An Update On IPv6

health and Nirvana. But recently a listener asked me to give an update on the switchover to I p V six, which I thought presented a great opportunity to talk about what that actually means, why it's important, as well as what progress has been made in this switchover. Now. It all boils down to solving April saying problem, which is that the old set of rules we relied upon for the Internet just aren't quite sufficient to keep up with

the way we're using the Internet. So first we need to define what is i P. It stands for Internet Protocol, which sounds a little daunting, but really that just means it's the method or set of rules that the Internet follows in order to send data between computers or other devices that are interconnected within this network of networks. In protocol speak, we typically refer to these devices as hosts, and from a high level, the rules are pretty straightforward,

but they are absolutely necessary. Without these rules, it would be pretty challenging to find what you're looking for when you connect to the Internet. So let's take a very simple connection and that way we can build on that to understand what the challenges are. Back in the good old Bolton Board System days, the BBS days, you would typically have one machine acting like a server. Everything would live on that machine and it would serve data to

a single client at a time. So that meant you add one computer or host in the terminology of the Internet, and this one is the server that has all the message boards, files, games, whatever the Bolton Board System might have. The people who want to visit the BBS would use a dial up modem on their personal computer. This is another host in Internet parlance, but we also would call it a client and they would use that to call

up the server computer over a normal phone line. So it's just like if you were to make a phone call to somebody else. The connection was direct, or at least as direct as a connection can be when it crosses over telecommunications infrastructure like phone lines. The point is the client and server communicated directly with one another. The data didn't need to pass through any third party hosts. A good analogy is the old two cans and a

string method of communication. You know, you got a can on one side and a can on the other, string connecting the two, and you speak, and the vibrations go through the string and are amplified on the other side, and thus you can talk well. The two computers in this example are those cans, and the phone infrastructure would be the string. But the Internet consists of millions of computers, plus routers and switches and other devices that are all

interconnected in various ways. Whenever you visit a website or you send an email, you're sending and receiving data to and from other devices connected to that network, and some of them might be across the world from each other, which means there have to be rules in place for your messages to get to the right computers, and those computers have to know where to send the data back to you in response to your requests. Part of the Internet protocol addresses this very issue, and yes, that is

kind of a pun. It's called IP addresses. An IP address is a unique identifier for every computer or connected device on the Internet. Uh, anything that's directly communicating with the Internet has to have an IP address, and it's similar to a physical address that we would use for mailing things in that it provides a means for computers to locate the right destination for data. But unlike a physical address, a machine's IP address doesn't necessarily always stay

the same. It can it can be a static IP address, but it's pretty common to run into dynamic IP addresses, which means they can change depending upon the local network that it connects to the router that's in charge. Lots of different things can determine what a machine's IP address is at any given time. Now, the idea for IP addresses goes back much further than most people's experience with the Internet, unless you were a researcher or an engineer.

The debut of the Worldwide Web in the early nineties kind of helped usher in a new era of computer users, but the Internet itself had been around for a decade before the Web was a thing. Those rules had to be in place for the Internet to work, and they were, in fact an evolution of the rules that engineers were creating when they built the predecessor to the Internet, called our Bonnet. Our Bonnet was a Department of Defense project.

Those working on the project, we're creating the framework and rules within which different computers built on different architectures could meaningfully communicate across a network. This was a pretty hefty undertaking, but I've talked about it in other episodes of Tech Stuff, so I'm not gonna go all the way through that again right here and right now. You can find the

episodes on our Bonnet and listen to those. But in a Request for Comments document and r FC document dating from nineteen eighty, the team working on Internet protocols detailed the necessity for addresses and helped clarify their role. Part of that meant explaining what an address is and is not. So here's a quote, a direct quote from RFC seven sixty and yes, fasten your seat belts because it's really exciting speech. A distinction is made between names, addresses, and routes.

