A technical look at the OSI layer 3 will focus on IPv6. We will be looking at packet sniffers and Android apps that help us discover IPv6 traffic. Why is IANA so important to IT professionals? How is IPv6 allocation going to work? Understanding IPv6 prefix concepts. What role does ARIN play in IPv6, IANA and its role, the massive address space of IPv6, and how much is already allocated, and looking at Global unicast concepts? Learn how to simplify the IPv6 address and understand the IPv6 address prefix.
Hello and my name is Lowell Vanderpool and this channel is dedicated to IT professionals,
IT students and anyone who's interested in technical subjects.
Well welcome to OSI Layer 3.
Listen carefully to this description of Layer 3.
I'm a big proponent of understanding the big picture of technology
and this description really gives you that.
Here we go.
Datagram delivery using connectionless communications
involving addressing additional services like messaging and diagnostics
and routing across complex internet networks.
There it is Layer 3.
It's really important to understand the principal mission of Layer 3.
Getting data from one computer to another whether it's local or on a remote network
and a router was a critically designed piece of equipment to assist in this mission.
Pushing for even a more clear understanding of Layer 3,
we could say Layer 3 is about transferring variable length packets
from a source to a destination host via one or more networks.
Those variable link packets could be from a minimum theoretically of 21 bytes
to an incredible 65,535 bytes.
Surprisingly today, internet load balancers from Google, Facebook,
mega ISPs pretty much limit this to about 1,472 bytes.
You'll rarely get one bigger than that.
Think of Layer 3 as FedEx, UPS, and United Postal Service all rolled into one.
Getting packets from your computer to any place in the world.
Layer 3 was designed by Vint Cerf and Bob Kahn for efficiency and performance.
Early routers, key equipment for Layer 3 delivery were expensive, slow, and had limited memory.
We will see this design criteria as we study Layer 3.
Another important concept as we look at Layer 3 is that we are on the last of the OSI layers
whose actions and functions are responsibility of the network.
Layers 4 through 7 will be a function of the host machine.
This is a diagram of the Windows network architecture
and Layers 1 through 3 are going to be found in kernel mode.
Layers, the OSI layers 4 through 7 are found in user mode.
The main reason why Layers 1, 2, 3 are found in kernel mode is because of stability and security.
The hourglass graphic that you see is no mistake but is a realistic representation of protocols
and functions of the OSI model.
Higher and lower layers explode in the variety of protocols and functions, features, and technology.
But at Layer 3, it's a function of IPv4 and IPv6 and very few additional services or features.
Layer 3 functions are logical addressing, IPv4 addressing, IPv6 addressing, routing,
datagram, encapsulation, fragmentation and reassembly, error handling, and diagnostics.
Layer 3 has IPsec for security if it's enabled, NAT, network address translation,
which saved IPv4, ICMP, which provides error handling and diagnostic,
and routing protocols that install with a router's operating system improving
that router's ability to effectively determine packet forwarding in complex internetworks.
And of course, the Band-Aid protocol ARP, which is sandwiched between Layer 3 and Layer 2,
helping IP and Ethernet to finally work together.
With IPv6, Neighbor Discovery Protocol, NDP, will replace and enhance functions of ARP
in IPv4 and rid us of the nasty impact of broadcast on our network.
IANA is a global organization responsible for a lot of critical information on the internet,
especially Layer 3. It is known as the Internet Assigned Numbers Authority,
and it is responsible for three critical components, domain names,
this would be all our top level domains, and our root servers.
If you want to know more about this functionality of IANA, if you go to my channel,
you can scroll down under videos and scroll down until you get to DNS root servers.
In this video, I cover a lot of the technology that goes on in the DNS root servers under IANA.
So IANA disperses IPv4 and IPv6 addresses to the regional registries.
The various ISPs then go to the regional and obtain allocations of IPv4, IPv6 addresses.
When you and I go to our internet provider, that internet provider provides you
one IPv4 address to your cable modem or DSL modem.
That will change radically when we get an allocation of IPv6.
But right now, IPv4 is primarily what most of us are getting.
IPv4 addressing was implemented in the 1980s.
And at the time, they were pushing the envelope of technology,
implementing a network address space 32 bits long.
It gave IPv4 2 to the 32 power of available IP addresses, which was over 4 billion.
But keep in mind, in the 1980s, this was the type of computing that we were using.
This was 8-bit. Some of these were 4-bit.
Most of them didn't have an address length over 16 bits.
When you're using this type of technology and you're trying to implement 32-bit address spaces,
it really presses the envelope of technology.
And yes, some of these I actually used.
