WEBVTT

0:00:03.100000 --> 0:00:08.220000
 All right, so we've already talked about
 this, how to write IPv6 addresses

0:00:08.220000 --> 0:00:10.120000
 so we can skip that.

0:00:10.120000 --> 0:00:15.940000
 Now, just like in the world of IPv4, in
 IPv4, there were different categories

0:00:15.940000 --> 0:00:19.940000
 or different uses for different
 types of addresses.

0:00:19.940000 --> 0:00:23.100000
 For example, link local address.

0:00:23.100000 --> 0:00:30.740000
 Now, in IPv4, the equivalent of this
 was something called an epipa, an

0:00:30.740000 --> 0:00:38.320000
 address, which stood for automatically
 provisioned IP address.

0:00:38.320000 --> 0:00:44.120000
 Now, in the world of IPv4, an epipa
 address typically started with 169

0:00:44.120000 --> 0:00:53.820000
.254.something, these were dynamically
 derived, but 169.254 was your epipa

0:00:53.820000 --> 0:00:58.280000
 address. Now, in IPv4, when you
 saw that, that was a bad thing.

0:00:58.280000 --> 0:00:59.840000
 You didn't want to see that.

0:00:59.840000 --> 0:01:05.740000
 169.254 typically meant, hey, my device
 tried to do DHCP, it sent out

0:01:05.740000 --> 0:01:08.760000
 as DHCP discover, and we
 didn't get anything back.

0:01:08.760000 --> 0:01:09.760000
 We never got an offer back.

0:01:09.760000 --> 0:01:13.540000
 The DHCP server is either down or
 unreachable or something happened.

0:01:13.540000 --> 0:01:18.380000
 So instead of just giving up and not
 having any IP address at all, we

0:01:18.380000 --> 0:01:22.480000
 automatically provisioned ourselves
 with an IP address starting with 169

0:01:22.480000 --> 0:01:27.900000
.254, and then we just dynamically
 generated the last two bytes.

0:01:27.900000 --> 0:01:34.080000
 So that was a bad thing, because that is
 supposed to be a link local address.

0:01:34.080000 --> 0:01:39.000000
 In other words, routers really should
 not route packets coming from or

0:01:39.000000 --> 0:01:45.020000
 going to 169.254, because there's
 no manageability of that.

0:01:45.020000 --> 0:01:46.900000
 That's not an assigned network.

0:01:46.900000 --> 0:01:50.000000
 It's not like, oh, this link over here
 is the only link that's got 169

0:01:50.000000 --> 0:01:55.840000
.254 on it. No, 169.254 could show up
 anywhere in the network, because

0:01:55.840000 --> 0:01:59.300000
 we don't know when hosts are going
 to fail DHCP and when they're going

0:01:59.300000 --> 0:02:00.320000
 to be successful.

0:02:00.320000 --> 0:02:02.560000
 So it's not really routable.

0:02:02.560000 --> 0:02:07.740000
 So if you have a 169.254 address, the
 idea is, okay, maybe on your link,

0:02:07.740000 --> 0:02:10.160000
 everybody in your link is
 having DHCP problems.

0:02:10.160000 --> 0:02:12.260000
 Nobody was able to get DHCP.

0:02:12.260000 --> 0:02:16.260000
 So on your link, on your VLAN, everybody
 ended up automatically provisioning

0:02:16.260000 --> 0:02:18.360000
 themselves with 169.254.

0:02:18.360000 --> 0:02:20.340000
 So at least you can
 talk to each other.

0:02:20.340000 --> 0:02:23.020000
 You can talk on your local link with other
 devices, but you're not supposed

0:02:23.020000 --> 0:02:27.300000
 to be able to be routed beyond your local
 link unless you have a legitimate

0:02:27.300000 --> 0:02:32.540000
 IP network that you got via DHCP
 or that was statically assigned.

0:02:32.540000 --> 0:02:36.280000
 So in IPV4, the APIPA address is a
 link local address, but it's a bad

0:02:36.280000 --> 0:02:38.180000
 thing. That's something you
 need to troubleshoot.

0:02:38.180000 --> 0:02:41.640000
 You need to figure out what
 happened with DHCP.

0:02:41.640000 --> 0:02:47.200000
 Well, in IPV6, every network interface
 card, whether it be on a smartphone,

0:02:47.200000 --> 0:02:51.680000
 a tablet, or a server, comes up with
 a link local address by default,

0:02:51.680000 --> 0:02:56.680000
 it has to. So in IPV6, link local addresses
 are recognized as beginning

0:02:56.680000 --> 0:03:04.040000
 with FE80. And so just like I described
 with IPV4, when a packet is created

0:03:04.040000 --> 0:03:09.340000
 with a link local address as the source
 and or the destination, that is

0:03:09.340000 --> 0:03:14.100000
 not routable. Now the router's interface
 itself has a link local address

0:03:14.100000 --> 0:03:18.680000
 as well. So a lot of times in IPV6, you
 will address packets to your router

0:03:18.680000 --> 0:03:21.700000
 going to his link local address.