A name indicates what we seek, an address indicates where it is. A route indicates how to get there. The Internet protocol deals primarily with addresses. Is the task of higher level i e. Host to host or application protocols to make the mapping from names to addresses, and so a device would need a name and an address. The route would be the pathway the data would take between the hosts. Now, this isn't all that different from sending

snail mail in the real world. You can't just put an envelope in a mailbox with a person's name on it and expect the postal service to be able to deliver it, unless I guess the name is Santa clause you need to include an address on the envelope. The postal service then takes the collected outgoing mail and to turn emens how to deliver it, which is not that

different from how computers send data across the Internet. Now, one thing that is a little different is how devices on the Internet package data when they send it across the Internet. Devices on the Internet send data in small batches called packets. These packets are sort of like puzzle pieces. They don't all have to take the same pathway to get to their destination, where they are then reassembled into

whatever it is that you were sending. It's a pretty ingenious design that helps avoid problems when a particular machine on the network goes down. But let's get back to addresses, because that's kind of a separate topic. In this request for comments, the group also mentioned that addresses were a fixed length of four octets or thirty two bits. Now, a bit is a single unit of information. It can

take the form of a zero or a one. So the team would refine this a bit in future RFC documents, but for the time being four octets or four uh eight bits four bytes. In other words, the thirty two bits gave the team a lot of potential addresses to play with. So each bit can have one of two states, that being a zero or a one, and there are thirty two of them total. So what you would do is you take to the states and raise it to

the power of thirty two the number of bits. That gives you a grand total of four billion, two hundred ninety four million, nine hundred sixty seven thousand, two hundred ninety six addresses. Almost Now why do I say almost, because not all of those addresses are actually available to the general world. The protocol reserves nearly two hundred ninety million addresses for specialized purposes, but still more than four billion addresses. Seemed like they were going to be plenty

way back in nineteen eighty. Now, the way we typically see those addresses written out is in a series of decimal numbers separated by periods. For example, you might see an address like this two hundred sixteen dot dot sixty one dot seven. Those are your four octets, except they're being represented as decimal numbers, not as bits. So you'll never see a number higher than two hundred fifty five

in any of those four spots. And why is that, Well, it's because the highest number eight bets can represent would be two fifty six. So you start with zero, you can go up to two five. If we started with just one and built up, we could go all the way up to two fifty six. But we start with zero because we're talking about base ten, so once you hit nine, you need to be able to flip back over to zero. So two five is the largest number

that can occupy any of those four spots. If we wrote out that same IP address I just mentioned the to sixteen dot seven dot sixty one dot one seven, it would look like this in binary code, and I apologize ahead of time because you're gonna hear a lot of ones and zeros, guys. But this is why they decided to switch over to a decimal based system for notation.

The binary address would be one one zero one one zero zero zero, dot zero zero zero one one zero one one dot zero zero one one one one zero one dot one zero zero zero one zero zero one. And yes, it sounds like I just did the binary solo from the humans are dead by fly of the concords. It doesn't exactly roll off the tongue, so you can see why the team decided to go with that decimal

based system. Now, at a higher level than the Internet protocol, there are rules in place that allow us to use stuff like domain names and file pathways to make it easier to direct emails or go web browsing to where we want to go. In tim Burners, Lee introduced the concept of uniform resource locators or u r l s. But all of that is just a way to make it easier for us humans to navigate the Internet and the Web. On the machine level, it's all about the bits,

all right. So we have our four octets of bits, each representing a decimal number in that series of four numbers. This set of rules was part of the fourth version of the Internet Protocol, so we tend to call these i p v four addresses or just IP address for short. Typically, if someone says IP address, they're talking about i p v four. So what happened to versions one through three, Well, those were all experimental versions of Internet Protocol that were

never rolled out for general public use. Now, originally IPv four divided up the octets to designate networks and hosts. In other words, part of the address would tell the system which network your computer was on, and the rest of the address would be specific to your actual computer on that network. Because remember, the Internet is a network of networks. It's not like each computer is just independently plugging into a giant web of data. These are hierarchies

that we're looking at. So this is not again that different from an address on an envelope, because on an address and an envelope you'll have general information such as a country, a state, a city, that kind of thing, and then more specific information like a house number and a street that sort of thing. So again, very similar to the way physical mail works. But even early on

while before the Web. The team working on Internet Protocol realized that this was going to limit the useful number of addresses a bit too much, and so in nine they redefined the approach by creating five classes of networks. They designated the classes from A to E. This became known as class full networking. Now, classes D and E contained the reserved addresses I talked about before. Some of them were for uh experimental purposes and some of them

were reserved for other reasons. Classes A through C use different bit lengths to designate specific networks. This removed the tight restriction on the number of networks that could join the Internet using IPD four, which up to that point had been a measly two fifty six networks. If the Internet only consisted of two ft six networks, we never would have had to worry about adding this change. But clearly that was not going to stay the same now.