Starting with the 1980s, they began to design an address scheme of class A addresses,
which gave a lot of addresses to, say, large companies, large organizations,
class B addresses to medium size, and class C to smaller networks.
This sounded absolutely logical.
It was a very simple way of giving and allocating addresses.
But by 1990, IANA was quickly running out of available IPv4 addresses.
No one saw the explosion of the internet.
They just couldn't.
The Internet Engineering Task Force began to immediately look at what kind of solutions
they could design to slow the exhaustion of available IPv4 addresses.
There were three main ones.
Variable Link Subnet Mask.
This was called classless subnetting and basically allows you to further subdivide
subnets or subnetting subnets.
Then there was NAT, Network Address Translation.
There are many, many flavors of this technology because it caused problems.
There were many efforts to solve those problems.
Then there was classless inter-domain routing, which basically took the idea of Variable Link
Subnet Mask and applied it to all public IP addresses.
It introduced the CIDR notation or the slash value that represents our subnet.
This is one of my favorite network apps for Android.
I highly encourage you to take a look at this if you want to know more about the network
that your mobile phone is using.
This is a great tool.
My mobile phone has always had an IPv4 address.
And probably about six months ago, I ran this tool on my mobile phone, which my carrier
is T-Mobile.
And lo and behold, I had an IPv6 address.
This was awesome.
IPv6 is slowly coming to me.
Let's take a look at a Windows 10 virtual machine.
And let's explore the kind of IPv6 that's going on in this classic installed Windows
10 enterprise.
So I've got PowerShell open.
I'm going to do ipconfig.
I'm going to do the slash all.
And I can see a link local address.
So this does not have a global IP address assigned, but it does have a link local.
And notice it does have this preferred.
That means that it has, we've done some multicasting to determine that there's not another computer
that has the same link local address.
So it has been tested, verified that this link local is not duplicate.
Notice we also have this percent nine, which is a zone ID.
Now I'm going to go ahead and launch Wireshark.
Wireshark is a protocol sniffer.
And I'm going to filter.
I'm going to look for any IPv6 traffic on my PC here.
So I'm going to go ahead and launch this.
And you can begin to see it's capturing any kind of IPv6 traffic.
And I see that I've got a link local talking to FF02 colon colon FB.
So let's just let this go and see what kind of IPv6 traffic we have on this virtual machine.
So after a while, you can begin to see I've got a lot of link local talking to what looks
like some kind of multicast address FF02 colon colon one FF02 colon colon 16.
Well, how do I know what they are?
Well, my Wireshark gives me some indication that some of that is ICMP version six.
But let's go take a look at the documentation.
Let me launch my website for IANA.
And remember, IANA is your go-to source for official documentation.
So I'm going to go to numbering resource.
And I'm going to scroll down.
And here we have all the documentation for IPv6.
I'm going to look up multicast.
And I'm going to go ahead and scroll down here.
And here I see multicast addresses FF01 and then 12C, FB.
And I can see and identify what they are.
Let's come down here.
And we see FF021, FF022.
And we can see 6789, ABC.
Let's go back to our Wireshark.
And we've got FF021.
And we can look at our protocol chart.
FF021, that's an all nodes broadcast or multicast.
Back to our Wireshark, FF02 colon colon 16.
And if we scroll down here, we can see 16.
This is an all MLD version two capable routers.
A lot of home routers are coming with this multicast capability.
So my router obviously has it.
And so that multicast address is there.
Wireshark and see what else.
We also have FF02, FB.
Back to IANA.
So FF02, FB is MDNS version six.
And if I want to go look at the RFC, I can just click and go to that site.
This multicast DNS, MDNS provides the ability to perform DNS like operations
on the local link in the absence of any conventional unicast DNS server.
So there you go.
Here we have FF02 colon colon C.
Let's go back to our documentation.
And that's SSDP, which is universal plug and play.
There's always a lot of concern about universal plug and play.
Security professionals break out in a sweat when they see universal plug and play.
It's a great feature.
It helps you discover things on your home network.
But from a security standpoint, a lot of people are not happy about universal plug and play.
All right.
So let's stop a minute and kind of go back to some of the basics of IPv6.
Number one, are we going to run out of IP addresses like we did with IPv4?
With 128 bit IPv6 address, are we going to find ourselves in the same problem
that we found ourselves with a 32 bit address length?
And the answer is unequivocally no.
Just to give you an example, that circle represents all the IPv6
addresses that are capable of being produced by 128 bit address.
The green portion of the pie is called the global unicast address.
They took a chunk and they've already given some of those addresses
to the five regional internet registries.
