IPv6 Adoption: Myths and Realities

The Internet and all of the world’s data networks are undergoing a fundamental change: The IPv4 protocol is slowly being phased out, while its younger sibling, IPv6, is gradually being adopted. In fact, you could say that IPv6 is not even all that young any more—in human terms, it’s legally an “adult” now! Despite this, however, the rate of adoption has been very slow—so slow, in fact, that many are concerned that IPv6 presents challenges too great to overcome efficiently, making the transition prohibitive, at least for the time being.

Such concern is not without merit. The IPv6 protocol was first officially published in 1995. Seventeen years later, June 6, 2012 was declared IPv6 Launch Day by The Internet Society, a day meant to officially inaugurate the beginning of the transition of the Internet to IPv6. However, over a decade later, the adoption rate of IPv6 at the time of writing is just shy of 40%, as measured by Google based on the percentage of users who access its services using IPv6.

Considering this timeline, many believe, and even advocate, that the adoption of IPv6 is plagued with problems that cannot or should not be overcome simply because it’s not worth the effort. There are several implied obstacles that have often been quoted as reasons not to adopt IPv6. These include the claims that IPv6 is slower, less secure, and much more expensive to deploy than IPv4.

These and other “challenges” are actually misconceptions that have inappropriately discouraged IPv6 adoption and hindered its progress. In this article, we’ll take a deeper look at some of the most often cited misconceptions of this type to ensure that you won’t be taken in by their seemingly truthful appearance. 

Summary of key IPv6 adoption concepts

The following table summarizes the key concepts that will be discussed in the article.

The current state of IPv6 adoption

There is no question that IPv6 adoption is proceeding at a very slow rate. It has been over 25 years since its official publication, and we still haven’t broken 40% worldwide adoption. Telecom technology advancements have developed at breakneck speeds, especially over the past decade. Comparatively speaking, IPv6 adoption is taking an eternity!

You may even be surprised to find out that the adoption rate is still relatively low in what many would consider more advanced countries, with Japan at 47%, the US at 46%, Finland and the UK at 45%, Brazil at 44%, Canada at 34%, and Australia at 31%. Among the most successful IPv6 adopters are Saudi Arabia and Greece at 61%, Germany at 65%, India at almost 68%, and France topping it all off at over 74%. For more up-to-date statistics, take a look at Google’s IPv6 adoption results by country.

But why is IPv6 adoption so slow? We’ll give you the real answer to that question at the end of the article. For now, let’s take a look at some of the most common misconceptions that are cited as reasons for its slow adoption.

Misconceptions about challenges to IPv6 adoption

Some of the most commonly quoted erroneous reasons for the slow adoption rate of IPv6 include the following:

• IPv6 is not yet a mature protocol and still has some bugs that need fixing.
• IPv6 is inherently slower than IPv4.
• IPv6 is less secure than IPv4.
• IPv6 is more complicated and requires extensive retraining of technical staff for use.
• IPv6 is very expensive to deploy.

In the following sections, we’ll explain the reasoning behind each of these views and refute them.

Misconception: IPv6 is not a mature protocol

If the maturity of a protocol were measured by its age in years, then we have already refuted this fallacy. IPv6 was officially published in RFC 1883 in December 1995. That makes IPv6 a fully-fledged “adult” by human standards.

Of course, age alone does not ensure that a protocol is mature: it must also have undergone several stages of rigorous testing as well as several iterations of modifications and improvements. It must also have been deployed as a testbed for several years in extensive networks to prove its stability and maturity.

Of course, IPv6 has already undergone extensive testing and modification. IPv6 has gone through all of the officially prescribed development stages outlined by the Internet Engineering Task Force (IETF), the standards organization that defines and ratifies standards for the Internet, and in particular, the TCP/IP protocol suite. Hundreds of standards documents have been created to define and refine the protocol.

The following diagram depicts IPv6’s standardization journey, with key events and dates indicated.

In July 2017, all previous RFCs were superseded by RFC 8200, which elevated IPv6 to “Internet Standard,” the highest maturity level for IETF protocols.

Beyond being ratified by the IETF, IPv6 has been extensively used in production networks and large-scale network service providers. Remember IPv6 Launch Day? Well, on that day, Facebook, Google, Yahoo, Akamai Technologies, and Limelight Networks decided to blaze a trail by permanently enabling IPv6 for their network services. More than 400 additional companies followed suit, including some of the most heavily accessed destinations on the Internet, like YouTube, the BBC, Vonage, and Cisco, to name a few. Since then, IPv6 has been continuously operating on these networks, representing over a decade of opportunity to identify and fix potential problems that may have shown themselves over such extensive and heavy use. 

IPv6 is an unequivocally mature protocol both from a standardization viewpoint as well as an experiential one.

