Remote access: Speed tests
Doing remote data collection and transferring images to your home
institution obviously requires fast network connection. But how fast is
the 'fast' and is my connection speed sufficient? You can use Speedtests
by OOKLA to benchmark your connection. Point your web browser to:
and choose a server close to Argonne, for example "Chicago, IL - Frontier", or
"Chicago, IL - Comcast", or "Chicago, IL - Speedtest.net". You should see the
download upload speeds no less than 600Mbps and 400Mbps respectively.
These are the numbers we observe for connections from the University of Michigan
and which are just about appropriate. To observe these speeds you should
have a computer with at least a gigabit wired connection.
Wi-Fi should not be used under any circumstances,
even if you have 802.11ac claiming 1.7 Gbps or 3.4 Gbps speeds. The matter
is that Wi-Fi connections are prone to packet loss with subsequent re-sending.
Wi-Fi may work fast when transferring data between two local computers, but
the speeds of long-distance connections deteriorate drastically even when
only a short arm of the connection (e.g. within a room) is over Wi-Fi.
This is because corrupted packets have to travel all the long-distance way
to the endpoint which then discovers corruption and requests re-sending.
If your numbers are significantly lower, that may become a show stopper for
remote access. It is important to understand where the problem may occur. Such
task is not trivial because in most cases there is no a single bottleneck, but
the delays add up non-linearly. In other words, you may have a descent speed
between A and B and between B and C, but slow between A and C in spite of the
fact that the connection between A and C may be routed via B. It is a good
idea to work with your institution's network admins and bring them into
contact with the GMCA staff.
The test below is to verify downloading speeds of a large 88MB file from GM/CA.
Compare your results with those for the University of Michigan (you can also use a
web browser instead of wget):
2019-06-11 16:09:41 (64.5 MB/s) - libreoffice-core-22.214.171.124-3.el7.x86_64.rpm saved
Finally, if you already have an SSH access to the GM/CA computers, you
may test the real data transfer speeds. Make a directory, e.g. TEST with twenty
or so data frames produced by the detector and copy them to your institution.
Below are typical speeds for the University of Michigan obtained with 32MB
Rayonix-300 CCD frames:
[sstepanov1@watson ~]$ scp -rp email@example.com/EIGER_TEST_IMAGES .
series_1_master.h5 100% 297MB 75.1MB/s 00:03
series_2_master.h5 100% 297MB 101.2MB/s 00:02
series_3_master.h5 100% 297MB 129.7MB/s 00:02
series_4_master.h5 100% 297MB 125.0MB/s 00:02
series_5_master.h5 100% 297MB 122.7MB/s 00:02
c3_1_01_master.h5 100% 89MB 96.7MB/s 00:00
c3_1_01_data_1.h5 100% 31MB 72.3MB/s 00:00
c3_0_01_00003.cbf 100% 17MB 61.6MB/s 00:00
series_00001_4.cbf 100% 17MB 56.8MB/s 00:00
series_00001_5.cbf 100% 17MB 45.9MB/s 00:00
To estimate required speeds, you should consider that when data
collection is running permanently, the Rayonix-300 detector frames are typically
acquired at the rate of 22fpm (frames per minute), which corresponds to
32MB x 22/60 = 11.73MB/s. With Pilatus3-6m typical rates are much higher:
50fps (frames per second), which corresponds to 6MB x 50 = 300MB/s. With
Eiger-16M they may be even higher: up to 130fps, which corresponds to
18MB x 130 = 2.4GB/s. However, given typical overhead of crystal mounting,
centering and screening, average daily data rates are ~200GB/day with
Rayonix-300, ~1TB/day with Pilatus3-6m and ~3TB/day with Eiger-16m. This
amounts to required average speeds of 2MB/s, 10MB/s and 30MB/s with Rayonix,
Pilatus and Eiger respectively. With some experiments the amounts of data
may be up to several times higher than average. We strongly recommend using
Globus for transferring Eiger and Pilatus data.
Improving the speeds
It may be possible to improve network speeds by tuning the TCP
parameters on your computer, although the administrative access to your
computer and some computer administrating skills may be required.
Please read the fasterdata.es.net guide.