0:03:21.700000 --> 0:03:27.840000
 Like there's a lot of under the hood
 mechanics of how IPV6 boots up, how

0:03:27.840000 --> 0:03:32.680000
 it learns of addresses, how it does like
 there's no ARP in IPV6 that does

0:03:32.680000 --> 0:03:37.320000
 not exist, but the equivalent of ARP
 can use link local addresses, but

0:03:37.320000 --> 0:03:40.140000
 it is not a routable address.

0:03:40.140000 --> 0:03:44.140000
 I can't send a packet to you on a completely
 different BLAN or a completely

0:03:44.140000 --> 0:03:48.540000
 different network that's across the
 router to your link local address.

0:03:48.540000 --> 0:03:49.620000
 I just, I can't do it.

0:03:49.620000 --> 0:03:53.880000
 It's only for talking devices
 on my own local link.

0:03:53.880000 --> 0:03:55.160000
 And this is mandatory.

0:03:55.160000 --> 0:03:56.760000
 It's got to be there.

0:03:56.760000 --> 0:04:01.980000
 So whenever you enable IPV6 on an interface,
 only just demonstrate this

0:04:01.980000 --> 0:04:04.860000
 here real quick.

0:04:04.860000 --> 0:04:10.400000
 So not this. I don't care
 about that, but here.

0:04:10.400000 --> 0:04:18.780000
 So for example, if I go onto my Ubuntu
 device, okay, and I say IF config,

0:04:18.780000 --> 0:04:25.200000
 you'll see that it's already by default
 running IP version six, and it

0:04:25.200000 --> 0:04:27.640000
 came up with a link local
 address for itself.

0:04:27.640000 --> 0:04:32.300000
 So your Windows laptop, your Mac laptop,
 if you look on your NIC card,

0:04:32.300000 --> 0:04:33.660000
 chances are pretty good.

0:04:33.660000 --> 0:04:35.200000
 It's running IPV6.

0:04:35.200000 --> 0:04:39.160000
 And even if it doesn't have a globally
 routable address, and we'll talk

0:04:39.160000 --> 0:04:43.680000
 about that in just a second at minimum,
 it does have a link local address.

0:04:43.680000 --> 0:04:54.520000
 Now, we'll also see that if I go onto
 a router, nope, don't want putty.

0:04:54.520000 --> 0:05:00.960000
 You want secure CRT.

0:05:00.960000 --> 0:05:11.500000
 All right. So if I go onto a router,
 and I go to do show IP interface

0:05:11.500000 --> 0:05:17.120000
 brief, let's say I go to interface
 gigabit zero slash zero.

0:05:17.120000 --> 0:05:25.480000
 So there's two ways a link local address
 can be automatically enabled

0:05:25.480000 --> 0:05:26.780000
 on your interface.

0:05:26.780000 --> 0:05:31.580000
 One thing you could do is you could just
 give yourself a real IPV6 address.

0:05:31.580000 --> 0:05:41.640000
 Okay. So I've just given myself a
 globally routable IPV6 address.

0:05:41.640000 --> 0:05:51.320000
 Now if I do show IPV6 interface brief,
 you'll see not only does it have

0:05:51.320000 --> 0:05:54.840000
 that address I configured, it also dynamically
 gave itself a link local

0:05:54.840000 --> 0:05:58.800000
 address. The moment it said,
 oh, he wants to do IPV6 here.

0:05:58.800000 --> 0:06:02.240000
 It said, I need to derive a link
 local address for myself.

0:06:02.240000 --> 0:06:09.500000
 Another way you could do it is go to an
 interface and just say IPV6 enable.

0:06:09.500000 --> 0:06:14.060000
 This means, hey, I'm not going
 to give you a global address.

0:06:14.060000 --> 0:06:15.960000
 I just want you to have
 a link local address.

0:06:15.960000 --> 0:06:20.260000
 And now if we look at interface gigabit
 zero slash one, we see just that.

0:06:20.260000 --> 0:06:23.540000
 He doesn't have any global address, but
 he does have a link local address

0:06:23.540000 --> 0:06:28.920000
 because I did the command
 IPV6 enable on him.

0:06:28.920000 --> 0:06:34.920000
 Global unit cast addresses.

0:06:34.920000 --> 0:06:40.080000
 So technically if we think of the 128
 bits, the make up our IPV6 address,

0:06:40.080000 --> 0:06:44.820000
 any IPV6 address were the very first
 three bits going from left to right

0:06:44.820000 --> 0:06:49.460000
 are this pattern right there, zero zero
 one, and then we don't care what

0:06:49.460000 --> 0:06:56.500000
 the remaining bits are, then that
 is a global unit cast address.

0:06:56.500000 --> 0:07:00.220000
 Technically, if we think about our first
 nibble as being the first four

0:07:00.220000 --> 0:07:03.540000
 bits, that gives us two possible
 patterns, right?

