To help you out, 6connect COO, Pete Sclafani, and one of our esteemed guest editors, Ed Horley from Hexabuild, are here to talk shop. Continue below to see some common myths about IPv6 and what the reality is – so without further ado, let’s get started!
Myth #1: IPv6 is “coming.”
Pete: Ed – we have run across this one way too often. Service Providers get it and at least are making concerted efforts and progress on their IPv6 deployments, but Enterprises and even Cloud Service Providers have been dragging their feet. What are you seeing in the field?
Ed: Enterprises still lag substantially in IPv6 adoption, there is no question about that. I anticipate as they watch the rest of the market move more and more to IPv6 their wait on at least providing Internet edge services on IPv6 will change.
However, Cloud Service Providers have been working through their support requirements for IPv6 and making progress. You can deploy a full dual-stack VPC in AWS today (https://docs.aws.amazon.com/vpc/latest/userguide/vpc-migrate-ipv6.html), along with load balance, CDN, DNS and DirectConnect services. Azure has released its IPv6 support in VNETs (https://docs.microsoft.com/en-us/azure/virtual-network/ipv6-overview) and will have support in ExpressRoute soon. Oddly enough, Google Cloud seems to be the furthest behind in IPv6 support (https://cloud.google.com/vpc/docs/vpc) even though traditionally Google has had strong IPv6 support for all their other products. I imagine this will change as their cloud matures but it is unfortunate they do not have better support.
Pete: Agreed, there seems to be an “IPv6 is coming” mentality still – but the data says that IPv6 has arrived if you look at the consumer device sector. When someone like Comcast has deployed IPv6 over 95% of their network, saying “IPv6 is on its way” doesn’t seem to acknowledge reality.
Today, virtually all Internet-connected mobile and end-user devices on the market support IPv6, and more and more ISPs are deploying IPv6 in their production networks and services. Increasing numbers of content providers are making content available over IPv6, as well – Google, YouTube, Facebook, and Netflix are some examples of big players who now boast full IPv6 capability.
A few other noteworthy IPv6 facts and stats:
- Comcast has an IPv6 deployment measurement of over 71% (source).
- Global IPv6 traffic has grown more than 5000% since 2012.
- Nearly 30% of Alexa Top 500 servers now support IPv6 (data updated daily here).
- India currently has the highest IPv6 adoption rate, at over 58% according to Akamai.
- India has made the biggest strides in overall IPv6 adoption this past year, in large part “because mobile” (PDF report).
- 2019 has seen a monthly average of 64 IPv6 address allocation and assignment requests in North America alone – an increase of 33% if the trend continues, as reported by HexaBuild (PDF).
- T-Mobile, one of the first mobile providers to embrace IPv6, is nearing 100% adoption.
Myth #2: We don’t need IPv6.
Pete: I will admit – the current state of most networks makes this myth a little easier to believe. If everything still “works”, what does IPv6 bring to the table? We all hear about IPv4 runout, but what are the ramifications? Ed, what’s your gut response to this one?
Ed: I think the framing of this question is important, especially for organizations and companies that have an abundance of IPv4 addresses due to their early entry into the Internet. This is often Universities or larger Enterprise technology companies that received large IPv4 allocations, often before we had better Internet governance and fairer allocation planning. Many of these companies state they don’t need IPv6 because they have plenty of IPv4 addresses to meet their needs today and well into the future. While it is true they have met their IPv4 needs and they will likely not have any structure issues meeting their addressing needs anytime soon they are mistaken in thinking they need no IPv6 at all.
The core of the issue isn’t how much address space they continue to hold and use for themselves. It is the lack of IPv4 that the rest of the world has that makes their local abundance less useful over time. Other countries and institutions around the world who were not so fortunate are now at the point of not being able to acquire the IPv4 address space they need to build businesses, get online and build out Internet services and infrastructure. Because IPv6 is their only option, they will adopt and use it to grow and expand their network to suit their needs. They will not dual-stack and they may choose to skip providing any transition technologies for legacy IPv4 protocol users since it simply introduces costs, complexity and support problems. This means it is up to the legacy IPv4 only networks to decide if they wish to participate in the entire Internet or only a small subset.
If they stay on IPv4-only then their audience and participation and impact will diminish over time. It won’t be overnight but it will happen more quickly then they anticipate and likely by the time it is identified as an issue they will not have enough time to deploy IPv6 in a constructive way and they will lose a major competitive technology advantage. Those that do not plan ahead will become less relevant as the Internet moves on. Let’s look at an example: India now has approximately 60% IPv6 adoption, mainly due to Reliance Jio introducing a new mobile service (growth over the last 3 years has been from close to 0% up to 60%) and with a population of approximately 1.3B people that means approximately 780M people are actively participating via IPv6. Yes, this is not the US or Western Europe as a market, but it is a rapidly growing and expanding market that will become more important over time. China and other East Asia economies are in similar situations and you will continue to see them rapidly adopt IPv6 for this core reason, there is no other option. If the US and Western Europe do not feel adopting IPv6 to engage with the rest of the world is important then their influence and position will deteriorate over time. IPv6 is the only viable option we have to solve the lack of addresses on a worldwide scale. The protocol isn’t perfect but it is what we have and it clearly works so I think this articulates why we need IPv6.
Pete: This makes sense and highlights that there is still an attitude that IPv6 is a “work in progress” so there’s no need to pay attention to it yet. But with deployment where it is and plenty of production networks leveraging IPv6 around the world, this is a big shift that early adopters are going to capitalize on – thanks for pointing that out.
