How the RIPE NCC
Service Region Hostcount++ is Performed
The Hostcount++ is an improved version of Hostcount. In this version issues of accuracy, performanse were addressed. Several new metrics were introduced as well, including hosts visibility info, IPv6 addresses and hosts resolved using reverse method. Hostcount++ is run in the middle of each calendar month.
The Hostcount++ uses the following process:
Third Parties Collect Local Data
The Hostcount++ results are more accurate and more data is collected when data is collected on a 'local'
machine. This usually means a machine that is geographically
situated in the country that the TLD represents. The main reasons for
this are:
- Connectivity external to the geographic region has no influence on
the data collection
- The organisation collecting the data, specifically the IP address
range of their network, is usually 'well-known' in the area and therefore
more trusted. This means that [zone transfers] are refused less often,
so that more data is found.
Because of the increased accuracy of third party data, we prefer that
the data collection is done by a local site wherever possible, usually
a Network Information Centre (NIC), University or similar institution.
To collect the data, each local organisation runs the Hostcount++ DIY Kit
tool on their TLD code. The workflow of the whole process is as follows:
- Zone transfer the Top-Level Domain.
This means: ask one of the nameservers for the TLD to give out everything
in its zone file for that TLD. If a nameserver is refusing zone transfers or is down, use another nameserver from the list of authoritative nameservers for this TLD.
- Remember any Nameserver records that
are seen.
Namserver records indicate the existence of a sub-zone and tell us which
nameservers should know something about the sub-zone.
- Zone transfer all of the sub-zones.
- Remember any Nameserver records that
are seen.
- Are there any more subzones ?
- If Yes - Go to step 3.
- If No - Stop.
The RIPE NCC Collects Remaining Data
For all of the TLDs served by the RIPE NCC that do not have a local
data collector, the RIPE NCC collects the data from its site in Amsterdam.
The method used is the same as for the local data as described above. Additionaly, the Hostcount++
is running in reverse mode once the data is collected using the normal method. This approach allows
to count hosts which cannot be resolved using the normal forward method. Reverse mode is based on
enumerating zones based on Pointer Records assigned to RIPE NCC region by IANA.
Data Combined and Published
The RIPE NCC collects the files from the sites of the local collectors
using FTP. Once a complete set of files is gathered, the data is analysed
and summary statistics are published.
|