0:07:03.540000 --> 0:07:08.380000
 Zero zero one zero and zero zero one one,
 which means our very first hexadecimal

0:07:08.380000 --> 0:07:11.280000
 character could be either
 a two or a three.

0:07:11.280000 --> 0:07:15.560000
 Well, the internet assigned numbers
 authority has said three is reserved

0:07:15.560000 --> 0:07:18.280000
 for future use. Can't use that.

0:07:18.280000 --> 0:07:22.580000
 So all globally routable IPV6 addresses
 right now and for the foreseeable

0:07:22.580000 --> 0:07:25.980000
 future are going to start
 with the number two.

0:07:25.980000 --> 0:07:32.640000
 And like I said, most likely they will
 have a slash 64 as their subnet

0:07:32.640000 --> 0:07:41.620000
 mask. In the world of IPV4, we had private
 and public networks, private

0:07:41.620000 --> 0:07:46.660000
 networks were like the 10 network,
 the 172 16, the 192 168.

0:07:46.660000 --> 0:07:50.140000
 Well in the world of IPV6,
 it's a lot easier.

0:07:50.140000 --> 0:07:52.000000
 Anything with FC.

0:07:52.000000 --> 0:07:56.960000
 FC is your private, what we now
 call unique local addresses.

0:07:56.960000 --> 0:08:01.580000
 Also just so you know, technically
 FD is also private.

0:08:01.580000 --> 0:08:06.940000
 If you ever see FC or FD,
 that is a unique local.

0:08:06.940000 --> 0:08:10.840000
 Originally when they first came out,
 they were called site local.

0:08:10.840000 --> 0:08:14.940000
 So if you're reading some older documentation,
 you might see that as well.

0:08:14.940000 --> 0:08:16.720000
 They both mean the
 exact same thing.

0:08:16.720000 --> 0:08:19.120000
 These are addresses that are
 not globally routable.

0:08:19.120000 --> 0:08:26.300000
 You can't send them to
 the internet FC and FD.

0:08:26.300000 --> 0:08:30.620000
 Multicast addresses in the world of
 IP version four, you had to remember

0:08:30.620000 --> 0:08:34.920000
 the class D address space in
 the world of IP version four.

0:08:34.920000 --> 0:08:42.240000
 You had to remember that multicast being
 class D start out with 224 dot

0:08:42.240000 --> 0:08:46.740000
 anything going up to
 239 dot anything.

0:08:46.740000 --> 0:08:50.300000
 You had to remember that range
 IPV6, it's a lot simpler.

0:08:50.300000 --> 0:08:54.800000
 Any address beginning with
 FF, that is multicast.

0:08:54.800000 --> 0:08:58.440000
 FF has been reserved for
 multicast addresses.

0:08:58.440000 --> 0:09:07.660000
 Also knows here that it says IPV6 nodes
 and an IPV6 node is anything running

0:09:07.660000 --> 0:09:12.540000
 IPV6, smartphone, tablet, router,
 switch, PC, whatever.

0:09:12.540000 --> 0:09:16.960000
 Listen to several IPV6 multicast groups
 by default and you should recognize

0:09:16.960000 --> 0:09:19.780000
 these. In the next slide, I'm going
 to show you two very common ones.

0:09:19.780000 --> 0:09:21.800000
 You should be able to recognize.

0:09:21.800000 --> 0:09:25.380000
 This is the all IPV6 host address.


0:09:25.380000 --> 0:09:27.180000
 Now here's an interesting thing.

0:09:27.180000 --> 0:09:32.280000
 In IPV6, they said, we're not
 going to do broadcast anymore.

0:09:32.280000 --> 0:09:33.240000
 So think about this.

0:09:33.240000 --> 0:09:40.720000
 In the world of IPV4, let's say that
 we had a switch right here and we

0:09:40.720000 --> 0:09:43.360000
 had several devices
 on that switch.

0:09:43.360000 --> 0:09:44.220000
 Maybe here's the router.

0:09:44.220000 --> 0:09:47.120000
 Here's the default gateway that
 everybody's going to use.

0:09:47.120000 --> 0:09:51.000000
 Then we have some PCs
 connected to this.

0:09:51.000000 --> 0:09:55.540000
 This was all in VLAN, whatever.

0:09:55.540000 --> 0:10:03.120000
 Maybe VLAN 2. Let's say maybe we assign
 the 2.2.2.0 slash 24 subnet to

0:10:03.120000 --> 0:10:04.640000
 everything that VLAN.

0:10:04.640000 --> 0:10:11.860000
 Now, if this host right here wanted
 to send a broadcast to everybody,

0:10:11.860000 --> 0:10:18.360000
 most likely what he would do
 is send it to this address.

0:10:18.360000 --> 0:10:24.940000
 That's what we call the general
 broadcast address.

0:10:24.940000 --> 0:10:26.660000
 You send a packet to that.

0:10:26.660000 --> 0:10:28.740000
 It's supposed to go to everyone.

