT ingestion site to share the load.  Three to
five sites should go a long way towards providing GINI imagery access
to the Unidata community.  What happens when the NIDS stuff becomes free
next October is anyone's guess, however.

re: other sites with NOAAPORT ingestion systems
>Other than Wisconsin, I'm not too sure. Mike Dross/DukePower also has a
>Unisys NOAAPORT system. I believe that it's similar to ours. I'm not
>sure about their Unidata status, but I know Mike is on several Unidata
>e-mail lists.

Linda suggested that Blue Skies might have a 4-port setup.  I will be
contacting Jeff Masters about this either tonight or tomorrow.  I believe
that their downlink is located in San Francisco.  If this is true, and
if they are getting the imagery data, and if they are willing to share
with the community, then there will be three sites stretched across
the country: Plymouth, UPC, Blue Skies.  This would be very useful.

Tom

>From address@hidden  Mon Mar  6 16:38:25 2000

Tom,

I'll do some configuring to set you up tomorrow, as I've got to leave
for tonight. Currently, you can't su or telnet into the superuser with a
user account like mcidas. I'll set you up with a tom_y account that will
be in the right group to su.

I think we'll be using snow.plymouth.edu. It's a PIII 500MHz system that
isn't overly taxed unless we change it from backup to primary web
production machine. It's the one that you were working on in Long Beach,
so it shouldn't need much tweaking.

                        Jim
-- 
James P. Koermer             E-Mail: address@hidden
Professor of Meteorology     Office Phone: (603)535-2574
Natural Science Department   Office Fax: (603)535-2723
Plymouth State College       WWW: http://vortex.plymouth.edu/
Plymouth, NH 03264