Participation in the ITHI study

Thank you for considering participation in the ITH study. The measurements of the DNS displayed in metrics M4 and M6 rely on statistics collected at recursive resolvers. We suppose that the traffic observed at different sites varies. It may depend on location, time of day, day of the week, and type of resolver, whether it is an enterprise, a University, an ISP, or a mobile network provider. We are also interested by collecting data at authoritative servers, as this will enable us to derive statistics on the type of load observed by these servers. The more sites particpate in the study, the more interesting the statistics.

Participation in the study entails the following steps:

  1. Install the DNSCAP utility. Dnscap is maintained by DNS-OARC. One option is to clone and build from github. Another option is to get the tarball from DNS-OARC, and to follow the build step specified in the DNSCAP README on GitHub.
  2. Download and build the ITHI tools, as explained here.
  3. Run a test script to verify that the capture tools work.
  4. Obtain an SSH account on the staging server, ithi.privateoctopus.com. This requires mailing a copy of your SSH public key to Christian Huitema (huitema at huitema.net).
  5. Set up a "cron" job that capture statistics several times a week and upload them to the staging server using scp.

Once this is in place, the data that you send will be used to compute the monthly metrics. And we will also of course credit your participation on this site.

There are more details on the Wiki pages of the Github project "ithitools", including:

Privacy considerations

We understand that the DNS traffic is privacy sensitive, and contains Personably Indentifiable Information such as:

Our data collection methods are specifically designed to minimize the privacy issues. We do not collect IP the source IP addresses of queries, and only process the TLD part of queried domain names. A capture session will produce a CSV file that only contains statistical aggregates. The size may vary based on the amount and diversity of the traffic, but it is typically just a few kilobytes of text, maybe 8 to 16KB. An example of the file that we produce is available here.

The sources of our software are publicly available on GitHub, with an MIT license. They can be easily inspected, and if you wish you could use it to produce a local version of our web site, using your own data.