[ipv6-hackathon] Project Proposal

Hello participants,

As promised, where are details to the project proposal! (It's also on
our etherpad, so you can read all projects nicely collected in one
place: https://pad.riseup.net/p/ripe_ncc-ipv6-hackathon)

Drivers for IPv6 Deployment

>From organizations that successfully implemented IPv6 we know that IPv6
is technical feasible and practically nobody questions that IPv6 is
going to be
the future of the Internet. Yet, today's Internet is run on IPv4 and
companies seem to be reluctant to do the transition.
In this project I suggest to take a data-driven approach to identify the
factors that encourage (respectively discourage) companies to use IPv6.
Since we will have only a limited time to work on it, I'd like to
bootstrap this project with the recent work I did on this topic.

In a nutshell, I introduced a metric that measures the IPv6 readiness of
a company based on routing (BGP) data. The data crunching has been
carried out for all
~80,000 global ASNs for the past 10 years and is readily available for
the team. The analysis was done on global, regional (RIR), national and
industry level.

The project offers multiple, interdisciplinary items to work on but the
main ones are:

a) Produce reports based on the data:
Analyze and interpret global, regional and/or national developments.
This involves finding correlations with corresponding events.
No programming skills required.

b) Improve the metric
The current metric associates IPv4 addresses with /48 in IPv6. This is
very simple and does not take the specific requirements of an organization
into account.

c) Improve and extend ASN/organization to industry mapping
The analysis on an industry level required a data set that mapped an
organization to an industry type, which was done through the parsing and
analysis
of 22,000 company webpages for the RIPE NCC region. The scraping of
webpages as well as the parsing require improving as the current
precision is
roughly 30%. The outcome would not only be useful for this topic but
publishing this data set would benefit a whole group of Internet
researchers. (Your
name will be of course published along with the data set :)!
This item would benefit of a multidisciplinary team: analytical, natural
language, web development and data mining skills.


All the best,
Christian