Lee Howard: “In Just Four Years, Global IPv6 Penetration Will Reach 80%”

28/06/2018

While many companies and organizations do not consider IPv6 essential for growing their business, deploying this technology can require hard work and there is a chance they will only be ready once having this protocol has already become a pressing need.

With this argument, Lee Howard, one of the keynote speakers at LACNIC 29, discussed which considerations and factors should be taken into account in order to understand the urgency and relevance of moving forward with IPv6 deployment. For example, a company that might only need IPv6 addresses three years from now should begin analyzing IPv6 deployment today, as implementing this technology will most likely require practically the same amount of time.

In his dialogue with LACNIC News, based on current deployment trends, Howard estimated that in just four years 80% of the world will have IPv6. According to the expert, that will be an important milestone.

When we evaluate IPv6 adoption in a company, what considerations should we take into account beyond the technical issues?

I hear quite frequently that people need to adopt IPv6 because its adoption is critical, that it’s urgent, that the sky is falling, that the apocalypse is coming… And I don´t think that´s the case. I think every company needs to do its own evaluation of their business need for IPv6. For some companies, it may be a long time before they need to adopt IPv6. I was talking to a European bank who told me they probably won´t need to implement IPv6 for another six to ten years. But when they evaluated everything they had to do in order to get IPv6 deployed, they realized it might take them 10 years and that it was better to start now.

So, the biggest thing beyond the technical issues performing a needs assessment – it is important to figure it out how important and urgent it is, and how much work it is going to take. This will allow completing the project before it becomes necessary for business reasons.

Another important aspect that I think is not completely technical is that a lot of people are afraid to implement IPv6 because they feel it is a new technology and they haven´t been trained on it. I find this surprising, because I never took a class on IPv4 and I don´t think most people have been to a class on IPv4. People think it’s harder than it really is. There’s nothing to be afraid of.

When do expect to have 80% of IPv6 users in the United States? Why is this an important milestone?

I don’t know that 80% is a magic number. It’s a number that I use because we think of it as the 20/80 rule, also known as the Pareto principle: the bulk of the work happens by the time you get to 80%. So, it’s a milestone, but that doesn’t necessarily mean that it’s a critical milestone.

The projections I was showing used charts based on Google data with the percentage of users who access Google websites over IPv6 in each country. I like using the logistic curve because is an adoption curve. Technology adoption tends to happen slowly at first, then it speeds up and then it slows down again. So, in the United States we expect to have about 50% of IPv6 users before the end of 2018 and 80% by 2020, which is just two years away.

Worldwide, we hope to reach 50% around the end of 2019 or the beginning of 2020, and 80% in 2022. This means that in just four years, global IPv6 penetration will reach 80%.

For me, the important part is to say that I think a lot of people have been waiting to deploy IPv6 and that nobody wanted to go first. People were waiting until there was a lot of IPv6 before showing interest in deploying the protocol. We can see now that we are within a couple of years of huge deployment levels both worldwide and in several countries. In particular, I am talking about websites and content providers. I can assure you there is enough IPv6 coming to your country in the next couple of years that you should really deploy it know. It will improve your performance and you’ll be able to track your users better if you need to get back to them or maintain consistency across sessions.

From an economic point of view, how can we approach the business case for IPv6? What benefits will IPv6 bring and what investments will it require?

Not having to buy IPv4 addresses means you manage to avoid the costs. We know the prices for IPv4 addresses are rising. In some cases, I keep telling people that if they just use IPv6 they can sell some IPv4 addresses and make some money that way. That may fund their entire IPv6 deployment effort. CGN is widely deployed and the problems it creates are fairly well understood, so in many cases people don’t even think of it as being a problem.

I’ve talked about the difference in latency. Many content companies have measured an average of 15% faster speeds with lower latency when using IPv6. That seems like a competitive advantage that should be easy to advertise.

I think there’s also an argument to be made that an IPv6-only network is simpler to troubleshoot and probably simpler to run and upgrade than a dual-stack network. Therefore, it probably takes less effort to upgrade and maintain, and when something breaks it takes less time to repair. I think these are additional cost savings.

In your presentation at LACNIC 29 you showed some projections of the increasing cost of IPv4 addresses. Can you comment on that?

Over time, we’ve seen that prices have gradually increased from 7-10 to 16-17 US dollars per address. We’ve also seen that it used to be that you paid less per address when you purchased larger blocks. In other words, there was sort of a bulk discount. But that’s not the case anymore.

So, if the trend continues at a linear growth, I would expect addresses to cost 20 US dollars by the end for the year. However, in the past few months we’ve observed a few transactions at much higher prices, although I don’t know whether these are one-time anomalies or if they are indicating a new trend. If it is a new trend, then we will easily see prices at 25 dollars or more by the end of 2018 and maybe 35 or 40 by the end of 2019.

In your opinion, what would be a good strategy to generate interest among a company’s business managers?

I’ve observed companies all over the world where there’s a disconnect between the priorities of the executive staff and the priorities of the engineers. Sometimes the executives are not communicating what they really want or the vision for their strategy to the engineering staff, and sometimes the engineering staff know that they have a technical problem but don’t communicate it properly to the executives. So, what I would like to see is more engineers, especially senior engineers, knowing a little bit more about business and understanding how to make their business arguments, because I think that’s how business decisions should be made.

I believe many engineers have been scaring the population by saying that the sky is falling and that we have to hurry up and deploy IPv6, but without explaining why. I think that’s a fundamental problem. I don’t think everybody absolutely needs to have IPv6 right away. Every business needs to conduct its own assessment to figure out how urgent it is for them to deploy IPv6.

Click here to read Howard’s presentation at LACNIC 29.