Of course..and I should have mentioned this..
thanks to Sun, Linux would be the only OS
that you could get hardware support for. So
the total solution really is OS dependent.
-----Original Message-----
Sent: Friday, June 21, 2002 11:40 AM
By all means this is wonderful. It was really becoming
a problem. Although to be fair, it is really OS independent
as Solaris Intel, SCO, BSD..etc would have produced the
same result. The real key lies in the x86 hardware.
I am glad that NWS are recognizing that the reliability
gap is now closed between x86 boxes and very expensive
RISC/MIPS systems if you stick to high-quality, high end
components.
I lament everyday how slow our even fairly recent Sun
systems are in comparison to our Solaris Intel Dell
boxes.
-----Original Message-----
Sent: Friday, June 21, 2002 9:02 AM
As a followup to last week's switch over to Linux SBN uplink machines and
latency issues for NOAAPort and NOAAPort dependent systems:
Time on Queue Statistics - Units are in seconds, for NCF.
-----------------------------
Channel Max Average
-----------------------------
GOESEast Queue 102 40
GOESWest Queue 12 4
NWSTG Queue 45 1
DCP Queue 49 1
-----------------------------
These are the "fastest" times (lowest latency) that I have seen on NOAAPort.
The massive improvement with a Linux solution should benefit us all . . .
and
kudos should go to NOAA for having the foresight to implement a technically
and economically better SBN uplink facility. After last week's latency
issues - rooted on the old HP-RT systems and upstream facilities from the
NCF
- this is a pleasant come-around.
--
Stonie R. Cooper
Planetary Data, Incorporated
ph. (402) 782-6611