Misconception: IPv6 is slower than IPv4

The misconception that IPv6 is slower than IPv4 comes primarily from the fact that IPv6 uses addresses that are 128 bits in length compared to IPv4’s 32 bits. That means that the IPv6 headers are much larger than those of IPv4. Take a look at a comparison of the two headers to visualize this difference:

This diagram suggests two primary reasons why someone might assume that IPv6 would be slower than IPv4.

The first involves network overhead. A typical IPv6 header is 40 bytes in size, while a standard IPv4 header is 20 bytes. That means that every packet sent by IPv6 includes an extra 20 bytes of data compared to IPv4. Only 20 bytes may sound trivial, but consider that this is added to all packets. That’s a lot of overhead. Add to this the fact that IPv6 may add one or more extension headers, resulting in an even larger header and more overhead.

The second involves the CPU resources of IPv6 hosts and network devices. The larger IPv6 addresses and resulting larger headers mean that more CPU cycles will be necessary to decapsulate, process, and encapsulate IPv6 packets compared to IPv4. Once again, taking into account extension headers, CPU resources are taxed even more.

Now, these concerns are understandable, and they do have some merit. However, remember that a one-time increase of 20 bytes in the header is quickly offset by continually increasing network capacity and bandwidth in all parts of the Internet, including the core and all the way up to the access portions of the network.

The same is true of CPU power. Moore’s law describes how CPU resources continue to grow substantially and will continue to grow over time, quickly minimizing any disadvantage that the larger address or header sizes introduce.

Now having said all of that, strictly speaking, does IPv6 introduce more overhead than IPv4 in the header? Well, yes: you can’t get away from the fact that you need more bits to represent the larger address space. Examining just the header sizes in the absence of all other factors does give IPv4 the upper hand in this comparison. 

However, many other factors also come into play, and one of those is the actual practical experience from using these protocols in live production networks and benchmarking their performance. Let’s take a look at some measurements of IPv6’s speed: 

• Sucuri ran some tests on these two protocols several years ago and concluded that there is no measurable difference in their performance.
• At around the same time, Facebook ran extensive tests on its native IPv6 network that concluded that IPv6 actually delivers 10-15% faster access to users compared to IPv4.
• Akamai used a different approach and tested a single URL on the same iPhone connected to the same mobile network and found that IPv6 had a median load time 5% faster than IPv4.

The truth is that there are many variables that will affect speed, so it is difficult to compare performance unless you are executing very tightly controlled experiments. Many experts have attributed IPv6 sometimes outperforming IPv4 in speed tests to the fact that IPv6 no longer requires the use of NAT, which itself adds latency to any communication.  Others suggest that more simplified routing mechanisms make IPv6 routing lookups faster,  Still others suggest that there are fewer IPv6 routers out there, so there are simply fewer hops to get to any IPv6 destination.  However, these are just speculation, and require a more rigorous set of experiments to come to any solid conclusion.

Based on the above, there is no experientially perceivable difference in performance between IPv4 and IPv6. If there was in the past, it has quickly been reduced to insignificance by vast improvements in technology. In fact, IPv6 may actually be faster in some cases.

Misconception: IPv6 is less secure than IPv4

Security concerns are always of utmost importance, so examining the security of IPv6 is an important part of adopting this protocol. There is some uneasiness regarding the security of IPv6, but this is often simply a result of the fact that IPv6 is relatively new. 

IPv6 uses the same security mechanism as IPv4: IPsec, which was developed as the primary method of securing IP communications. In IPv4, IPsec is implemented as an additional mechanism that functions on top of IPv4, whereas IPv6 has integrated IPsec into its extension headers, allowing for a more elegant implementation. However, as explained more thoroughly in our IPv6 security article, IPv4 and IPv6 have the same level of security in all respects simply because it is the same security mechanism being used.

Now, having said that, there is one area of security that must be pointed out concerning IPv6 adoption. The various transition mechanisms that are used to allow a dual-stack arrangement where both IPv4 and IPv6 coexist can be a security risk. This does not mean that IPv6 is inherently insecure, however, just that the transition mechanisms can potentially increase the attack surface of a network. For this reason, care must be taken to ensure safely implemented dual stack environments; if put in place correctly, these environments are as safe as running either IPv4 or IPv6 alone.

An excellent resource to deal with these particular threats is a recently published NSA document called IPv6 Security Guidance.

IPv6 is just as secure as IPv4 because both use the same heavily tested IPsec suite for security. Dual stack arrangements, when implemented according to well established best practices, pose no additional security risks.

Misconception: IPv6 requires extensive staff retraining 

Network professionals who learned IPv4 so well that they can do subnetting in their sleep may find it disconcerting to initially look at IPv6 addresses, which seem foreign and indecipherable. Aside from the addresses, there are also other aspects of the new protocol that differ in important ways from its predecessor; one example, how IPsec is implemented, was mentioned above.

The transition from IPv4 to IPv6 has been considered by many to present a very steep learning curve for network admins and techs. Of course, every new technology does require some retraining, and in the telecommunications industry, retraining is part of the job.

