Iljitsch van Beijnum writes:
About 85% of the IPv6 address space is specifically left unused at this
time. And even within the 2000::/3 which is defined for global unicast
use *now* just 3/8192th is really used.
But that represents 5,192,296,858,530,000,000,000,000,000,000,000
addresses. Why do we need so many??
To put that in perspective, that's ten IP addresses for every square
_nanometer_ of the Earth's surface.
It doesn't matter if it's only 1/65536 of the total IPv6 address space;
it is still obviously too much. And that illustrates my point. Already
all these addresses are being allocated, even though it is unlikely that
every MOSFET transistor in the solar system is going to be assigned an
IP address. Virtually all of this allocated space is wasted; in fact,
even if we gave every human being on Earth one trillion IP addresses for
his own use, only 0.1 billionth of one percent of this address space
would actually be used.
Do you see the problem? IPv6 isn't even truly in place yet, and already
space is being consumed incalculably faster than any shortsighted
engineer ever dreamed. People have trouble understanding large numbers,
so they confuse them with infinity. And people have trouble
understanding the difference between a linear string of bits and
exponential series, so they think they can divide up a field of 128 bits
any way they want and yet still have some useful fraction of 2^128
addresses available. But it doesn't work that way.
What will happen next? This slice of address space will be wastefully
suballocated in the same way that the slice itself was wastefully
allocated from the total IPv6 space. Easily 99.999999999999% of all
possible addresses will become unusable, thanks to wasteful allocation
schemes. Companies, organizations, and individuals will be given
staggering amounts of address space because "there's plenty of space"
and "it'll never run out." And in no time--far sooner than any of the
clueless designers ever thought possible--it will all be gone. And then
another slice will have to be allocated. And another. And another.
The ultimate result: Instead of taking trillions of years to exhaust
the IPv6 address space, it will be consumed in a few decades. And I'll
still be around to say "I told you so."
I really have to question the intelligence of engineers sometimes when
they can make the same glaring mistake over and over, and even with
careful explanation they still do not see it.
From a cognitive standpoint, this may be the same sort of failure to
understand that underlies many assumptions concerning national debt and
budget.