This public dataset contains the data used to train our system to
learn regular expressions that extract geohints from router hostnames,
as well as the product of using our system.

 + *-midar-iff.routers files contain IPv4 routers inferred using
   MIDAR and Mercator, annotated with node IDs and hostnames.
 + *-speedtrap.routers files contain IPv6 routers inferred using
   Speedtrap, annotated with node IDs and hostnames.
 + *-minrtt.txt files contain RTT samples towards each router from
   vantage points with known locations, identified by Node ID.
 + *.geo-re.json files contain JSON-formatted rules that researchers
   can use to interpret hostnames.  We include hoiho-apply.pl for
   applying those rules.
 + geocodes.txt contains the geocodes we used, excluding the CLLI
   codes, which we licensed from iconectiv.

If you use this data supplement, you are required to cite:

 M. Luckie, B. Huffaker, A. Marder, Z. Bischof, M. Fletcher, and
 k. claffy.
 Learning to Extract Geographic Information from Internet Router
 Hostnames.
 Proc. ACM Conference on emerging Networking EXperiments and
 Technologies (CoNEXT).

You are also required to cite the ITDK, from which this data is
derived.  The instructions for citing the ITDK are included at:

 http://data.caida.org/datasets/topology/ark/ipv4/

The data is designed to be used with sc_hoiho, which is included
as part of scamper:

 https://www.caida.org/tools/measurement/scamper/

To obtain the inferred regular expressions which are included in this
dataset release, you will need to build sc_hoiho by passing
--with-sc_hoiho and either --with-pcre or --with-pcre2 to configure.
When building sc_hoiho, ensure pcre (or pcre2) is in the path where
your compiler looks for header files and libraries.  For example:

CFLAGS='-I/usr/local/include' LDFLAGS='-L/usr/local/lib' ./configure \
 --with-sc_hoiho --with-pcre2

and then run:

sc_hoiho -O learngeo -d best-regex -g geocodes.txt -R <training-set>.minrtt.txt public_suffix_list.dat <training-set>.routers

Other options to sc_hoiho are documented in the manual page for
sc_hoiho.

 https://www.caida.org/tools/measurement/scamper/man/sc_hoiho.1.pdf