In Part 1 of this series, we discussed what’s wrong with IPv4. Looking at our roadmap again, let’s move to the next part of the journey.
- What’s wrong with IPv4?
- What is causing the Internet to grow and are these things I care about?
- Is IPv4 address depletion a problem that could impact me?
- Is a new solution warranted and does it bring any value to me?
In 2008, the number of nodes on the Internet exceeded the human population (1). Since then Internet growth has continued with the momentum of a freight train. Let’s take a look at the drivers:
- World Population Penetration
- Mobile Devices
- Devices per User
- Internet of Things
Starting with the obvious, the current world population is a little over 7 billion. About 34% of them are currently online (internetworldstats.com). At present rates, this number will double again in 5 years (2) and we’ll be approaching 5 billion Internet users. In an addressing system designed for 4 billion entities, this is strike number one.
The gradual conversion of all cellular phones to smartphones in tandem with BYOD is causing torrid network growth. Globally, there are over 5.6 billion cell phones (3). In the US, the percentage of these that are smartphones passed 50% earlier this year and will do so globally next year (4). The year after that, over 2/3 of all phone sales will be smartphones. So around 2016 or 2017, we’ll exceed 4 billion phones on the net. Strike two. But wait – we haven’t even accounted for tablets yet.
Smartphones and tablets are only the start. Read Anywhere by the Yankee Group (5) who expects the average person to have 10 networked devices by 2015. Besides your laptop, smartphone and tablet, you may have an umbrella that lights up when it’s going to rain, a pill dispenser that reminds you to take your supplements or medicine, a car hooked up to ITS, an armband to track your exercise and vital stats, a headset that takes commands from you, glasses that let you watch movies on your flight, shoes that track your running times, and keep in mind we’re limiting this just to personal devices (6). As you go from home to work, you’ll expect this to work seamlessly. Don’t think so? Just ask any of the large multi-nationals who have depleted all RFC 1918 address space while rolling out BYOD. So then what? Do you use non-routed public address space and hope you never have to interact with the real owner? Do you have regions of RFC 1918 addressing with double NAT in between them? So what happens when you have to manage, monitor, and troubleshoot this? As you start thinking down this path you’ll see why the large ISPs started migrating to IPv6 about 10 years ago. That’s when they reached the problems enterprises are now hitting. They studied all the solutions and determined that IPv6 is the cheapest solution – but I’m getting ahead of myself.
Most interesting of all is the Internet of Things. Cisco’s Infographic (1) is a great start to get your imagination going. The Internet of Things is transforming society. There are over 10 billion new microcontrollers shipped out each year, with an increasing percentage of them networked. These microcontrollers, sensors, and machines are talking to each other to further monitoring, reporting, and automation. There is tremendous activity in the space – Gartner for example ranks it as a 2012 top ten strategic priority. Because it is obvious that the Internet of Things must scale to 10s of billions of devices, many devices in this category only use IPv6 (e.g. Smart Grid). For those that do use IPv4, you can see the immediate problem. With over 100 billion microcontrollers in the world, only a small percentage of them need to be networked to overwhelm the IPv4 addressing system. When you start to think about where we’re going with the Internet of things – pretty much all electronics being networked – you get an idea of just how hopeless it is in IPv4. Think about where you work. Imagine every light bulb, everything you can plugin, your HVAC/environmental controls, power feeds, projectors, corporate vehicles – all networked. Then imagine this for adjacent business, your city, your county, your state, your country. This is why we need IPv6. And it’s already here – there are already huge sensor networks running today.
So how does this impact me? Through two directions. You and your business depend on the Internet. We probably have about 2 years before things get really ugly. So why not wait to do something about IPv6 until then? Because it will take you years to fully deploy IPv6. Yes, really. That’s why it’s so important to start now. You don’t need to rush a deployment, but you need to assess your environment to understand where you’re at. You need to build out a lab to learn IPv6 – it’s a new protocol. You need to develop your implementation plan. The second avenue IPv6 will come at you is with mobile devices. LTE/4G uses IPv6 as the preferred protocol. This is because the designers believed LTE would outlive IPv4. In fact, all major US cellular carriers have deployed or are deploying IPv6 in their networks. So what – LTE has been out for a while with IPv4, right? But that’s because the end devices didn’t support IPv6 cellular connections until relatively recently. However, Apple’s iOS 6+ and Android 4.0+ both fully support IPv6. So as new LTE-enabled devices running iOS 6 or Android 4.0 start appearing on or interacting with your network, what will happen? For example, when your CEO uses his LTE mobile device to access your web sites or VPN, does it work? Are there any problems? You do realize that on a dual stack host, IPv6 is preferred, right? We’ll talk about default prefix policies in a future post.
One question still remains – clearly the Internet will eventually outgrow IPv4. But does IPv4 depletion affect me today? We’ll tackle that question next time.