0:10:28.740000 --> 0:10:34.260000
 Now we know in reality, everyone really
 means everybody on this side of

0:10:34.260000 --> 0:10:37.520000
 the router, because routers
 do not forward broadcast.

0:10:37.520000 --> 0:10:40.760000
 If I send a packet to that, it's not
 going to go to every conceivable

0:10:40.760000 --> 0:10:44.420000
 network in my company once
 it gets to the router.

0:10:44.420000 --> 0:10:47.120000
 The router will process it
 and then he'll drop it.

0:10:47.120000 --> 0:10:48.400000
 It won't go any further than him.

0:10:48.400000 --> 0:10:51.300000
 It's really for broadcasting on your
 own local network, but we call it

0:10:51.300000 --> 0:10:52.940000
 the general broadcast.

0:10:52.940000 --> 0:10:58.240000
 But there was also another way to send
 a packet to everybody on my network,

0:10:58.240000 --> 0:11:01.840000
 which was what's called the directed
 broadcast, which means I'm going

0:11:01.840000 --> 0:11:04.380000
 to send a packet to
 my network address.

0:11:04.380000 --> 0:11:09.160000
 Then I'm just going to fill in
 the host bits with all ones.

0:11:09.160000 --> 0:11:14.040000
 That's called the directed broadcast,
 where you're using the network bits

0:11:14.040000 --> 0:11:17.460000
 and the host bits are
 all just set to once.

0:11:17.460000 --> 0:11:18.820000
 That's the directed broadcast.

0:11:18.820000 --> 0:11:21.560000
 It does the exact same
 thing as this.

0:11:21.560000 --> 0:11:25.060000
 Everybody on your wire will see
 it, including the router.

0:11:25.060000 --> 0:11:26.560000
 Now that's not as common.

0:11:26.560000 --> 0:11:32.420000
 Most protocols that use broadcasts like
 DHCP, ARP, they use the general

0:11:32.420000 --> 0:11:35.680000
 broadcast. It's probably pretty rare
 that you'll see anything creating

0:11:35.680000 --> 0:11:38.980000
 a packet going to the
 directed broadcast.

0:11:38.980000 --> 0:11:41.580000
 What does this have
 to do with IPv6?

0:11:41.580000 --> 0:11:45.000000
 Well when IPv6 came out, they
 said, no more broadcasts.

0:11:45.000000 --> 0:11:48.480000
 There's no equivalent
 of this in IPv6.

0:11:48.480000 --> 0:11:51.620000
 Well, yes and no, because
 guess what?

0:11:51.620000 --> 0:11:57.320000
 If I send a packet in IPv6 to this
 multicast address, this is the all

0:11:57.320000 --> 0:12:02.520000
 IPv6 host address, which means that
 all IPv6 hosts have to be listening

0:12:02.520000 --> 0:12:03.780000
 to it. So guess what?

0:12:03.780000 --> 0:12:05.420000
 It's going to do the
 exact same thing.

0:12:05.420000 --> 0:12:10.060000
 Every host in my VLAN, including my
 router, is listening automatically

0:12:10.060000 --> 0:12:11.920000
 to that multicast address.

0:12:11.920000 --> 0:12:16.140000
 It's the same effect as if
 I was sending a broadcast.

0:12:16.140000 --> 0:12:19.600000
 So yeah, they got kind of tricky by saying,
 oh, there's no broadcast anymore.

0:12:19.600000 --> 0:12:24.620000
 Well, yeah, that's true, but sending
 a packet to FF02 colon colon one

0:12:24.620000 --> 0:12:28.420000
 for all intents and purposes,
 you're sending a broadcast.

0:12:28.420000 --> 0:12:31.480000
 Yeah, technically we call them multicast
 because it starts out with FF.

0:12:31.480000 --> 0:12:36.800000
 But hey, if I have a multicast that
 everybody's listening to, if I send

0:12:36.800000 --> 0:12:40.100000
 a packet to that, it's going to get
 to everybody, which is kind of like

0:12:40.100000 --> 0:12:41.600000
 what a broadcast does.

0:12:41.600000 --> 0:12:45.880000
 And then there's another special multicast
 address that only routers listen

0:12:45.880000 --> 0:12:51.800000
 to, which is called the all
 IPv6 routers address.

0:12:51.800000 --> 0:12:57.500000
 So if a host like a laptop or a PC
 needs to discover its router, if it

0:12:57.500000 --> 0:13:01.160000
 doesn't know if a router exists, it says,
 hey, is there a router out there?

0:13:01.160000 --> 0:13:05.060000
 It'll start by sending it to
 this FF02 colon colon two.

0:13:05.060000 --> 0:13:08.500000
 And if there's an IPV6 router on your
 segment, it has to be listening

0:13:08.500000 --> 0:13:14.380000
 to that. And notice it says this
 is used with IPV6 slack.

0:13:14.380000 --> 0:13:15.700000
 You need to know what
 that term is.

