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

20030715: upgrade to LDM-6 at NMT (cont.)

>From: "D. J. Raymond" <address@hidden>
>Organization: NMT
>Keywords: 200307041809.h64I9TLd018509 IDD LDM-6

Hi Dave,

>I have changed my setup as you suggest -- how are we doing on
>latencies now?

The latencies for IDS|DDPLUS and UNIWISC have dropped to near zero:

http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?IDS|DDPLUS+heron.nmt.edu
http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?UNIWISC+heron.nmt.edu

but the HDS latencies remains high:

http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?HDS+heron.nmt.edu

This is consistent with your postulating that there is possibly some
sort of packet shaping being done on your campus, but it is also consistent
with some sort of outbound packet shaping being done at your upstream
feed site.  The behavior we are seeing right at the moment is the same
as we saw at the University of Louisana-Monroe a couple of weeks ago.
Their problem turned out to be feed limiting of some sort (the exact
cause was never determined) at or near the upstream LDM site (LSU).  I
say at or near because the LSU IT folks and the LSU SRCC contact said
that nothing was being done in networks they are overseeing that would
purposefully limit the feed.  There was, however, an open trouble ticket
on the network that connected LSU to I2.

Since the LDM on heron has not been running for a sufficient length of
time to really know how well the HDS ingest is working, I would ask you
to leave the feed as it is for at least a day.  After a day (4 model
output cycles), if the HDS latencies remain high, I would like you to
change your HDS request from yin.engin.umich.edu to
rainbow.al.noaa.gov.

We know that rainbow.al.noaa.gov does not have any restrictions on
their outbound traffic, so if your HDS latencies remain high after
switching to it, then the problem is at or near NMT.  If the latencies
drop to zero, the problem is closer to U Mich.

Tom

>From address@hidden Tue Jul 15 18:51:09 2003

>Hi Tom,

>OK, I'll leave it set up this way for a day.

>Dave