When people first start to contemplate IPv6, they tend to get the false impression that they must make a full migration from using IPv4 exclusively to using IPv6 exclusively. They believe that their organization has to make a complete “flip-the-switch” cutover from using one protocol to the other. It turns out that this is not the case. The IETF knew early-on that the migration from using IPv4 to IPv6 would take time and, therefore, they created a variety of transition mechanisms (RFC 4213). There were manual and automatic/dynamic tunneling and translation mechanisms created, however, the predominant migration strategy was a dual-stack implementation. This is where both protocols are run in parallel for some time and then, eventually, IPv4 might be turned off at some point. Organizations will make their systems “bilingual” by adding IPv6 to the existing IPv4 network, run both protocols for many years then eventually start to pare back their IPv4 network, eventually leaving an IPv6-only network.
IPv4 Address Exhaustion Has Occurred
Meanwhile as organizations work on deploying IPv6 onto their networked environments, IPv4 address depletion is continuing. The Regional Internet Registries (RIRs) have differing policies about how each of them has managed their finite IPv4 address resources and handled address allocations. Geoff Huston’s “RIR IPv4 Address Run-Down Model” graph has been referenced frequently over the past decade. APNIC and RIPE are already in an IPv4 address depletion situation. The LACNIC region has experienced tremendous Internet growth and has exhausted their supply of addresses. It is expected that ARIN run-out will occur sometime middle of this year. AFRINIC still has a supply of addresses that should satisfy demand for a few more years.
IPv6 Deployment Continues
The good news is that the use of IPv6 has been growing substantially over the past five years. One look at the Google IPv6 Statistics – IPv6 Adoption graph shows the use of IPv6 from their vantage point. From this graph, it appears that IPv6 usage has been doubling every year. However, the vast majority of Internet traffic is still IPv4 and the vast majority of the Alexa 1M still use IPv4 only. The Hurricane Electric “Global IPv6 Deployment Progress Report” states that while 19% of ASNs are advertising IPv6 routes, only 6% to 7% of web sites are running both IP protocols. From this data, we can conclude that if anyone wants to access the Internet, they will be unable to reach most destinations using an IPv6-only configuration.
No Plans to Turn Off IPv4
Internet Service Providers and large enterprises are struggling with how they will continue their IPv4 operations for the indefinite future. The vast majority of organizations will need to continue their IPv4 usage for the next ten to twenty years. There are many IPv4-only systems within buildings and data-centers that will not be aged-out rapidly. Virtually all organizations have no current plans to disable IPv4. Therefore, organizations should ask themselves: Do we have enough IPv4 addresses to sustain our business for the next ten years? How do you even know what networking will be like in ten years from now?
Feeling the IPv4 Address Crunch
There are many organizations that are running out of public IPv4 addresses and are starting to feel the crunch. Some very-large organizations are even running out of private addresses (RFC 1918) inside of their NATed environments. It is not uncommon for organizations to divide up their 10.0.0.0/8-space into /28s and smaller subnets. IPv4 prefix lengths are getting smaller and smaller. If your organization is either running out of public IPv4 addresses or starting to run out of private IPv4 address space, you are already behind the proverbial 8-ball. Most service providers will have no choice but to use some form of Carrier-Grade NAT/Large-Scale NAT (CGN/LSN) in the coming years, even though it may negatively impact applications (RFC 7021). As Lee Howard and Wesley George from Time Warner Cable have said, “Therefore the best method to reduce the cost of Carrier-Grade NAT (CGN) deployment is to work to dep...”
IPv6 Deployment Won’t Help Your Near-Term IPv4 Crisis
All organizations should come to the conclusion that they must act quickly to deploy IPv6. The faster they deploy IPv6, the sooner they might be able to switch some systems to using IPv6-only. Any system that can be migrated to IPv6-only will have a benefit of freeing up an IPv4 address. For example, moving cable modem management to IPv6-only saves cable broadband providers IPv4 addresses that could be used for other purposes.
It is conceivable that an organizations can try to aim to hit the sweet spot on the graph where IPv4 connectivity starts to get really bad, and IPv6 deployment/implementation is starting to become more ubiquitous. Aiming for that point is a bit like steering the Millennium Falcon through a narrow gap between two asteroids.
For most organizations, the slow rate with which they are deploying IPv6 will initially have no effect on their increasing short-term need for IPv4 addresses. Their systems will need to be deployed using both IPv4 and IPv6 to maintain dual-protocol connectivity and facilitate connectivity to the largest population of clients. Failure to deploy IPv6 will only make things worse and only delay the inevitable. Those organizations that are on the tail-end of the bell curve of companies deploying IPv6 will experience fear, anger, hate, suffering, and eventually, the dark side.
The best approach is to aggressively deploy IPv6 in an effort to try to reduce the duration your organization spends running two Internet Protocols. As more IPv6-capable systems come online, your organization will already be ready to take advantage of that. Being on the leading edge of the bell curve of companies deploying IPv6 puts your organization in an advantageous position that reduces operating expenses and your dependence on IPv4 sooner.