0:13:15.700000 --> 0:13:19.100000
 That's the stateless automatic
 address configuration.

0:13:19.100000 --> 0:13:23.600000
 This is something that IPV6 can
 do that IPV4 could not do.

0:13:23.600000 --> 0:13:28.340000
 In the world of IPV4, your host, let's
 just say your laptop, there was

0:13:28.340000 --> 0:13:32.700000
 really only two ways your laptop
 could get an IPV4 address.

0:13:32.700000 --> 0:13:37.160000
 You could go into it, go into the control
 panel or system preferences,

0:13:37.160000 --> 0:13:41.160000
 go to your nip card and you could statically
 assign an IPV4 address and

0:13:41.160000 --> 0:13:45.880000
 subnet mask. Not advisable unless you're
 dealing with like a server, something

0:13:45.880000 --> 0:13:48.300000
 that never moved that
 had a static address.

0:13:48.300000 --> 0:13:49.500000
 But you could do that.

0:13:49.500000 --> 0:13:54.240000
 The other alternative was using DHCP,
 right, which is what most mobile

0:13:54.240000 --> 0:13:57.460000
 hosts like laptops and smartphones
 and tablets used.

0:13:57.460000 --> 0:13:58.720000
 DHCP. That was it.

0:13:58.720000 --> 0:14:00.740000
 Those were your two alternatives.

0:14:00.740000 --> 0:14:03.140000
 In IPV6, you do have
 those two methods.

0:14:03.140000 --> 0:14:05.860000
 You can statically assign
 an IPV6 address.

0:14:05.860000 --> 0:14:10.000000
 There is DHCP version six, so you
 should know how that works.

0:14:10.000000 --> 0:14:13.920000
 But there's a third alternative as well
 called stateless automatic address

0:14:13.920000 --> 0:14:17.760000
 configuration. It's actually
 pretty cool the way it works.

0:14:17.760000 --> 0:14:20.640000
 So imagine that this
 is my laptop here.

0:14:20.640000 --> 0:14:23.960000
 He's probably connected to a switch
 and we'll just say the switch is a

0:14:23.960000 --> 0:14:28.080000
 layer two switch for all intents and
 purposes and a switch is connected

0:14:28.080000 --> 0:14:31.240000
 to a router. Okay.

0:14:31.240000 --> 0:14:34.660000
 Well, in stateless automatic address configuration,
 first thing that happens

0:14:34.660000 --> 0:14:39.180000
 is that host, when I bring up his nip
 card, let's say the host comes online,

0:14:39.180000 --> 0:14:42.380000
 he's going to give himself
 an FE80 address, right?

0:14:42.380000 --> 0:14:45.040000
 He's going to give himself
 a link local address.

0:14:45.040000 --> 0:14:50.000000
 Then he's going to send out a
 special type of ICMP message.

0:14:50.000000 --> 0:14:52.680000
 Remember ICMP? Most people
 think of that as ping.

0:14:52.680000 --> 0:15:00.620000
 Well, there's ICMP before which uses
 IPV4 addresses and ICMP V6, which

0:15:00.620000 --> 0:15:03.080000
 uses IPV6 addresses.

0:15:03.080000 --> 0:15:08.760000
 Now, in ICMP V4, we had the echo request,
 echo response that was for ping.

0:15:08.760000 --> 0:15:11.220000
 We also had ICMP redirects.

0:15:11.220000 --> 0:15:14.820000
 We had ICMP host unreachable.

0:15:14.820000 --> 0:15:18.380000
 We had all kinds of
 ICMP message types.

0:15:18.380000 --> 0:15:22.980000
 Well, in ICMP V6, that
 still exists, right?

0:15:22.980000 --> 0:15:29.500000
 IPV6 ping, you're doing ICMP V6,
 echo request and echo response.

0:15:29.500000 --> 0:15:34.900000
 There's ICMP V6 redirects and
 host unreachable messages.

0:15:34.900000 --> 0:15:41.220000
 And there are four new types of ICMP
 messages that only exist in ICMP

0:15:41.220000 --> 0:16:01.760000
 V6. They are the neighbor solicitation
 and neighbor advertisement.

0:16:01.760000 --> 0:16:07.660000
 These are the replacements
 for ARP.

0:16:07.660000 --> 0:16:12.760000
 If I need to find, if I know that a host
 I'm trying to reach like my router

0:16:12.760000 --> 0:16:17.020000
 is on the same network as me, the same
 prefix as me, and I need to discover

0:16:17.020000 --> 0:16:21.980000
 his MAC address, I will send him
 a neighbor solicitation message.

0:16:21.980000 --> 0:16:25.620000
 He will send back to me a neighbor
 advertisement message.

0:16:25.620000 --> 0:16:27.480000
 So that's the replacement for ARP.


0:16:27.480000 --> 0:16:36.680000
 And then we also have
 router solicitation.

0:16:36.680000 --> 0:16:41.620000
 And router advertisement.