So they've given away some of those.
There's enough in the 001, that green section, to take care of a decade of IPv6 needs.
You'll see the little blue portion of the pie chart.
That's unspecified loopback.
We'll get into that.
Very small.
Multicast is purple.
You see just a small portion of IPv6 addresses allocated to multicast.
Link local unicast is orange.
And then the red is unique local unicast.
We'll get into what those are.
But you can see all the rest of the white portion of the pie chart is reserved.
We haven't even begun to give away or use any of those reserved IP addresses.
What we've allocated is in the green and that's enough for a decade.
Let's go back to an IPv6 address and take a look at it.
First of all, it's 128 bits.
And in binary, as you can see, it becomes an incredibly difficult thing for a human being
to work with 120 bits binary.
Computers can handle this all day long.
But you and I, this is very, very difficult.
So software engineers decided to use the hexadecimal numbering system to represent
all these binary bits that we have to deal with in IPv6.
What we do is every hexadecimal digit represents four binary bits.
So when we have four hexadecimal digits, like you see,
2001, that represents 16 binary bits.
If you'll notice our example, 2001, and a whole lot of other hexadecimal values beyond that.
But this is a global unicast IPv6 address.
Anytime you see an address, an IPv6 address that begins with a 2 or a 3,
that is a fully routable, usable IP address on the internet.
So first of all, look at our example.
We know right away it's a global unicast address.
2 begins with a 2.
The first four hexadecimal digits represent a total of 16 binary bits.
So here we see 2001.
1 represents four binary bits.
And we also call that a nibble.
Every four hexadecimal digits, we put a colon and then we do another four hexadecimal digits.
So every group of four hexadecimal digits represents 16 binary bits.
You should always have eight.
So we should have eight hex tets divided by colons.
That's how you can always check to see, is your IPv6 address correct?
When you're dealing with IPv6, make sure that you count your hexadecimal digits divided by colons
and make sure you have eight hex tets.
That's how you can always check you have a correct IP address.
IT pros have to learn to simplify IPv6 addresses.
First of all, it reduces errors and it eases the use of an already complex 32-bit hexadecimal
number.
But you have to learn and follow some rules.
We have to learn how to simplify hexadecimal addresses because all operating systems will
simplify these addresses to the lowest, most simplest form that obeys all of the RSE
specifications.
So the first rule is remove leading zeros.
So this is actually a legitimate address.
It's called unspecified address.
And it's a whole lot of zeros.
If you look at the hexadecimal value at the top, that's got four hexadecimal values, even
though they're zeros, divided by colons.
And we have eight of those hex tets.
Removing all the leading zeros, we end up with zero colon, zero colon, zero colon, and
you know the rest.
So when we remove all the leading zeros, we end up with our sample as shown below.
Now our sample number two is actually another real IP version six address.
It's called the loopback.
And it's very similar to the unspecified, except it has a one in the last hex tet.
This is just like we did in IP version four.
We had a loopback IP address that we could ping ourselves.
So this is the same thing.
Now, if I remove all my leading zeros, you can see again, I end up with zero colon, zero
colon, zero colon.
And at the last hex tet, I'm going to end up with a one.
And you can see the most simplified is at the bottom.
My sample number three is a global unicast address.
This one can be used on the internet.
You know that because it begins with a two.
So I have, I can take all the leading zeros in each of the hex tets, and I can remove
all of them except one.
You can see we end up with two zero, zero colon, zero colon, zero colon, zero colon,
all the way to one.
And at the bottom, we see the most simplified form.
Now this is rule number one, removing leading zeros.
So here's one more global IP six address.
And you can see in the blue version of the IP address, we've taken red and highlighted
all the leading zeros.
And so those can be removed.
If you look at our example at the bottom, you can see the fully, the simplified version
of that same IP address.
Now, one thing we cannot do, we cannot remove trailing zeros.
So all of our trailing zeros in each hex tet, you cannot remove.
Don't forget that.
All right.
Rule number two in simplifying an IP version six hexadecimal address is the use of the
double colon.
We can use a double colon to represent a single continuous set of all zero hex tets.
It can only be used once in any IP version six address.
Let's look at sample six.
It's a link local address.
It has an FE80 colon, and then you can see the rest of it.
We see a section of contiguous zeros in our hexadecimal IP address.
Those can be removed and replaced by a double colon.
So you can see here at the bottom, this is a simplified IP version six address where
those contiguous zeros in the hex tet have been removed and replaced by a double colon.
Now, we still have the center section of zeros that could be reduced.
The leading zeros could be reduced to a single zero.
All right.
