I admit that being an IPv6 advocate for as long as I have been can cause some self-doubt and loathing at times. IPv6 has had a longer ramp-up than expected and us IPv6 advocates have often been frustrated by the pace of adoption. The good news for me was that 2016 was a really great year for IPv6. You can check out my end of year prediction reviews and see just how much really did change for IPv6 overall.
But we are in a new year and I wanted to share what I think will change for IPv6 in 2017. In a similar format to my IPv6 predictions for 2016 I am simply stating what I think will happen.
The majority of container solutions (Docker, Kubernetes, Mesos*) will have IPv6 support by the end of 2017
IPv6 growth worldwide will again outpace the US
Major private cloud solutions (OpenStack, AzureStack, VMware) will have production ready IPv6 support
Security will finally start figuring out IPv6
Early IPv6-only data center solutions will start happening
Let’s walk through each of these in more detail and unpack what I mean.
I think 2017 will continue to see the mass adoption of containers for both developers and operators. There will likely be a period of disillusionment during 2017 (especially by those looking to push function as a service solutions) but containers and those that have solutions around containers will see continued massive growth. The container solutions around management of workloads and managing the clusters will be super important. They are effectively the next generation of operator platforms replacing VMware vCenter or OpenStack Horizon (as an interface at least). These solutions, I believe, will provide the ability to run and operate with IPv4 and/or IPv6. The reality is that Mesos already provides this capability (Mesos was written with IPv6 support from day one) but production level support for IPv6 in Docker and Kubernetes are some ways off (see Shannon McFarland's IPv6 and Docker blogs):
It shouldn’t come as a surprise to anyone that the rest of the world will start adopting IPv6 at a faster rate, and here’s why: The US is above 33% of native IPv6 services (mainly due to mobile operators) and that rate will grow steadily over 2017. But because so many countries outside the US have not had high adoption rates (with a few exceptions like Belgium) they have a much higher initial deployment growth curve to leverage. Often a single service provider enabling IPv6 for a country will cause a massive deployment of IPv6 almost overnight. I think several other countries are poised to do just that in 2017. Also, if China or Russia decide to do that it would instantly become the largest deployment increase of the year. So keep your eyes out for such announcements!
Private clouds will become more important in 2017 as more and more customers determine that an all-in public cloud strategy does not address all their business requirements or concerns. As a result, you will see an uptick in hybrid-cloud solutions that will require deployment of private clouds. These private clouds will want to be as tightly integrated with their public cloud counterparts as possible to allow for low friction utilization of both public and private cloud. AWS and Azure both have native IPv6 capabilities and I expect Google to have the same shortly. With that in mind, it just makes natural sense that the on-premises version will need to have IPv6 support too. It will take until the end of 2017 to see anything close to a production level quality deployment of IPv6 for on-premises but I do believe it will happen this year.
Because of IPv6 support from the major public cloud providers the security industry overall will be forced to more deeply embrace and start taking IPv6 more seriously. You will start seeing IPv6 specific capabilities within security product portfolio but also event correlation and matching for dual-stack hosts. Understanding the relationship between IPv4 and IPv6 and what events are happening on either protocol will become more important. Turning off IPv6 will no longer be the standard request from IT security and gaining skill and insight into what IPv6 is doing will become more common among IT stakeholders.
My crazier prediction is that we will start seeing some more cutting edge companies decide to start deploying IPv6-only data center solutions to meet the demand of the mobile market. I don’t think there will be a huge amount of them but I do believe you will start seeing companies take a very serious look into the option of doing an IPv6-only solution to meet their primary customer needs. They may adopt protocol conversion or proxy functions for IPv4 so that an IPv6-only solution can still provide resources for an IPV4-only host. At some point, as IPv6 adoption grows, it will be far more cost-effective to deploy and operate a new data center with IPv6-only compared to a dual-stack solution. I think that tipping point will happen in late 2017.
So, in summary, cloud, containers and global adoption will be the big IPv6 stories and my fingers are crossed that security and IPv6-only data center will become an important part of the 2017 story too.
You can find me on twitter as @ehorley and remember…