0:16:41.620000 --> 0:16:46.560000
 And this is what's used in stateless
 automatic address configuration in

0:16:46.560000 --> 0:16:54.140000
 Slack. So once I've got my link local address,
 I will send a router solicitation.

0:16:54.140000 --> 0:16:57.280000
 And I will send it
 to this address.

0:16:57.280000 --> 0:16:59.660000
 Because I don't know if there's
 a router out there or not.

0:16:59.660000 --> 0:17:03.940000
 I'm assuming there is, but if there is
 one, I know he's listening to FF02

0:17:03.940000 --> 0:17:05.100000
 colon colon two.

0:17:05.100000 --> 0:17:08.360000
 So I'll send my router solicitation
 to him and say, hey, if there's a

0:17:08.360000 --> 0:17:13.580000
 router out there, tell me about yourself
 and tell me what global prefix

0:17:13.580000 --> 0:17:19.000000
 beginning with two is on this wire because
 I need an address for myself.

0:17:19.000000 --> 0:17:23.940000
 The router will send back to
 me a router advertisement.

0:17:23.940000 --> 0:17:26.320000
 And that's where he'll say, hey,
 let me tell you about myself.

0:17:26.320000 --> 0:17:29.200000
 Let me tell you about my link
 local address, what I've got.

0:17:29.200000 --> 0:17:32.800000
 Let me tell you about
 my global address.

0:17:32.800000 --> 0:17:34.160000
 Whatever that is.

0:17:34.160000 --> 0:17:40.580000
 Oh, and by the way, the prefix that
 we're using on this wire is this.

0:17:40.580000 --> 0:17:44.400000
 That's the prefix we're using.

0:17:44.400000 --> 0:17:48.680000
 Now that host has got a third
 way to learn of the prefix.

0:17:48.680000 --> 0:17:50.420000
 He now knows what the
 global prefix is.

0:17:50.420000 --> 0:17:52.300000
 He can assign that
 to his NIC card.

0:17:52.300000 --> 0:17:55.360000
 And the remaining 64
 bits is interface ID.

0:17:55.360000 --> 0:17:58.700000
 He can dynamically come
 up with himself.

0:17:58.700000 --> 0:18:02.640000
 So that's a method that has no equivalent
 in the world of IPV 4.

0:18:02.640000 --> 0:18:05.560000
 We call this stateless automatic
 address configuration.

0:18:05.560000 --> 0:18:10.340000
 This exchange of router solicitation and
 router advertisement, which allows

0:18:10.340000 --> 0:18:14.020000
 hosts to dynamically
 configure themselves.

0:18:14.020000 --> 0:18:22.180000
 Now, the reason we call this stateless
 is because DHCP V6 is state full.

0:18:22.180000 --> 0:18:23.460000
 What do we mean by that?

0:18:23.460000 --> 0:18:27.060000
 Well, just like in the world of IPV
 4, if I reach out to a DHCP server

0:18:27.060000 --> 0:18:31.540000
 with a discover message and he sends
 me an offer, once that four step

0:18:31.540000 --> 0:18:35.240000
 process is done, that
 server knows me.

0:18:35.240000 --> 0:18:36.640000
 He's got a record of me.

0:18:36.640000 --> 0:18:42.760000
 He knows that he has allocated to me
 the IPV 4 address of 2.2.2.2.2 for

0:18:42.760000 --> 0:18:46.180000
 whatever the lease is,
 maybe for 24 hours.

0:18:46.180000 --> 0:18:49.460000
 So I could go to that DHCP server and
 I could see all the IP addresses

0:18:49.460000 --> 0:18:52.780000
 he's assigned to everybody and
 how long they're good for.

0:18:52.780000 --> 0:18:55.140000
 We have state information
 about that.

0:18:55.140000 --> 0:18:57.200000
 That's called a stateful address.

0:18:57.200000 --> 0:19:02.660000
 In Slack, the router has no idea, once
 this process happens with router

0:19:02.660000 --> 0:19:06.020000
 solicitation, router advertisement,
 the router doesn't keep a record of

0:19:06.020000 --> 0:19:10.140000
 that. He doesn't remember that
 he gave me that prefix.

0:19:10.140000 --> 0:19:12.680000
 He doesn't know what address
 I've been given.

0:19:12.680000 --> 0:19:14.660000
 And there's no lifetime on this.

0:19:14.660000 --> 0:19:19.000000
 Once I do Slack, that addresses good
 for as long as I'm connected to that

0:19:19.000000 --> 0:19:22.340000
 network. If I'm connected to that network
 for the next week, I will still

0:19:22.340000 --> 0:19:26.700000
 use that address that I dynamically
 derived via the Slack process.

0:19:26.700000 --> 0:19:28.240000
 And no one's keeping track of it.

0:19:28.240000 --> 0:19:28.880000
 There's no state.

0:19:28.880000 --> 0:19:33.060000
 So that's why we call it stateless
 address configuration or automatic

0:19:33.060000 --> 0:19:35.340000
 address configuration.

