Chapter 1 – A History Lesson
IN THE BEGINNING, the Internet Protocol was part of TCP. There was no separation. There was no layer-2 and layer-3. It all worked as one…
When I say ‘the beginning’, I’m talking about 1973, and the protocol was TCP version 1.
Fast forward a little to 1977, when an American computer scientist named Jon Postel split the TCP protocol into TCP and IP. Two separate distinct layers, and the birth of the Internet Protocol.
After a few revisions, TCP/IP finally went live in 1982, when the US Department of Defence made TCP/IP the official standard for military networking.
The fourth generation of IP, the Internet Protocol, was the first to be formally used across the globe. The internet has officially been born!
Even though IPv4 has been in use for forty years, it’s still the most popular version of the internet protocol.
But IPv4… has a dark side…
In the decade following IPv4’s introduction, a critical flaw was discovered. The original IPv4 architecture used 32-bit addresses. This supports about 4.2 billion addresses.
That would have seemed like a lot back in 1983. But then the internet started becoming popular. It wasn’t just for the US military and universities anymore. It was gaining global appeal.
The world-wide-web was invented. Businesses got online. Some people even had internet access in their own homes!
By the end of the decade, it was predicted that the global IP address pool would soon be exhausted.
Something needed to be done! In 1992 the first official document outlined some conservation strategies to combat the unexpected ‘explosive growth’ of the internet.
Additional attempts to extend the IPv4 pool were made. 1993 saw the introduction of Classless Interdomain Routing. IP addresses were no longer allocated in massive classful blocks.
In 1994, NAT was introduced. NAT, along with private IP addressing, further conserves the number of public IP addresses that a site needs.
Because of NAT, not all devices need public IP addresses. A business could support hundreds or thousands of devices with a handful of public IPs.
NAT, CIDR, and other techniques have certainly helped extend the life of IPv4. But these are still just workarounds. They are not permanent solutions.
The permanent solution is IPv6. It can work alongside IPv4. But the ultimate goal is to replace IPv4.
You might have heard of IPv6 before. Some people think it’s intimidating. Trust me, it’s not so bad.
In this section of the CCNA series, we’re going to demystify IPv6. We’re going to cover:
- The basics of IPv6
- The new address format
- The different types of addresses, such as public and others
- IPv6 configuration
- And, routing IPv6 traffic
Chapter 2 – Introducing IPv6
One of the most obvious changes in IPv6 is the format of addresses. We’ll see how they work in detail in the next video.
The reason for this change is the size of the address space. IPv4 addresses are 32-bit numbers, while IPv6 addresses are 128 bits.
This might sound like it’s four times bigger. But when you think about how binary numbers work, you’ll realise that IPv6 addresses are a LOOOOT bigger.
Let’s put that in perspective.
There are nearly 4.3 billion unique IPv4 addresses. Of course, some of these are reserved for special uses.
At the time I’m recording this, earth’s population is approaching 8 billion people. So we don’t even have enough IPv4 addresses for everyone.
Let’s think bigger to get our heads around this. Let’s look at the Milky Way galaxy. This has about 100 billion stars.
That’s a drop in the bucket compared to the whole universe, which, according to one estimate, has about 200-billion-trillion stars.
Now, how does this compare to IPv6? Any guesses? There are 340 undecillion IPv6 addresses. That’s about 1.7 million billion IP addresses for every star in the universe.
So if IPv6 solves the exhaustion problem, as well as adding other features, why don’t we just use it? Why didn’t we all start using IPv6 back in the 90s?
Well, there’s a problem. IPv4 and IPv6 are independent of each other. That is, a host with only an IPv4 address cannot communicate with a host that’s only IPv6.
So it’s not simply a matter of turning it on. All networks across the globe need to migrate to IPv6.
There are a few ways this can be done. One of the simplest options, called dual-stack, is to configure devices to have both IPv6 and IPv4 addresses.
Ok, so, we need to migrate… Why don’t we do that then?
Well, we are doing that. But it’s a very slow process. Imagine a large company or service provider with links across the country. There’s a lot of planning, effort, and expense in changing the configuration of their entire infrastructure.
Another reason is that there are a lot of applications that haven’t been written with IPv6 in mind. If you’re already working in IT, you’ve probably noticed that people don’t like change all that much. A lot of people are still using old applications, which need to be updated.
And unfortunately, many people are frightened of IPv6. Seeing an IPv6 address for the first time can be intimidating since it looks quite different. But you don’t have to be scared. It’s just something new. By the end of the next few videos, you’ll feel much more comfortable, I guarantee it.
So even though IPv6 adoption has been slow in past, it’s gaining momentum.
This is a graph of the IPv6 devices that connect to Google’s services. It really speaks for itself. More and more people are using IPv6.
So, love it or hate it, as network engineers, we need to know how it works, and how to support it.
Chapter 3 – Exhausted!
Here’s a common question that often comes up when we discuss IPv4. Are we really running out of IPv4 addresses? Haven’t we been hearing about this for years?