If you connect your computer directly to your Internet service provider, also known as your I s P, your computer would get an IP address from that I s P. It would have some sort of designation for this, but many people are using routers or connecting to the network through another network that has its own router, and in these cases, the router has an IP address specific to it and typically creates a subnetwork for all the devices or nodes

that connect through that router. So let's take our little example again. So let's say your computer's IP address is currently set to to sixty seven dot sixty one dot one seven, and in this example, we'll say your network's identity is represented in the first three octets, so that means your network's identity is to sixteen dot dot sixty one.

The final octet designates the node on that network, in other words, your computer, So you could theoretically have any value from dot zero as the final one to dot to. Now I say theoretically because this is a Class C network and as such may not have an ending in zero or two fifty five. But that gets a little too technical. The computers on the network use a set of rules called a subnet mask to separate the octets that indicate networks versus nodes. Uh sub net masks in

I p V four follow up patterns. So the first such mask is two five five dot zero dot zero dot zero. This mask reserves eight bits to designate the networks. That eight bits is in that first octet, the number that's represented in there is two five, and then it reserves the other twenty four bits to designate nodes, So that would be a Class A network. On the flip side, you have two five five dot two five five dot

two five five dot zero. That would mean you would have twenty four bits the first three octets to designate the network and only eight bits to designate the various nodes. This would be a Class C network. These are called masks because they guide routers to look at specific numbers within the addresses, and by looking only at the numbers that are pertinent to the network, the routers can save some time. They don't have to process an entire thirty two bit address. They just look at the bit of

the address that's important. Again, going back to the postal service, this would be like if you had UH two major piles and one of them is for local mail and the other is for non local mail, and you look at the address and just by looking at the bottom line where you look at the the state, you know is it local or nonlocal, and so you just do a quick sorting that way. It's kind of similar to that.

But why would we have all these different classes. Anyway, let's all in how a network work gets set up. So a network administrator would make these sort of determinations. Uh, if a network is going to have a lot of subnets, you need more bits to designate networks. If the subnets are going to have a lot of nodes, that is, computers connected to those subnets, then you would need to dedicate more bits for the node or a computer side

of things. It's all dependent upon the infrastructure of the network. I'll explain a little bit more, but first, why don't we take a quick break to thank our sponsor. All Right, on the Internet itself, only the network part of an IP address is important. Uh, the the Internet machines don't care about your specific computers address. They just need to know what network it needs to go to. Data will move toward that appropriate network, and once it's there, the

host part of the address then becomes important. It's it's kind of like once you get the mail to the right state, then you need to start dividing it up by a city and thus neighborhood and that kind of thing. IP addresses can be static, which means they don't change, or they can be dynamic, meaning they get assigned every time the device connects to the Internet. Most devices in the hands of consumers use dynamic IP addresses, and many are connected to sub networks or subnets, which are networks

that share a common network address. A single company might use subnets to share a single network address across all subnets, even if they are geographically distant from one another, and that helps conserve the number of network addresses that are generally available. Even so, it didn't take very long for the engineering team to realize that thirty two bits was a big limiting factor, and so they set about to

solve this problem. Otherwise, the world would reach a limit on the number of useful addresses, and no new networks would be able to join the Internet. It would have hit capacity. So with IPv four addresses again, we have a little bit more than four billion available ones once you take away the ones that have been reserved for other purposes. A billion is admittedly a very large number,

but not nearly large enough. Even in the late eighties and early nineties, engineers were saying this is gonna be a problem. Things were changing quickly, and it soon became clear that we'd hit a real crunch with IP addresses. For one thing, there was an emerging trend of moving from an on demand connection to the Internet to pervasive connections.