However, training a network administrator or engineer in IPv6 is not as difficult as it originally may seem. While there are obviously some differences, IPv6 uses many of the same principles as IPv4 and has been engineered, based on the experience gained from decades of IPv4 usage, to be more elegant and streamlined in its configuration and implementation. Indeed, many administrators who have spent the time to learn IPv6 have actually declared that it is more intuitive and much easier to work with. This is primarily because no more complex subnetting needs to be performed, and the need for the often cumbersome and troublesome NAT is eliminated. It only requires a little bit of time and effort to be committed to learning and implementing the newer protocol.

We’re not suggesting here that learning IPv6 is trivial.  However, the learning curve for IPv6 is not as steep as some may perceive it to be.  It can easily fit into the training regimen that is part of any job in telecommunications. 

Misconception: IPv6 is very expensive to deploy

IPv6 is actually not that expensive to deploy because virtually all network equipment and end devices already support IPv6, and this has been true for at least a decade. Unless your network has very old equipment, you won’t need to invest in new hardware (and if your equipment is indeed that old, you will have to replace it soon anyway…)

The same is true of software. Network services and software of all types have supported IPv6 for a very long time now, so even if your enterprise is running exclusively on IPv4, it is unlikely that you will require any equipment or software upgrades to transition.

The only expected expenses in transitioning are those involved with network administrator training, something we’ve talked about already, and the labor required to implement the transition.

Some additional costs may also arise from your ISP, which could involve purchasing public IPv6 addresses to replace your IPv4 addresses (if you have them). Another consideration is the additional expense of designing and implementing a dual-stack arrangement for the coexistence of IPv4 and IPv6. This may be needed for your own internal enterprise network as well as for the IPv4 and IPv6 address ranges you may advertise to the Internet for your public-facing services. However, these costs depend on the specific requirements of particular networks and can vary significantly.

In most cases, IPv6 is not expensive to deploy: The cost typically falls well within the normal operational costs of any business. 

So, why has IPv6 adoption been so slow?

It’s important to remember that IPv6 is not a new product on the market or a new service that was made available to users worldwide. It is a standard. For this reason, we cannot compare the adoption rate of IPv6 to that of, say, the newest smartphone or the latest social media service. 

Standards typically take years to be fully adopted into common usage, a pattern that we can see in other areas, such as mobile cellular network technologies. For example, take a look at this timeline of milestones for the fourth generation (4G) mobile networks standard:

• 2002: A strategic vision for 4G is published by the ITU.
• 2008: The ITU publishes the official standard for the technology.
• 2009: The first commercial 4G network is deployed in Oslo, Norway and Stockholm, Sweden.
• Today: 4G penetration has still not peaked worldwide (even though 5G has already started being deployed).

As you can see, the 4G standard has already had a life cycle of over 20 years, and its rollout is likely to continue for many more years. This example simply shows that standards, unlike other entities within the telecommunications realm, are expected to take years or decades before they peak.

The IPv6 adoption rate simply feels like it’s slow because of the pace of change we are used to in other areas, but in actuality, it is understandably so. IPv6 is the foundation upon which the world’s interconnectivity is built, so it makes sense that such a complex entity would require that much time to transition from IPv4 to IPv6.

Look at IPv4: It is over 40 years old and still going strong. 

Motivation is key

There is a famous old adage that many individuals believe in, including many technical people: “If it ain’t broke, don’t fix it!” Many think that there’s nothing really wrong with IPv4 at this time, so why should they change it? Administrators don’t really like to make fundamental changes to their networks unless they see something that motivates them. 

Things like performance and cost are indeed motivators: if you tell an administrator that a new protocol will increase performance by 50% or reduce cost by the same amount, that will lead to immediate enthusiasm. However, that’s not what IPv6 was about. It was about fixing something that wasn’t “broke” now but was clearly going to break in the near future due to its limitations: IPv4. Changes to avoid future problems are critically important, but they are hard to get people excited about. 

There is little promise of an immediate tangible benefit here, so many simply say, “Let me stick with IPv4 until I actually need to go to IPv6.” That may be short-sighted, but it is one additional reason why this change is taking so long.

Conclusion

IPv6 adoption has been declared by some to be occurring too slowly. Many have attributed this slowness to challenges that are too difficult or unprofitable to overcome. Some believe that IPv6 is not yet ready to replace IPv4, and for this reason, IPv4 is still the dominant Layer 3 protocol worldwide.

However, many of these declared “challenges” are misconceptions concerning IPv6 related to incorrect ideas about maturity, protocol speed, security, and costs. It’s essential to understand these mistaken notions before seriously considering deploying IPv6.

IPv6 is a standard, and unlike other advancements in telecommunications, standards take time to be adopted and to gain penetration into user markets. In fact, the rate of adoption of IPv6 is proceeding at a rate consistent with that of other standards. 

IPv6 will continue to gain ground and eventually replace IPv4 over the coming decade.

FAQs