My partners and I had breakfast with John Curran, CEO of ARIN, at Interop. ARIN is the organization responsible for managing IP addresses for North America. John is on a mission to spread the word about the exhaustion of the IPV4 address space and the need for a rapid transition to IPV6.We have been hearing about this problem since 1993, when the IETF realized that the existing 32-bit address space in IPV4 would be insufficient. The IETF worked out the details for IPV6, which has a much larger address space and should allow us to address every molecule in the universe. But, as my wife says, transitions are hard. I always thought that real IPV4 address exhaustion was still out there in the future. John convinced me otherwise.
As of my writing, the current prediction for when ARIN, and its partner companies that serve the rest of the world, will run out of available IPV4 address blocks is next April 30!
ARIN recalculates the exhaustion date every day, based on a moving average of the amount of addresses requested to date. This calculator is published, and is available here:
http://www.inetcore.com/project/ipv4ec/index_en.html
Here is the calculator itself. Depending on when you are reading this posting, the date may have changed from what I show above.
So IP address exhaustion is only a year away! What does this mean? It means that anyone who wants addresses beyond that time will have to use IPV6 addresses.
Of course folks will be scrounging old unused IPV4 addresses, and there will be (is already?) a black market in IPV4 addresses. Some folks have released blocks of addresses back to ARIN, such as the DoD. But even with some big releases the address pool will soon be depleted given the current growth rate of the Internet.
Some of the most frequently accessed sites on the Internet now operate with both IPV4 and IPV6. The University of Pennsylvania (UPenn) has written a software test program to search through the top 1 million + sites as rated by Alexa, and determine how many have an IPV6 address as well as an IPV4 address. Their current results, shown here, indicate only 0.17% of the top sites are running IPV6.
So if new endpoints today were only running IPV6, they would require translation from IPV6 to IPV4 to reach 99.83% of the top 1M sites on the Internet. This IPV6 to IPV4 function is like a NAT translation which will have to be implemented by the service providers to allow the network to continue to work.
If I were a service provider, I would be worried about this transition. I see lots of opportunity for network failure, congestion and low performance until we get a good dual-mode system worked out. I don't expect people will stop using IPV4 addresses for a long time, so we will have to work with both IPV6 and IPV4 addresses simultaneously.