Now by that I mean in the early days, you would typically use a dial up modem to connect to the Internet, and frequently this meant that you were engaging your homes one phone line so that you could call up an Internet server and browse the Internet. So you wouldn't stay on it forever because you need that phone line for other things, and frankly, you could end up racking up big charges if you were on for a really long time, so you would typically hang up when

you were done. Now hanging up meant that you were you had just freed up an IP address. You didn't need that IP address anymore because you were no longer connected to the Internet. So even if you had more people than you had I P addresses, not everyone was going to be connected at the same time, and so you can kind of fudge things a little bit. But then we gradually began to move toward more broadband connections using things besides just dial up modems, and we also

started to make a move towards always connected devices. Now this is even more true today when you have everything from set top boxes like video game consoles, to wireless devices like smart thrum, the stats and home security devices to add in. Then there's an added problem that any device that has multiple ways to connect to the Internet

needs a different i p address for each of those methods. So, for example, if you have a smartphone that connects to an LTE network as well as to a WiFi router UH, it has to have an IP address for each of those, Or if you have a laptop that has both a wired and a wireless connection, has to have an IP address for each of those, So a single device can take up more than one IP address. And of course the other big challenge was that more of the world

was going online. The Internet when it first got started was kind of an exclusive club, really nerdy exclusive club, but still exclusive. But now it was going global, so there were just a lot more people trying to connect and a new method of addressing was needed. In the Internet Engineering Task Force developed the successor to i p V four. This one was called i p V six, So that immediately raises a question, why the heck did

we jump from four to six. Well, what would have been Internet Protocol five was an Internet stream protocol that was in development as far back as the late seventies. Ultimately, that protocol never rolled out for public use, and it also was never adopted as i p V five officially. But back then no one was really sure how things were going to turn out, and so they went with

i p V six to be safe. It expanded that bit number for addresses from thirty two to one twenty eight, and they don't look like i p V four addresses either, and I V six address consists of eight groups of four hexadecimal digits. The hexadecimal system is a base sixteen system. So with a base ten system, you start at zero, you go up to nine, and then you start over right, ten is just one zero, and then you can go up to nineteen, which is just one nine, and then

you start over again. Now you go to twenty, which is to zero, and over and over again. Now hexadecimal goes up to sixteen, but you really can't show that using just decimal numerals. It doesn't work that way. So with hexadecimal, you start at zero, you go up to nine, and then to fill in the additional digits you use a sequence like a B C d, E and F, so, in other words, and hexadecimal it goes zero to f before repeating. Each hexadecimal digit in this sequence of thirty

two represents four bits, also known as a nibble. So while each decimal is and I p v for in an IPv for address represents an octet, each digit and a hexadecimal address represents a nibble. And now I'm not making any of that up. Four times thirty two is one eight, So that's where you get the one eight bits. What this means is that the effective number of addresses sky rockets. If none of the addresses were reserved for special use, you would have three point four times ten

to the thirty eight power of addresses. So what is that in real numbers? I'd love to tell you, but the closest I can really get is three hundred forty trillion two d eighty two billion, three hundred six million, nine thousand eight followed by twenty four zeros. It's more than enough for the foreseeable future. Even with such developments as the Internet of Things, an IPv six address might

look something like this. And bear with me, because again, this is a long and somewhat confusing one f e eight zero, colon zero zero zero zero, colon zero zero zero zero, colon zero zero zero zero, colon zero two zero to colon b three f f, colon f e one e, colon eight three to nine. Each sequence of four digits has a colon separating it from its neighbors, so it's essentially the same sort of purpose as the

dot in the I p v four addresses. In addition, there are some nice tricks that you can use to abbreviate those long addresses. If you have a leading zero, in other words, you've just gotten past the colon, or maybe you're even in the very first uh octet um, then you get not octete. I'm sorry, you're in the first section. Any leading zeros you can drop, you can just go to the first number that is not a zero. Also, if you have a single range of zeros, you can

drop those and replace it with a double colon. If you have several series of four zeros, you can drop all of those if they're all consecutive, and replace them with a double colon. So you might remember that when I was reading off that very long I p v six address, there were a few uh segments that went zero zero zero zero, Well, that means we could drop

