2-node-supercomputer.net 2021-08-27
Recently, there was a discussion on HN about IPv6 adoption.
The main driver behind IPv6 is the increased address space, from 32 bits in IPv4 to 128 bits in IPv6.
And, as everybody should know, we have run out of IPv4 addresses, and only a combination of IPv6 in Asia and various tricks like NAT keep the internet running.
A common way to visualize the number of IPv6 addresses available is to compare it to the number of stars in the observable universe. How is that intuitive? Idk. Or relative to the number of grains of sand on the Earth? You know how many there are? Me neither.
Better comparisons use the number of people. However, why would data centers and super computers be limited by the number of people on the planet? Can we have something less fluid to compare to?
Here is my take: How many IP addresses are there for every square kilometer on the surface of the Earth?
The calculation is relatively simple. The area of the Earth is 4πr². The number of IPv4 addresses is 2³². Therefore, there are about 8.4 IPv4 addresses per km².
However, 70% of the Earth’s surface is ocean, so really about 28 addresses per km² of land.
But only maybe a third — give or take — is arable, so we might as well say about 100 addresses per km² on the liveable part of the Earth.
Compare that to IPv6 addresses. There, including oceans and all, we have about 6.6×10¹⁷ addresses per mm². Yes, millimeter squared.
For IPv6 it even makes sense to see how many IPv6 addresses are available for every mm³ throughout the Earth: about 10⁸ per mm³.
Should last a while. No?
The cynic realist might say that address ranges will be given away too freely, so that we will be in the same conundrum again soon enough.