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

Re: 20010806: feedtypes question for non-standard data

chris calvert wrote:
> 
> >
> > Would you please describe the topology further for me?  How many
> > machines we talking about, and who feeds from whom?
> >
> 
> I think I misspoke when I said we participated in IDD. We're bottom
> feeders. The radar data is collected at various radar points, shipped to
> the National Climatic Data Center (NCDC) for archiving, then shipped to
> the machines on which we derive our precip products. Right now there are 3
> machines (2 in AZ, 1 in OK). Soon we will add a machine in NC, AL, and
> another in AZ.
> 
> Our "network" of machines is at the bottom of the food chain. NCDC would
> be considered the "top". There's no chance that Unidata (or anyone outside
> of our NSSL domain) would get our products. Visually speaking (as best
> I can using a text only e-mail program), our products would flow
> horizontally on the topology tree, and not upstream or downstream.
> 
> CRAFT is still rapidly evolving and things will change. Basically, the
> NSSL machines would be considered a "private" network shipping products
> back and forth to one another. No one else will be getting this stuff.
> 
> Chris

Hi Chris,

So, you're like a carp! :-)

I guess that sounds safe enough.  You could use EXP to start, and/or you
could use any existing feed types.  If you move beyond EXP, I suggest
using the FT?? names, e.g., FT7 or FT10 so that confusion doesn't arise
over the names.  Also, you might want to avoid the popular feeds of both
now and in the future, just in case your sites may ever want to get that
data.  Those would be: UNIDATA and all its constituent feed types (IDS,
DDS, PPS, HDS, and MCIDAS), and FT21 - FT27, i.e., NIMAGE - NNEXRAD.

You might need to make it clear to the sites that you are feeding that
they face restrictions on using their ldms, i,e., they can't participate
fully in the IDD.  Of course, that's assuming they know what the IDD
is...

Also, please have your sites *not* run pqbinstats and 'ldmadmin dostats'
- the reports originating from your sites wouldn't really reflect the
latencies on those feed types since you're using them for your own
purposes. 

Good luck!

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