0:19:35.340000 --> 0:19:38.500000
 Now, what about the interface ID?

0:19:38.500000 --> 0:19:40.520000
 I just got a global prefix.

0:19:40.520000 --> 0:19:42.860000
 How do I dynamically come
 up with interface ID?

0:19:42.860000 --> 0:19:45.680000
 And that brings us to
 our next section.

0:19:45.680000 --> 0:19:52.380000
 In IPV 4, if I statically configured
 an IPV 4 address on my laptop, I

0:19:52.380000 --> 0:19:57.400000
 had to statically type in both the prefix
 and the interface ID or in IPV

0:19:57.400000 --> 0:20:00.460000
 4, we said the network bits
 and the host bits, right?

0:20:00.460000 --> 0:20:03.860000
 I couldn't go into my laptop and I
 couldn't say, hey, your network is

0:20:03.860000 --> 0:20:08.360000
 2020, figure out for yourself what
 the last eight bits are going to be

0:20:08.360000 --> 0:20:09.500000
 for your host bits.

0:20:09.500000 --> 0:20:10.640000
 That didn't work.

0:20:10.640000 --> 0:20:13.380000
 If I statically configured an address
 on something, I had to give it the

0:20:13.380000 --> 0:20:17.500000
 whole 32-bit address and then manually
 give it the mask so it knew where

0:20:17.500000 --> 0:20:18.780000
 the dividing line was.

0:20:18.780000 --> 0:20:23.340000
 I could use DHCP, but even with DHCP,
 what did I get from the server?

0:20:23.340000 --> 0:20:27.460000
 I got the whole 32-bit address
 and a subnet mask.

0:20:27.460000 --> 0:20:32.520000
 Well, in IPV 6, like with stateless
 automatic address configuration, I'm

0:20:32.520000 --> 0:20:37.000000
 getting the prefix and I have to come
 up with myself what the last 64

0:20:37.000000 --> 0:20:39.960000
 bits are of the interface ID.

0:20:39.960000 --> 0:20:44.920000
 Now, originally the way that this worked
 was using something called EUI

0:20:44.920000 --> 0:20:51.520000
 64. So EUI 64 worked this way and
 you definitely want to know this.

0:20:51.520000 --> 0:20:53.520000
 It was based on your MAC address.

0:20:53.520000 --> 0:21:00.420000
 So let's say that I figured
 out that my prefix was this.

0:21:00.420000 --> 0:21:08.260000
 All right, that's
 my global prefix.

0:21:08.260000 --> 0:21:13.060000
 Either I got that via DHCP, I got that
 via Slack, or maybe I went into

0:21:13.060000 --> 0:21:17.880000
 my network interface card and I said,
 IPV 6 network is this, and then

0:21:17.880000 --> 0:21:21.500000
 I told it EUI 64.

0:21:21.500000 --> 0:21:25.720000
 All right, this is the indicator that
 my system needs to dynamically derive

0:21:25.720000 --> 0:21:31.880000
 via the EUI 64 process, the last
 64 bits of its interface ID.

0:21:31.880000 --> 0:21:33.460000
 So here's how it's going to do it.


0:21:33.460000 --> 0:21:36.280000
 It's going to look around and
 say, okay, what's my net card?

0:21:36.280000 --> 0:21:38.060000
 What's the MAC address on there?

0:21:38.060000 --> 0:21:47.860000
 Maybe my MAC address is 002111222.


0:21:47.860000 --> 0:21:53.200000
 All right, so there's
 my 48-bit MAC.

0:21:53.200000 --> 0:21:58.600000
 The EUI 64 process says, okay, we need
 to expand this to a 64-bit number

0:21:58.600000 --> 0:22:01.000000
 so we can shove it into
 our interface ID.

0:22:01.000000 --> 0:22:01.720000
 So here's what we're going to do.

0:22:01.720000 --> 0:22:04.460000
 We're going to cut that MAC
 address right in half.

0:22:04.460000 --> 0:22:11.780000
 All right, and we're going
 to say 002 colon 11.

0:22:11.780000 --> 0:22:13.480000
 So there's the first half.

0:22:13.480000 --> 0:22:17.860000
 And then we're going to
 have a colon right here.

0:22:17.860000 --> 0:22:22.420000
 Then we're going to
 say 11 colon 2222.

0:22:22.420000 --> 0:22:25.800000
 All right, now we're
 missing 16 bits.

0:22:25.800000 --> 0:22:29.700000
 To expand from 48 to 64 bits,
 that's a difference of 16.

0:22:29.700000 --> 0:22:33.660000
 So we got to shove 16 bits in the
 middle here to expand this.

0:22:33.660000 --> 0:22:35.020000
 How do we do that?

0:22:35.020000 --> 0:22:37.920000
 Well, EUI 64 says, here's
 what we're going to do.

0:22:37.920000 --> 0:22:42.300000
 We're going to put F, F, so
 that's two nibbles, right?

