NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.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.
Hi Gary, Here are some numbers from the real-time statistics that Marcus' machine is sending back to Unidata: current data ingestion on gwoemul.wiu.edu: Data Volume Summary for gwoemul.wiu.edu Maximum hourly volume 481.502 M bytes/hour Average hourly volume 247.684 M bytes/hour Average products per hour 47720 prods/hour Feed Average Maximum Products (M byte/hour) (M byte/hour) number/hour HDS 179.680 [ 72.544%] 391.790 12378.295 IDS|DDPLUS 40.204 [ 16.232%] 55.577 35288.318 UNIWISC 22.617 [ 9.131%] 29.568 26.659 DIFAX 3.360 [ 1.356%] 12.090 4.841 FSL2 1.823 [ 0.736%] 2.076 22.227 Comparison of all of the data that a site like Marcus' could request given sufficient bandwidth: Data Volume Summary for idd.unidata.ucar.edu Maximum hourly volume 14155.751 M bytes/hour Average hourly volume 6973.663 M bytes/hour Average products per hour 249825 prods/hour Feed Average Maximum Products (M byte/hour) (M byte/hour) number/hour CONDUIT 2471.870 [ 35.446%] 7245.773 64061.738 NEXRAD2 2175.907 [ 31.202%] 2917.550 62035.357 NEXRAD3 820.510 [ 11.766%] 1048.914 51436.000 NGRID 690.929 [ 9.908%] 1406.051 15980.810 HDS 253.852 [ 3.640%] 393.416 17838.524 FNMOC 226.990 [ 3.255%] 1312.805 1960.190 NIMAGE 134.981 [ 1.936%] 241.821 168.571 FNEXRAD 87.758 [ 1.258%] 109.766 72.405 IDS|DDPLUS 40.365 [ 0.579%] 55.577 35461.214 EXP 37.829 [ 0.542%] 65.342 330.786 UNIWISC 22.599 [ 0.324%] 29.568 26.786 DIFAX 3.305 [ 0.047%] 12.090 4.833 GEM 2.807 [ 0.040%] 26.388 192.476 LIGHTNING 2.129 [ 0.031%] 4.094 232.762 FSL2 1.830 [ 0.026%] 2.076 22.357 A careful examination of both listing above show a couple of key features: 1) Marcus is requesting a small fraction of the data that is available 2) even with the current data request, not all of the data being requested is getting through One of the key things we look at is a comparison of delivery latency for two different streams of data, one of relatively low volume, and one of a relative medium to high bandwidth: IDS|DDPLUS receipt latencies on gwoeuml.wiu.edu (relatively low volume): http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?IDS|DDPLUS+gwoemul.wiu.edu HDS receipt latencies on gwoemul.wiu.edu (relative moderate to high volume): http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?HDS+gwoemul.wiu.edu I want to draw attention to the latencies seen at ~18Z each day: - the latencies for the low volume feed are acceptable - the latencies for the moderate volume feed are large enough that data disappears from the upstream feed machine before it can be sent to gwoemul These two observations taken together are a classic signature of artificial bandwidth limiting by packet shaping. Cheers, Tom -- **************************************************************************** Unidata User Support UCAR Unidata Program (303) 497-8642 P.O. Box 3000 address@hidden Boulder, CO 80307 ---------------------------------------------------------------------------- Unidata HomePage http://www.unidata.ucar.edu **************************************************************************** Ticket Details =================== Ticket ID: GOK-818269 Department: Support IDD Priority: Normal Status: Closed