Let's combine rule number one, which is get rid of leading zeros.
And number two, any contiguous group of hex tets that are zeros, you can replace with
double colons, but you can only use it once.
Let's look at sample seven.
This is a multicast address.
We know that because it begins with FF.
Anytime you see FF in any IP six address, it's a multicast address.
But notice this one's tricky.
We have FF00 colon and then a series of zeros contiguously.
And then we have ACAD and then another set of zeros contiguously.
So we could use the double colon either in the front or in the back of this particular
IP address.
And notice the author has shown this by using option one and option two.
We still have leading zeros that we can get rid of.
So look below.
We have option one where we end up with double colon in the front, taking care of the two
sections of zeros that are the two zero hex tets that are contiguous.
And then at the very bottom, option two, we see that the double colon is used in the rear.
There is an RFC that specifies that the option one is the preferred simplification method.
When working with IP version four, we use CIDR notation or what is also known as a prefix.
So here I've got an IP address 192.100.10.33 CIDR notation or forward slash 28.
And the 28 simply told me how many ones were in the subnet mask.
It helped me quickly identify what portion of my 32 bit address was network and subnet
and what bits were host.
And I had to have that CIDR notation in order to determine that.
There's no way I could determine it without it.
That is going to totally change in IP version six.
Here is the layout of a typical global unicast IP version six address.
The first 48 bits are called the global routing prefix.
The next 16 bits are the subnet ID.
This is typically set up so there's at least 65,000 subnets available to most organizations
when they get an allocation of IP six addresses.
Now look at the last 64.
Remember with CIDR, you used an IP four, you could, you use that CIDR notation to figure
out what bits were host, what bits were network.
Well, looky here, 64 bits are dedicated always to the host bits.
You never guess as to how many bits are your host bits.
It will always be 64.
In fact, you never subnet below 64 ever because the minute you do, you lose a number of significant
IP version six features if you do that.
So IP version six addresses always use prefixes, but remember they never have anything to do
with how many bits are for the host because the host always has 64 bits reserved for itself.
Now, regional internet registries get a prefix of slash 23.
Their left most 23 bits are network.
The rest of the bits are IP addresses that they can allocate to the next level, which
is ISP.
They get a slash 32.
Most large organizations get a slash 48 prefix.
So the left most 48 bits are their network and all the rest of the bits are available
to allocate to host networks, infrastructure, et cetera.
Most medium businesses are going to get a slash 56 prefix.
Small homes typically are going to get a slash 60 and the smallest subnet you can give out
is a slash 64.
Remember the last 64 bits are your host bits.
Let's go back to our pie chart that represents all the IPv6 addresses that we possibly can
use, and that's a total of 340 unidecillion.
I have no idea how big that is.
It's huge.
IANA has allocated already 001, the first three bits of that IPv6 address to a pool
that they can give out so far.
That's 42.5 unidecillion IP addresses.
But we've only given 1% of that to the five regional global internet registries, and they
have 415 dexillion IP addresses already, and that's for at least the first decade.
I don't think we're going to run out of IPv6 addresses for a while.
At the end of this video, I'm going to show you some of the resources, YouTube videos
that I found super, super helpful in helping me get up to speed with IPv6.
IPv6 was one of those things that I just kind of pushed on down the road because I had too
many other things to learn, and I really had to do my homework, and I found these resources
very helpful.
If you're looking for great resources for IPv6, I encourage you to start with Dan's
course's channel.
He has a video called IPv6 Basics for Beginner.
This is an excellent video that covers the fundamentals of IPv6, all the kinds of things
that you really need to know to grasp as you begin to look at IPv6.
He also gets into Slack, which is the stateless, automated configuration of TCPIP information.
So I encourage you, if you want a good video that gets you started, take a look at this
video.
At the top of this slide is a URL.
I encourage you to go there.
This takes you to a channel called ITFreeTraining.
The author has done a really good job on IPv6.
There's just a series of videos.
This series of videos that he has on his channel for IPv6 are very well done.
They're systematic, they're very thorough, they do a good job in presentation, and he
shows a lot of troubleshooting and practicality, command line tools that you can use to test
and verify and configure your IPv6.
A good set of videos, well worth watching.
Charles Kosrock wrote the TCP Guide.
This is an excellent, thorough overview of TCPIP.
This is his website, and you can go to the Read the Guide online.
You can buy a full PDF version of this book.
I would encourage you to go to the Table of Contents right here and use this to navigate
the book.
The website is very dated, but the material covering IPv4, IPv6, almost any topic related
to TCPIP, this is one of the best resources that you can use in your study.