It is a well established fact that we will run out of IPv4 addresses in the next few years. I know that people are tired of hearing it - its been the IPv6 mantra for year. “Get IPv6 because the IPv4 sky is falling!!” Even when the IPv6 advocacy shifted towards a more business oriented view of IPv6, the underlying principle that IPv6 was created to address (no pun intended) a future depletion of the IPv4 address pool remained valid.
So here we sit at the beginning of the 4th quarter of 2008. Projections on when the IANA IPv4 address pool will be exhausted vary. If you sit the Geoff Huston camp, then exhaustion time frame stretches from October 2009 out to July 2012. If you are interested in a more detailed discussion on the matter, I’d encourage you to peruse Geoff’s most recent posting. Mind you, this is from IANA to the Regional Internet Registries (RIR). The RIRs will then distribute those resources down to their members - ISPs and organizations that can make the justification for the v4 resources - until their resources are exhausted.
Stephan Lagerholm of Infoweapons also just released a study of IPv4 address exhaustion. In his statistical modeling, he predicts the IANA pool will be exhausted on August 9, 2010, with RIR exhaustion occurring on August 19, 2011. You can find Stephan’s paper here. The paper is good, if a bit heavy on the mathematics used to derive the exhaustion dates. One item that he approaches that I don’t recall seeing in Geoff’s paper is discussion about “tainted blocks” - a reference to IPv4 /8 address blocks that have not been allocated by IANA, but for various reasons will likely create more problems than they solve if released to the world for use in the DFZ (Default Free Zone).
The real issue here is what will happen once that pool is exhausted? Global Geothermal Nuclear War? Well, not quite - but a fair amount of chaos, privateering, black markets, legal battles over address space, and new and improved complexity added to the global infrastructure - like carrier grade NATs. All-in-all, it should be an amusing period with all the charms of the Dot.Com implosion, crushing innovation and forward momentum in IT for a few more years.
Interestingly enough, the solution to this problem is the same one that was invented in 1995, when IPv6 was first standardized. Being an IPv6 advocate I realize this sounds self serving, but both rationally and technologically speaking it makes sense. IPv6 is not new. It’s been around for over a decade now. Despite being battered constantly by market naysayers, proverbial technology ostriches, and vendors who didn’t want to spend the cycles to integrate the technology, IPv6 is still here. Integration may be challenging - even more so now than it would be if you had started 3 years ago.
One of the biggest push backs on IPv6 adoption has always been the cost. But there is a mounting pile of evidence that suggest the cost for the alternative - fighting to keep IPv4 alive. Here is what we’ll see over the next few years:
1. IPv4 resources become more costly as they become more scarce – basic supply/demand economics. That increase cost will flow to the users or to the overall costs of running a network.
2. More complexity to the network. Carrier-grade NATs cost money to deploy and maintain and will have to be integrated with billing systems (oh yeah!). Services and applications will have to work through nested NAT infrastructure – meaning more expensive development and support costs.
3. Breakage. Some stuff just doesn’t work well when two endpoints can’t talk to each other directly. Continued reliance on a 1970’s technology base will inhibit the deployment of new services and technologies that will help drive down costs and prohibit the cost effective deployment of new services, which means less ARPU.
4. IPv6 adoption costs increase. For those that wait until the last minute and have to do a “quick” move over will find themselves facing exorbitant costs for equipment and technical know-how. Systems that were just deployed on the “old” network weren’t vetted for compatibility in IPv6-enabled networks.
Even though IPv4 address exhaustion has never been particularly sexy, access to IP numbers is a critical function of making the network run. And IPv6 integration is no different than any other technology transition – say IPX to IP or from ATM to MPLS – it requires thought, planning, and implementation over time. When it comes to IPv6 though, some geographies are stuck at step one – thought. In the meantime, the sky continues to fall at an increasing rate of speed.