Penetration varies by economy. Taking 5% (btw, I disagree with this figure,
its unadjusted for population stats from the ITU and we (APNIC) get 3% but
thats another story) is masking that we have economies above 30%, and
economies on 0.01%.
I think this is a time when a single figure is actually not helping. The
net is bifurcating into a two class model: the dual stack world which is
well provisioned, and everyone else stuck behind ageing CPE and bad
investment issues
On Thu, Dec 4, 2014 at 8:30 AM, Fred Baker (fred) <fred(_at_)cisco(_dot_)com>
wrote:
On Dec 3, 2014, at 9:15 AM, Dave Crocker <dhc(_at_)dcrocker(_dot_)net> wrote:
So, after 25 years of effort, we've achieved 5% penetration. Wow.
I don’t think that’s fair.
Yes, the protocol was designed 20 years ago (draft-ietf-ipngwg-ipv6-spec
and RFC 1883 are dated in 1995), and if we were doing it today we would
separate location from identity (for multihoming reasons and in order to
not need mptcp) and perhaps enable the use of very short IIDs (to not need
6lowpan). From that perspective, we are where we are.
But serious allocation/deployment didn’t start until 2007 (
http://www.oecd.org/internet/ieconomy/44953210.pdf figure 2), Microsoft
(the largest end system vendor) didn’t have a credible IPv6 offering until
Windows 7 at the earliest (2010), and non-trivial traffic levels didn’t
happen until perhaps 2012 (14 years after 1995). By comparison, what we
today call IPv4 was well underway in 1978 (look at the header in section
3.1 of IEN 28), RFC 760 was deployed in Korea in 1982, the Internet as a
whole turned up IPv4 in January 1983, in June 1988 had 173 routes in the
route table, and didn’t see heady growth until the mid-1990’s (about 15
years after IEN 28). If you’re starting from the first publication of a
protocol specification, I’d say we’re about on track. 20 years after 1978
is roughly when IPv4 took over from IPX, DecNET, and the rest, give or take
a few years.
And that's for a single, special service provider.
Again, let’s be accurate. That’s “as measured by a single, special
services provider”. They’re not the only provider measuring it, just one
that is particularly noisy about it. Akamai says that 2Q14 was the first
quarter in their history in which the number of IPv4 addresses accessing
their service decreased, and that they are seeing significant IPv6 usage,
as do Yahoo and others.
And in all of these cases, they are not measurements of “folks having IPv6
addresses and trying to use them”, they are measurements of “folks
successfully using IPv6 end to end, through one or more ISPs, and engaging
in communications with” whoever is reporting the measurement.