Reply to comment:

Stéphane Bortzmeyer
Note there is also a interesting discussion about this article on the mailing list of the RIPE IPv6 Working Group http://www.ripe.net/ripe/mail/archives/ipv6-wg/2013-June/002153.html