CFHT - CADC transfer rates

From: kanoa <kanoa_at_cfht.hawaii.edu>
Date: Mon, 31 Mar 2008 18:33:12 -1000

Hi,

Jean-Charles mentioned earlier that one factor in deciding to postpone
reprocessing the CFHLS data was a lack of network bandwidth and I
thought I would add a short note to explain that situation.

Along with the rest of the Mauna Kea observatories, CFHT recently
upgraded our network connection to the global research networks from 45
megabits to 1,000 megabits. That upgrade has gone well and we were
hoping to use the bandwidth to download old science data from the CADC,
detrend it using the latest recipes, and then re-upload it. It turns
out, however, that we have great network performance to everywhere
_except_ the CADC. Actual performance to the CADC is the same or
slightly worse than it was before the upgrade.

I still do not have a convincing explanation for why. My best guess is a
peering agreement between one of the three involved research networks
(Australia, USA, and Canada) is not working properly. At this point I
have made contacts with representatives of each network and have done
throughput tests from points across Hawaii, Canada and the US mainland
to trace the bottleneck. I'm confident we can get the problem resolved
but until we discover what the problem is I can't predict how long it
will take. Hopefully not much longer than it already has.

In conclusion, if you choose to do a global reprocessing at a later date
we should have better network infrastructure to support it. I will also
post another message to this forum when we get the network bandwidth to
what it should be.

Aloha and thanks for your patience,

-Kanoa
Canada France Hawaii Telescope
Received on Mon Mar 31 2008 - 18:33:12 HST