those and just put a double colon in there. So that last address I read a minute ago, if we abbreviated it, we could say was F E eight zero double colon two zero two colon B three f f colon f e one e colon eight three to nine. Simple as pie, or at least easier than what the previous one was. Now, I mentioned that the Engineering Task Force developed the strategy in but it wasn't until two thousand seventeen that I p v six was officially adopted

as a standard. And I p v four and I p v six are not fully compatible on their own, and I p v six address can contain an embedded I p v four address, but the I p v four format is not forward compatible with I p v six because well, that's the way time works. When you think of something first, you don't necessarily make it where it's automatically compatible with the next thing, because you haven't

thought of the next thing yet. Sadly enough, it also takes a lot of time to switch an established infrastructure over into a new set of protocols, So that has what's one of the other reasons why it took so long to adopt as a standard and to even make some progress on this. Add to that the fact that the Internet is not a single entity. It's not like there's one big building where the Internet is. It's not like the I T crowd where they put it on top of big ben The Internet is a distributed network

across lots of different machines. So rolling out a comprehensive change to the protocol is a colossal task, and it's a big complicated mess, and there's no one entity in charge to say, hey, go out there and do this. Everyone agrees that we need to deploy i p v six across all networks and adopt it moving forward, but

actually doing it isn't that easy. According to the website World i p v six launch dot org, the percentage of i p v six adoption among the top one thousand sites as measured by Alexa Internet, which is a subsidiary company owned by Amazon uh the numbers just below, So all the traffic going to the one thousand top websites of those top websites support i p v six, which means that more than sevent of the most popular websites are not yet reachable via i p v six,

and that is a bit of an issue. I will say that a lot of the I s p s, particularly the big ones in the United States, are pretty well prepared. They have done a fairly decent job at switching over their infrastructures to support I p v six. But we still have a long ways to go. Speaking of a long ways to go, let's take another quick break and then we'll conclude our discussion about I p

v six. So every day we're adding more devices, including routers, to the Internet, and we've established that even in the nineties, engineers recognized that the number of addresses that I p v four could handle was not up to snuff. How in the blue blazes have we managed to avoid a catastrophe for so many decades? So does that just mean the problem wasn't as bad as we thought it was. Because if we haven't switched completely over to I p V six, and we knew that I p v four

addresses running out, why haven't we hit a massive crash? Well, the problem is evident. The four plus billion IP addresses are actually all gone now. That is, they've all been snagged by various companies and institutions like Internet service providers or unversities. Now those entities can assign out dynamic IP addresses on their own respective networks, but no new network could pop up and request a spectrum of addresses from the Internet Assigned Numbers Authority or I A n A.

That's the authority that oversees the stuff. You couldn't request any new ones from I A n A because there are no new ones left. What you might be able to do is negotiate with some other entity that actually has unused IP addresses and purchase them that way. But there aren't any new ones coming out because i P

V four is all dried up. Below the I A n A R five Regional Internet Registries UH they are called r I R s. These, as the name suggests, oversee I P address assignments over specific regions in the world. Those five entities, in turn can assign banks of IP addresses to local Internet registries, which can include stuff like the aforementioned I s p s and universities and other institutions. And I'll so just to clarify, while an I s P might be a local Internet registry, not all I

s p s fall into that category. Some I s p s belonged to a larger entity, which in turn is the actual local Internet registry. So this stuff gets pretty complicated. Back in January two thleven, the Asia Pacific Network Information Center better known as ap NICK, A P and I C, which is one of those five regional Internet Address registries I talked about, requested and received the

last two unreserved blocks of IP addresses. There were only five reserved blocks remaining, and so the I A n A ceremoniously granted one block to each of the regional Internet Address registries, and then all of those IP addresses were technically out in the wild. There were none left

in reserve. By April two thousand eleven, app NICK ran out of freely allocated IPv VOORA addresses, which means that sometimes in that region you can't get an IPv for address when you need one, I mean, you can't connect to the Internet. So why hasn't that happened everywhere? Well, it's largely because engineers are very clever at figuring out workarounds or problems. There have been a few temporary measures that have extended the useful life of ip V four

