Since it's inception in 1995, the PingER project has amassed more than
16 years worth of historical ping data from several tens of
around the world to many hundreds of
remote sites in over 160 countries.
PingER provides some canned tabular and graphical reports. If further insight
is needed then the PingER data has/can be accessed and used by collaborators and
others in the industy to research network related theories.
When access to archived data is needed for a user's research/use, there
are a few options available for the user to selecting the data to download.
When using this data, please acknowledge the
PLEASE NOTE: All times associated with the rawdata downloaded
from PingER are measured with respect to GMT.
- As a user you can visit our online distribution of data accessible at
selecting various options in the form based on the user's
needs, to view the data. From the web page, the data is available for download in
tab-separated-value (.tsv) format which makes it easy to import into Excel.
Another option via the Pingtable is that the user can write a script to use an
HTTP GET with the appropriate QUERY_STRING in the URL and work with the data
The example above will get you the hourly data from SLAC to all the sites in the
world for Feb 22, 2004. Stepping through all the days for the last X days
will get you all the data per the request. It will come back in tab
separated format which can be imported into Excel or your favorite spreadsheet
If you as a user are looking for a large amount of collective data for example
from all monitoring sites for all time monitored based
on one or more specific
(see fig.1), please contact us by
Please describe what data (metric, ping packet size (100 or 1000 Bytes), whether
the remote (monitored) nodes should be aggregated by site or independent)
you need and what your intended use is.
We will get back to you and probably prepare
a zipped tar file of the data and make it available by anonymous ftp. Contained in
this tarball are zipped files, one for each day, for each metric, for each
packet size (100 or 1000). If no metric is specified, only a tarball of
the 'average_rtt' for 100 Byte pings for each host (not aggregated by site).
will be returned. The tarball will be stored for a few days
in a read only FTP directory
where you can simply perform an FTP download of the data.
A typical tarball for one metric is over a GByte.
Once the tarball is detarred and the files unzipped, you will find
that the individual files are space separated and contain a line header with
the hour numbers. Each following line contains the structure:
source_host_name destination_host_name metric_for_1sthour ... metric_for_nthhour source_host_name destination_host_name (see fig. 2).
Figure 1: Metrics
See the Tutorial
for more on the metrics.
Figure 2: Example of format
For example the 1st 2 lines of file: packet_loss-100-by-site-2004-02-24.txt
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
pinger.slac.stanford.edu www.jinr.dubna.su 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 pinger.slac.stanford.edu www.jinr.dubna.su www.jinr.dubna.su
(i.e. the losses were always 0.000, no packets lost out of 10.)
- You can also use
to choose the monitoring and remote host pair of interest and click on the ? to use
find out more about the pair. From the dbprac.pl page you can click on
Add CSV file
to get the min/avg/max RTTs for the specified dates and times for the specified monitor-remote site pair, plus a time series plot.
If the user is in need of specific rawdata values (as gathered by getdata.pl from the monitoring hosts)
seen from SLAC you can use
HTTP GET and a URL like:
http://www.slac.stanford.edu/cgi-wrap/ping_data.pl to enter the begin and
end dates for which you are searching for data. The user can also enter
and write a script to get the specific data needed.
You will have to replace dd,mm, and yyyy respectfully to
reflect the begin/end dates you are requesting data for.
The contents of each line returned for the request are as follows:
source_host_name source_host_addr destination_host_name destination_host_addr size unix_epoch_time sent rcvd min avg max seq_rcv(i=1,rcvd) rtt_rcv(i=1,rcvd)
- size =100 or 1000 bytes in packet
- unix-epoch_time is the
GMT/UTS time when the measurement was made.
- sent is number of ping packets sent
- rcvd is number of ping packets received back
- min=min of ping packet RTT received
- the seq numbers of the packets received are indicated in sqes_rcv
- the RTTs of those packets are given in rtt_rcv
pinger.slac.stanford.edu 126.96.36.199 multivac.sdsc.edu 188.8.131.52 100 1077235276 10 10 28.682 32.445 35.427 1 2 3 4 5 6 7 8 9 10 31.4 32.4 33.3 35.1 35.4 34.9 32.6 28.6 31.2 29.1
There are more details on the
Monitoring Data Format.
The raw data is saved locally at slac in files of the form:
/nfs/slac/g/net/pinger/pingerdata/hep/data/<node>/ping-<yyyy-mm-dd>.txt.gz for example
Revised 12 September 2014.