As we discussed in a previous post, the IPv6 protocol has been endorsed by all Internet technical standards bodies, who view it as the standards-based solution to the IPv4 address shortage. The inevitability of a worldwide transition to IPv6 means that immediate adoption is strongly recommended by virtually every major authority.
So ignoring the situation is untenable. If you remain IPv4-only, you will eventually run into serious problems such as lack of address space to meet your needs, the inability to migrate to IPv6 once it becomes necessary, a loss of connectivity with the Internet, and being locked out of possibilities for innovation and competitive edge. There is no “last-mover” advantage here – neglecting IPv6 has no benefit and plenty of risk.
Myth #3: IPv6 replaces IPv4.
Pete: Ugh – this one is tough as it requires context. Technically IPv6 could replace IPv4, but that’s not something that should be an immediate concern and should definitely not hold up IPv6 adoption. The reality is that IPv6 and IPv4 are supposed to coexist in a normal way (dual-stack) and it should not be a one or the other choice. That being said, with environments where IPv4 runout is more advanced, IPv6 Only networks are just going to be a reality. Ed, for environments where there is confusion about “IP replacement”, what’s your go-to response?
Ed: The reality is, over the long term we are supposed to end up in an IPv6-only world with IPv4 running as a service (over the top) or in isolated islands for legacy gear that can’t support IPv6. It is the only way to address the issue of IPv4 address exhaustion for public IPv4 space. It is true that for many internal networks they can continue to operate and run IPv4 and IPv6 (dual-stack) but the operational burden of running both protocols plus the troubleshooting and support costs over time will push many to want to single stack their networks again. In the near term (next several years) I continue to think that most of our deployments will be dual-stack in nature but a few customers we are working with are seriously looking at IPv6-only and how to get there. A great example of this is what Microsoft is doing.
Pete: This makes good sense – designing for an IPv6-only world doesn’t mean you need to overanalyze at every level and ignore IPv6. This just seems like a recipe for disaster since when migration keeps getting pushed further out, it just makes the process harder when it does happen. It will be interesting to see when we hit the “Let’s go IPv6-only” inflection point and the legacy applications running IPv4-only are thrust into the spotlight. I have a feeling CFOs will see the additional cost for the legacy architecture as a driver to migrate it to service/application/stack that supports IPv6.
Myth #4: IPv6 is expensive to implement.
Pete: Another context one. While in our experience there are definitely costs that can come up on the hardware side, we see a big need for training of engineers and non-engineers, in addition to programmers/developers that may not be aware of the implications (security, compatibility or otherwise). Ed, with your time in the field, you can definitely point to costs as an obvious objection, how would you respond to this concern?
Ed: I think it is important to understand the other side of that equation, how much will it cost you to NOT deploy IPv6. If you don’t understand your business, your potential market and what audience you need to address you have no idea if your lost opportunity costs are actually higher than the effort to adopt and support IPv6 (dual-stack or IPv6-only). The challenge for most technical teams in a company is that they don’t have access or the experience to do that sort of analysis. Typically a company will hire business or market analysts to help figure out that question. Framing the discussion with your financial and business colleagues to get them to help put the entire picture together is important. From a technology perspective, it is critical that you understand what is happening in the market and why IPv6 is a key technology to adopt. You may discover, for instance, that your next big market expansion will be in India, which makes it easy to argue that (at a minimum) your website, email and other Internet-facing services should be dual-stacked. Over time, adoption and migration will grow out of necessity but a multi-phased and budgeted approach should make the adoption more palatable for many organizations.
Pete: I like this approach to responding to the objection. This also helps immensely when speaking to non-technical audiences. With some simple re-framing, it can connect with a much larger audience inside a company where technical roles aren’t the majority.
Myth #5: IPv6 services are more costly.
Pete: This one is my favorite – we have heard that since IPv6 services can require education, hardware, and other resources, that the cost is just too high. What is just starting to hit home is that with IPv6 deployments already out there, the clock is running on IPv4 service costs and fees. Chances are your static IPv4 address costs have gone up over the years, it would stand to reason that these legacy services are going to become a premium feature (i.e. extra cost) to encourage adoption of newer IPv6 services. Ed, when it comes down to pure cost arguments, how do you educate someone on the benefits of IPv6 for their current infrastructure, but their future plans as well?
Ed: The good news is that often investments made over the last few years in hardware, software and licenses will continue forward with IPv6. Many modern operating systems, network gear, and even Internet services have IPv6 support. This means that the cost to implement, adopt or consume those services as IPv6 really doesn’t change substantially for many organizations. This may not be the case for service providers or large multinational companies who may have older equipment they must support in which case building transition technology solutions will have a cost that is not associated with IPv4 at all. But these are normally one time fixed costs to address their needs and are often a catalyst for upgrading or replacing legacy solutions. If you can pivot your spending from a transition technology solution to a replacement or upgraded service that supports IPv6 then the issue has been addressed.
Pete: Agreed – In fact, by putting some effort into bringing your IPv6 implementation forward, you will be set up to save money as IPv4 service pricing goes up. The American Registry of Internet Numbers (ARIN) has stated that there are now “extra costs associated with staying IPv4-only, which will likely increase over time.”
As IPv6 deployment steadily increases worldwide, service providers naturally have an incentive to charge a premium for customers who avoid the new protocol. Just to give an example, one UK-based ISP advertises that “IPv6 connectivity comes as standard,” whereas “IPv4 connectivity is an optional extra.” Various organizations (and governments) will even offer discounted or free IPv6 resources simply to encourage adoption.
The bottom line: your network costs will increase if your services are not IPv6 compatible. So the real question is, can you afford NOT to implement IPv6? Think of it as an investment rather than a cost, and bring your business into the future!