despite the growing number of connected devices. So remember when I mentioned that IP addresses could be divided up by classes based on how many bits are dedicated to the network address versus the host address. Well, there's also something called classless inter domain routing or c I d R

or CIDER. The Internet Engineering Task Force introduced c I d R in as a way to simplify how data moved through routers on the Internet and to extend the useful life of IP v four, and it mostly has to do with the big drawback of the class full system. So the smallest allocation of addresses using the class full system is two hundred fifty six addresses, which is a pretty small number when you remember how many devices need more than one IP address. So that's the most addresses

that eight bits can support. If you were in that class, if you moved up a class, suddenly the number of IP addresses you would get or for this particular class would go from two fifty six to sixty five thousand, five hundred thirty six. There's no in between there. You go to fifty six to sixty thirty six. That's a huge number. That's more than what most organizations need for the devices that are connecting through their networks. So there

was no way to step between those two. From a protocol standpoint, you either ended up with too few addresses where people were not going to be able to connect their machines properly to too many addresses where a whole bunch we're just gonna go unused and wasted. Um, And that was a real issue. That's when c I d R was able to solve this problem. It was a new method to step around it. Rather than define networks and hosts by octets for bites, you know, sequences of

eight bits, it divides networks into variously sized subnets. So when setting up a network, engineers can aim for a range of addresses that best suits the organization's needs and not go beyond that. Now you have to go for a consecutive addresses if you're using c I DR notation. Otherwise you have to keep using the notation repeatedly. Think it's pretty messy. We use c I DR notation to

represent IP addresses in this way. If you've ever seen what looks like an I P address followed by a slash and then a decimal number, you've likely seen an example of c I d R notation. The decimal number represents the number of leading one bits in the routing mask. Essentially, it's a shorthand notation to express the range of addresses represented in a network, and it's not limited to octets the way class full representation was. One big benefit of c I d R was that fewer IP addresses would

go to waste unused. If you were up a large but not ginormous network, you can set up a range of IP addresses sufficient to meet your needs without going overboard by tens of thousands of excess addresses, which in turn meant that those unused addresses could be freed up for other networks, and it allowed for a better distribution of IP addresses. In other words, Another solution is network

address translation or in a t net. It's a set of rules that allows a single device to act a sort of a liaison between a specific network and the Internet. So remember the Internet is a network of networks, so a device like a router could act as the gateway between your local network, the one that just has a bunch of computers directly communicating with one another, and then everything else out there on the Internet. So I'll use

my office as an example. When I log in from how stuff works, all traffic between the Internet, and my computer passes through our company's neat router. First, the router has a range of IP addresses that ultimately come from I, A, N A. You have to go up a couple of levels, several levels, as it turns out, but I A and

A ultimately was the agency that granted this. My computer, on the other hand, has a non unique and therefore non routable address that works fine for communicating with other computers on my network, but wouldn't work if I could somehow bypass the router and try to communicate directly with the Internet. So what do I mean by that? Well, within a network, you need unique addresses, otherwise data is not going to know where to go to get to

the right destination. But network A could have a series of addresses, and network B could have the exact same series of addresses, And as long as A and B are completely self contained, it doesn't matter. Right, all the computers on B know what it means to go to this particular address, and all the computers on A know what it means to go to that exact same address, because they only belong to their respective networks. When you connect through the Internet, you then have to have another

layer something else to actually uniquely identify the machine. Otherwise it would be as if the house I live in and the house you live in have exactly the same physical address, but are in two different parts of town. That would be incredibly confusing, and we would constantly be receiving each other's mail. I think you have my socks, but my computer is behind the router. From the perspective of the Internet, so my computer is not communicating directly

with the rest of the Internet. It's communicating through the router. It sends data to the router, and then the router in turn routes that out to the Internet. So on my side of the router, on the company side, I don't have a need for a unique IP address. Whenever I try to communicate with a machine that is not on my private network, non the house stuff works network, that message passes through the router, which uses an available IP address that it has assigned to it, and then

sends that message out into the world. The router then has to consult what is called an address translation table whenever data is coming back to determine which machine on our network is the intended recipient. So if I go out to say look at a web page. That message will go out to the router, which then will assign an available IP address to send that out to the Internet. The response will come back, the router will look at the IP address on that that you know, the intended recipient.

