NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Dr. Koermer, Interesting approach; it would be a good mental game to see how much the SBN time differs from the internet provided atomic clock NTP. Certainly, NOAAPort receive paradigms differ from vendor to vendor - especially when it comes to injecting data from the SBN into an LDM. One reason that our time stamps on some of the first NEXRAD products for a volume scan are the same as the header - i.e. a 1730Z product being completely written by 1730Z on the system - is that our LDM injection facility streams the data directly to the queue, as opposed to buffering and pqing'ing. I can't say that one method is better than the other . . . just different. It does help explain differences in latency, though, when troubleshooting larger issues external to the NOAAPort receiver - such as NCF uplink issues, internet drops on IDD, etc. We've sufficiently beat that dead horse. Stonie On Thursday 29 November 2001 22:16, Jim Koermer wrote: > Stonie, > > We've got things reversed--our NTP server is the final receptor and our > four NOAAport systems synch off that server. The times in our filenames > are pulled right from the headers. The directory listing times are times > that the files were received by the server. As I indicated, the NOAAPORT > machine feeding the LDM on the server buffers the data in one minute > files and then dumps the buffer files at one minute intervals for LDM > ingest on the server. It seems to do this at 30 seconds after the > previous minute's buffer has been built. Hence, the data actually do > come in "at least" a minute (since seconds aren't recorded) earlier. > It's certainly close enough to real-time for our purposes. > > We used to grab the data via NFS directly from the NOAAport ingestor, > but our hard drives are now way too small to hold very much data because > of the deluge from the new NIDS products, so we now move them over to a > server where we have a large RAID system, which allows us to store all > products from all the NIDS sites for a much longer period. > > Jim > -- > James P. Koermer E-Mail: koermer@xxxxxxxxxxxxxxxxx > 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 -- Stonie R. Cooper, Science Officer Planetary Data, Incorporated 3495 Liberty Road Villa Rica, Georgia 30180 ph. (770) 456-0700; pg. (888) 974-5017; fx. (770) 459-0016
ldm-users archives: