[arin-ppml] Just a reminder of some quick mathematics for IPv4 thatshows the long term impossibility of it
tedm at ipinc.net
Thu May 12 18:36:33 EDT 2011
On 5/12/2011 1:44 PM, Mike Burns wrote:
> Hi Ted,
> I don't think anybody made the claim that IPv4 had enough unique
> addresses to give one to each person on the planet.
> Nor was there a claim that the Invisible Hand would create any address
> However, there is enough space to give everybody their own unique IPv4
> address + TCP port.
> Even their own thousand unique IPv4 address+TCP ports.
> 3.8billion times 65,000 is a very large number.
> Food for thought?
The stats I posted show penetration of existing IPv4 into the
population. It is not a 1-for-1 mapping of the IP address space
You can look at headcount, look at total number of IPv4 in use and
figure out what the ratio of IPv4 numbers to each person is. As with
any math formula, if you change a variable, things happen.
For example if you decrease the ratio of IP addresses per person
(which is what NAT does) then you can increase the population that
is using the same amount of IP addressing. That is what we did 10
years ago and it worked.
If you add a new way to use IP addressing that allowed for
port numbers on a single IP to be spread around, then you could
further decrease the ratio of IP addresses per person. That is
what co-called "carrier based" nat or NAT4444, or "stacked nat"
(translators behind translators, behind translators) essentially
Notice though that each iteration of that trick increases complexity
and increased complexity decreases reliability. That is why NAT is
regarded as played out, we have hit the area of diminishing returns
with it. Also notice that NAT tends to push servers from the "leaves"
of the Internet to the "trunk" So instead of Joe Blow Company
running his webserver off a process on his regular desktop that he
runs his company on, now he has to pay someone to do it who is
closer to the trunk. Same for e-mail and other services. That
isn't what attracted people to the Internet in the first place.
So you might say that the popularity of the Internet is putting
pressure on promoting solutions that run contrary to the fundamental
properties of the Internet that got it popular to begin with.
Much like when you build a highway, it is empty, people buy a car
and drive on it, it is enjoyable to do so because there is no
congestion, so now everyone buys a car, they all get on the highway
and now it's congested and unpleasant to drive on, thus you have
killed the golden goose that laid the egg that everyone originally
> ----- Original Message ----- From: "Ted Mittelstaedt" <tedm at ipinc.net>
> To: <arin-ppml at arin.net>
> Sent: Thursday, May 12, 2011 4:23 PM
> Subject: [arin-ppml] Just a reminder of some quick mathematics for IPv4
> thatshows the long term impossibility of it
>> From the following website:
>> Notice - as of 2009/2010 (last time these figures were updated):
>> WORLD INTERNET USAGE AND POPULATION STATISTICS WORLD TOTAL
>> Population Internet Users Latest Data Penetration (% Population)
>> 6,845,609,960 1,966,514,816 28.7 %
>> From the beginning of public access of the Internet in 1995 we
>> are now about a year away from complete global IPv4 runout. Assuming
>> maybe by then we are at a global penetration of 35% of maybe 7 billion
>> people, that's still only about 2.5 billion people on the Internet.
>> So how exactly do we get the other 4.5 billion people on the Internet
>> using IPv4?
>> The last "legacy" IP block was handed out in 1997. To assume that
>> unused legacy IPv4 will cover the remaining 4.5 billion people assumes
>> that 65% of the usable IPv4 space was handed out before 1997 and the
>> remaining 35% of it was handed out from 1998 until next year. It is
>> a ridiculous assumption completely unsupported by the math.
>> But, don't let something like mathematics bog your day down! It's much
>> more fun to believe that Adam Smith's "Invisible hand" will come sailing
>> in at the last minute and manufacture IPv4 out of thin air! ;-)
>> You are receiving this message because you are subscribed to
>> the ARIN Public Policy Mailing List (ARIN-PPML at arin.net).
>> Unsubscribe or manage your mailing list subscription at:
>> Please contact info at arin.net if you experience any issues.
More information about the ARIN-PPML