Use a trend a network translation table to say, all right, well who did I give this to? Again? Which which computer has this temporary IP address? Oh, it's Jonathan's machine. He's the one going to red versus Blue dot com and watching cartoons. That's typical, And then it would get to me, probably without the actual commentary. Network address translation has a lot of other uses, but for the purposes of this episode, it really kind of sums up what

it does via visa the the IPv for shortage. So this is a lot of of useful treading a water, you might say. In fact, a lot of engineers said that the development of stuff like c I, d R and n A T extended the useful life of I p V four by about twenty years. That's pretty cool, but we still clearly have a need to move to I p V six, which is going to create so many addresses that is very difficult to imagine a time

when we will run out of them anytime soon. It may one day happen, but we're talking trillions upon trillions of addresses here, so the deployment is going pretty well. It's been a slow process when you consider that it was back in that they were first proposing I p V six, But it's largely because, at least in the commercial world, a lot of entities don't make a move

until it's absolutely necessary. So once I A and A allocated all those remaining I p v four addresses, that's when companies said, WHOA, we might need to start working on this I p v six thing. We might need to start rolling that out and making sure our websites are accessible, that our machines can communicate through I p V six, because otherwise we're going to run into some

pretty tough situations. The migration requires software and firmware updates, sometimes hardware updates to stuff like routers across the Internet, but many I s p s, particularly those big ones like the ones that we have here in the United States, have largely addressed this. That no one's at a percent as far as I can tell, but uh, there are a lot of of companies that have gotten to you know, the eighties and ninety percentiles of deployment, which is pretty impressive.

On the consumer side, smartphones pretty much are on the i p v six train already, and all current versions of operating systems support the i p v six protocol, so we're mostly waiting on web servers at this point, if I'm being honest. Now. Back in two thousands thirteen, cloud Flare CEO Matthew Prince projected that based on the adoption rate of i p v six at that time, we could celebrate a full migration to i p v

six on May tenth of so market calendars. Also, I should point out it's quite quite likely we're never gonna see I p v four abandoned entirely. It's more likely that we're gonna see both sets of protocols continue side by side. It's just that I p v four addresses will eventually be more or less completely locked down. But it's very rare to abandon completely a legacy infrastructure, and the i p v four framework as a particularly large one.

So we will continue to see this deployment, will see more development on the side of i p v six, and uh maybe in a few years it'll we'll see that percentage for the top one thousand sites go up to above the mark. But it's gonna take some time. And honestly, the the measures that have been put in place have created enough slack that a lot of people who obviously should be thinking hard about the future have

kind of put it off a bit. They procrastinated. Uh, that's not great for all of us, But the bright side is we're not going to see the Internet fail. I p V six will be more than enough to solve this problem. It's just a question of when the various entities involved get motivated enough to switch over to I p V six. So I don't think that they were headed toward a catastrophe at this stage, at least

not collectively. There might be individual companies out there that find themselves scrambling once it gets to a certain point, but maybe that'll just mean they'll learn a valuable lesson and then the board of directors will change the CEO and then the whole thing will start over again. That's cynical.

We're not gonna talk about that. Let's wrap this up, Hey, guys, if you have any suggestions for future episodes of tech Stuff, I have a solution for you send me an email address is tech stuff at how stuff works dot com, or you can drop you a line on Facebook or Twitter. The hand over both of those is tech Stuff hs W. You can watch me live on twitch dot tv slash tech stuff most Wednesdays and Fridays. I'm recording episodes. I'm glad to have you guys watch me as I stumble

through this. I'm so glad that right now I don't have anyone watching me because cold medication makes me do weird things. Oh and uh, make sure you follow our Instagram account. All right, that's it for this episode. Join us next time when we talk about something completely unrelated to Internet protocols, and I'll talk to you again really soon. For more on this and thousands of other topics, how stuff works dot com