0:22:42.300000 --> 0:22:45.880000
 That's eight bits, F, E.

0:22:45.880000 --> 0:22:47.420000
 There's another eight bits.

0:22:47.420000 --> 0:22:50.920000
 So now we've just added a 16
-bit pattern in the middle.

0:22:50.920000 --> 0:23:02.680000
 So now we've got 002
 colon 11FF colon F.

0:23:02.680000 --> 0:23:05.620000
 E 11 colon 2222.

0:23:05.620000 --> 0:23:07.680000
 But we're not quite done yet.

0:23:07.680000 --> 0:23:08.900000
 We're almost done.

0:23:08.900000 --> 0:23:13.940000
 So now we've taken our 48-bit MAC address
 and we've expanded it to a 64

0:23:13.940000 --> 0:23:17.460000
-bit interface ID by putting this
 pattern of FF FE in the middle.

0:23:17.460000 --> 0:23:20.240000
 Now, don't ask me, why do
 they choose that pattern?

0:23:20.240000 --> 0:23:23.380000
 Why not A123? I have no idea.

0:23:23.380000 --> 0:23:24.520000
 It is what it is.

0:23:24.520000 --> 0:23:25.900000
 You just have to memorize that.

0:23:25.900000 --> 0:23:28.740000
 Now, here's one other thing that's
 like, huh, why do they do that?

0:23:28.740000 --> 0:23:30.400000
 I can't really explain it.

0:23:30.400000 --> 0:23:31.180000
 I've Googled it.

0:23:31.180000 --> 0:23:34.980000
 I've never found anybody who explained
 in plain English why they did this.

0:23:34.980000 --> 0:23:37.700000
 So just another freaky little thing
 they did and we just have to know

0:23:37.700000 --> 0:23:40.900000
 what it is. So here's what we do.

0:23:40.900000 --> 0:23:46.140000
 We take our first byte, which is 000,
 000, so I'm expanding it into binary

0:23:46.140000 --> 0:23:50.980000
 now. 000, 000. So that's
 our first two zeros.

0:23:50.980000 --> 0:23:52.080000
 And here's what we do.

0:23:52.080000 --> 0:23:53.040000
 You're going to say, what?

0:23:53.040000 --> 0:23:54.380000
 I'm going to say,
 oh, live with it.

0:23:54.380000 --> 0:23:55.060000
 It's just the way it is.

0:23:55.060000 --> 0:23:59.140000
 And now we're going to take
 this bit right there.

0:23:59.140000 --> 0:24:02.660000
 Whatever that bit is, flip
 it to its opposite.

0:24:02.660000 --> 0:24:04.980000
 So if it's zero, flip it to a one.


0:24:04.980000 --> 0:24:10.720000
 If it was a one, flip it to a zero,
 which ends up giving us 000, 000,

0:24:10.720000 --> 0:24:17.860000
 000, 000, 10. So now our final
 interface ID is not 0002.

0:24:17.860000 --> 0:24:24.480000
 It ends up being 0202,
 102, 11FF FE11, 222.

0:24:24.480000 --> 0:24:29.180000
 Because we had to flip that bit that
 was in this nibble right there.

0:24:29.180000 --> 0:24:32.540000
 Take a look at that nibble,
 flip that bit.

0:24:32.540000 --> 0:24:38.860000
 And that gives us our EUI64 interface
 ID based on the MAC address.

0:24:38.860000 --> 0:24:49.000000
 All right, and we've already talked
 about this, the neighbor discovery

0:24:49.000000 --> 0:24:53.260000
 protocol. So I didn't actually give it
 this name, but I did tell you that

0:24:53.260000 --> 0:24:57.620000
 with ICMP, there were four
 new ICMP message types.

0:24:57.620000 --> 0:24:58.920000
 So we talked about those.

0:24:58.920000 --> 0:25:04.520000
 Those four message types taken together
 as a group is what we call the

0:25:04.520000 --> 0:25:07.560000
 neighbor discovery process, NDP.

0:25:07.560000 --> 0:25:11.840000
 So if you ever get a question on what
 is comprised of the neighbor discovery

0:25:11.840000 --> 0:25:16.520000
 protocol, you know, oh, those
 are four ICMPB six messages.

0:25:16.520000 --> 0:25:18.060000
 And these are the messages
 right here.

0:25:18.060000 --> 0:25:21.840000
 Neighbor solicitation, neighbor advertisement,
 that's the replacement

0:25:21.840000 --> 0:25:26.260000
 for ARP. Router solicitation, router
 advertisement, that is brand spanking

0:25:26.260000 --> 0:25:31.860000
 new. That's what we use for Slack.


0:25:31.860000 --> 0:25:40.460000
 All right, and that concludes this
 video refresher on an intro on IPv6

0:25:40.460000 --> 0:25:43.260000
 and addressing types and
 all that good stuff.

0:25:43.260000 --> 0:25:45.300000
 I hope you found this
 video informative.

