Is it another Y2K? We have less than a year before we run out of allocatable Internet addresses. Does this matter? What can be done?
We have known for some time that the number of IP4 Internet addresses isn't sufficient but we have reached a tipping point. IPv4's countdown has fallen below 365 days which means that at the current rate of consumption there will be no more IP addresses to hand out in less than a year. Of the 4 billion usable IP addresses less than 6% are left free and these are about enough for the coming year.
The fact that this is going to happen can be no surprise as many Internet organisations and luminaries - Vint Cerf for one - have been warning of the coming crunch for some time. However the reasons for the problem becoming critical is that there is an ongoing explosion of web connected data - mostly due to sensor grids and web enabled devices including mobiles.
There are a number of well known ways of reusing IPv4 addresses so that they stretch to meet the demand. The most common is to simply reallocate IP addresses as users connect and disconnect using DCHP. The problem with this is that this sharing of addresses is becoming less effective as users adopt always-on devices.
The most effective method of sharing IP addresses is to use NAT - Network Address Translation - which converts multiple private IP addresses into routable public IPv4 addresses as required. This allows a private network to share a pool of IPv4 addresses for packets that are routed on the public network. It effectively increases the size fo the IPv4 address by using the port number to identify the private address that the packet belongs to. The main problem with NAT is that it only helps for small networks attached to the Internet - you still need one routable address for each home user even if they have just one machine. A better use of NAT is at the IP level where a small pool of IPv4 addresses can be shared between all of its users.
It has to be admitted that the final solution to the problem has to be the widespread adoption of IPv6. IPv4 has a 32-bit address field which provides 232 (around 4 billion) addresses which once seemed like more than enough. IPv6 has a 128-bit address fields which provides for 2128 (around 300 trillion trillion trillion) addresses which is roughly 5x1028 addresses for every person on the planet. As well as having a much larger address space, IPv6 has big advantages in terms of routing and re-configuring networks.
The change to IPv6 is mostly a matter of changing software but the changeover isn't simple and both IPv4 and IPv6 are likely to coexist for some time. Many major companies, like Google and Facebook, are making progress to IPv6, many are not. If large companies move to IPv6 then they can free up blocks of IPv4 addresses to keep things going. and avoid a possible black market.
Many commentators are referring to the coming IPv4 address crunch as another Y2K. Given that Y2K was less of a disaster than predicted we might as well hope that this is correct.
To track the time left follow: IPv4 countdown on Twitter