It has been years, yes. As we’ve said, IPv4 exhaustion was predicted in the 80s, and a lot of steps were put in place in the 90s.
Thanks to the work being done in the 90s, IPv4 exhaustion was pushed out by years.
Let me illustrate… There are a bit over 4.2 billion IPv4 addresses. Some are reserved for special uses, but let’s just work with 4.2 billion addresses for now.
There are nearly 8 billion people on earth. If half of the population had only one device, like a smartphone, that’s 4 billion devices right there. IPv4 is exhausted.
Fortunately, NAT and private IPs helped with that, which allows several devices to connect to the internet with a single public IP.
But most of us have more than one device, don’t we? We have phones, PCs, laptops, tablets, TVs, and other devices. That’s on top of the infrastructure devices like routers, switches, and servers.
I found an article from 2014 that said that 10 billion devices are connected to the internet.
We now have a technology called IoT, or ‘Internet of Things’. This dramatically changes the number of connected devices.
In 2019 there were nearly 8 billion IoT devices connected to the internet, with a prediction of 16 billion by 2025. And that’s just IoT
The population is growing. Our reliance on internet connectivity is growing. And the number of devices on the internet is growing.
So, just how close to the edge are we? In order to understand this better, let me explain where IPv4 addresses come from.
You see, when two protocols love each other very much…
…nah, just kidding, it’s not like that.
At the very top of the family tree is an organization called IANA. They hand out large blocks of IP addresses to Regional Internet Registries.
These RIRs handle IP address allocations in their geographical regions. They hand out blocks of IP addresses to large organizations, and internet providers.
Smaller businesses and home users get their IPs from their internet providers. This can be as little as a single IP address.
OK, so back to the main question. How close are we to running out of IPv4 addresses?
Well, in January 2011, over 10 years ago, IANA handed out the last of their /8 blocks to their RIRs.
APNIC was the first to exhaust its allocation of freely available IP blocks. That means they were down to a single /8 block.
It doesn’t mean there is nothing left, it just means the quantity they have left is extremely limited and difficult to acquire.
This happened later in 2011.
RIPE followed in 2012, LACNIC in 2014, ARIN in 2015, and finally AFRINIC in 2017.
That doesn’t mean IPv4 is completely dead. There are still some attempts to keep it going.
For example, some ranges that were previously reserved for special use have been made public.
Some institutions, including Stanford University, have returned /8 blocks that they were given in the early years of the internet.
And as you can imagine, there’s a bit of a marketplace for buying and selling blocks of IPv4 addresses.
So IPv4 exhaustion is happening slowly. IPv4 exhaustion sometimes seems slow, like climate change. But, it is inevitable.
The long term solution, of course, is IPv6.
Exhaustion isn’t the only reason you should know IPv6. It’s getting more and more popular. So like it of hate it, we need to know how to support it.
Oh, and quickly, was there ever an IPv5? Yes! But it didn’t get past the experimental stage. Because it was designed for video media and didn’t consider exhaustion, it was a dead end before it even started.
Chapter 4 – What’s New?
So what’s new? What’s changed? What’s the same?
Well, IPv6 is still routable, just like IPv4. It’s still part of the Internet Protocol and it still works at layer-3.
We can even use IPv4 alongside IPv6. This is called dual-stack, and it makes migration much easier.
The most obvious change is the size of the address space, as we’ve seen in the video.
Along with that comes a different address format. The addresses look quite different, as they’re written in hexadecimal. We’ll see how this works in the next video.
There are no dotted-decimal subnet masks anymore either. All subnet masks are now in CIDR notation, as shown here.
NAT was intended to be a short term solution for IPv4 address exhaustion. So, there’s no NAT or PAT as we currently think of it, in IPv6.
There are some types of NAT, such as NAT64 and NAT46. These translate between IPv4 and IPv6 addresses. But these are only there to help with migration from IPv4 to IPv6.
There is a NAT66, which is like old-fashioned NAT, but it’s not really recommended. While researching this video, I even found an article calling NAT66 a type of ‘abuse’.
A few of the older technologies have been refreshed.
In place of ICMP is ICMPv6. We’ll see how this has grown in upcoming videos.
DHCP and DNS have been enhanced to make DHCPv6 and DNSv6. DNSv6 even comes with a new record type called the ‘Quad-A’ record.
And of course, routing protocols have been extended to support IPv6. In a later video, we’ll see how to use OSPFv3.
An interesting one is ARP. The Address Resolution Protocol. It’s gone now. It has been replaced by NDP, the Neighbour Discovery Protocol. We’ll definitely be looking at this one further.
And finally, I want to mention broadcasts. There aren’t any broadcasts anymore. “WHAT?!?!” I hear you say, “no broadcasts? How?”
Well, broadcasts really weren’t that efficient. Now, with better protocols like NDP, everything can be done with multicast messages.
You better believe that we’ll be looking at those too.
We’ve got a few more videos on IPv6 after this one. So, I invite you to continue on with the series when you’re ready to see how IPv6 works, starting with the new address format.