WEBVTT

0:00:02.060000 --> 0:00:05.740000
 So in this video, I'd like to talk
 about one of the ways of connecting

0:00:05.740000 --> 0:00:10.780000
 a router, or I should say a routing
 device to a VLAN, and this way is

0:00:10.780000 --> 0:00:12.960000
 called router on a stick.

0:00:12.960000 --> 0:00:17.980000
 So I'm going to talk about how we introduce
 a router to a VLAN, and what

0:00:17.980000 --> 0:00:19.860000
 are the requirements for that?

0:00:19.860000 --> 0:00:24.100000
 What does the router need to be able
 to do in order to be able to route

0:00:24.100000 --> 0:00:27.560000
 between VLANs? And we'll also look
 at how do we follow a frame?

0:00:27.560000 --> 0:00:31.040000
 When a frame is actually using this
 particular method, what does that

0:00:31.040000 --> 0:00:35.320000
 look like? What changes occur in that
 frame as it goes from one VLAN into

0:00:35.320000 --> 0:00:38.420000
 the other? And we'll look
 at those details.

0:00:38.420000 --> 0:00:45.880000
 All right, so we know that in order
 to have frames, I should say data

0:00:45.880000 --> 0:00:52.340000
 from one VLAN go into another VLAN,
 we need something connected to both

0:00:52.340000 --> 0:00:55.360000
 of those VLANs that's capable
 of routing that traffic.

0:00:55.360000 --> 0:01:00.020000
 Something that's capable of receiving
 an incoming Ethernet frame, and

0:01:00.020000 --> 0:01:04.240000
 then changing some of the Ethernet headers,
 changing some of the IP headers,

0:01:04.240000 --> 0:01:08.520000
 and then routing that frame out a different
 interface into a different

0:01:08.520000 --> 0:01:15.740000
 VLAN. So in order for a physical router
 to route between VLANs, what must

0:01:15.740000 --> 0:01:21.100000
 be true? Number one, the router has to
 have unique IP addresses appropriate

0:01:21.100000 --> 0:01:24.160000
 for each VLAN subnet.

0:01:24.160000 --> 0:01:28.280000
 We've learned that each VLAN needs
 a unique subnet because the devices

0:01:28.280000 --> 0:01:33.440000
 within a VLAN need to know that their
 subnet is unique and only reachable

0:01:33.440000 --> 0:01:37.140000
 within their circle, within their
 group, within their VLAN.

0:01:37.140000 --> 0:01:43.300000
 So that means that within that VLAN,
 we've got a unique subnet, and one

0:01:43.300000 --> 0:01:47.320000
 of the IP addresses, one of the host
 addresses from that subnet needs

0:01:47.320000 --> 0:01:51.420000
 to be put on the router's interface
 that's connecting to that VLAN.

0:01:51.420000 --> 0:01:55.400000
 So that router, here's something I
 say to a lot of my students, if an

0:01:55.400000 --> 0:02:00.080000
 interface, like an interface on your
 laptop or an interface on a router,

0:02:00.080000 --> 0:02:05.800000
 if you expect that interface to be able
 to create packets or to receive

0:02:05.800000 --> 0:02:11.360000
 and respond to packets, that interface
 must have a layer three address

0:02:11.360000 --> 0:02:15.420000
 in that same language,
 in that same protocol.

0:02:15.420000 --> 0:02:18.900000
 In other words, a router when you take
 it out of the box and you plug

0:02:18.900000 --> 0:02:23.100000
 it in and you turn it on, has no IP addresses,
 IP version four or IP version

0:02:23.100000 --> 0:02:24.980000
 six, it has none.

0:02:24.980000 --> 0:02:29.560000
 So if I simply stick a cable into that
 router connecting to my laptop,

0:02:29.560000 --> 0:02:33.840000
 and then I bring up that router's interface
 with the command no shut down.

0:02:33.840000 --> 0:02:38.900000
 If an IP packet, let's say IP version
 four, if an IP version four packet

0:02:38.900000 --> 0:02:42.380000
 hits that router's interface, that
 router's interface is not going to

0:02:42.380000 --> 0:02:46.520000
 understand it, because without an IP
 version four address of its own,

0:02:46.520000 --> 0:02:51.600000
 a unique host address, it's not speaking
 IP version four, it doesn't understand

0:02:51.600000 --> 0:02:54.240000
 that language, it'll have
 to drop that packet.

0:02:54.240000 --> 0:02:58.620000
 Similarly, if I was to send an IPV six
 packet to that router's interface,

0:02:58.620000 --> 0:03:03.320000
 it needs to have an IPV six address
 to be able to process and recognize

0:03:03.320000 --> 0:03:09.080000
 that packet. So the same thing is true,
 if we're doing router on a stick,

0:03:09.080000 --> 0:03:12.080000
 well let's just say a router connected
 to VLANs, let's forget about that

0:03:12.080000 --> 0:03:13.440000
 term router on a stick
 for a minute.

0:03:13.440000 --> 0:03:18.460000
 If I have a router's interface with
 a cable that's going into a switch,

0:03:18.460000 --> 0:03:24.440000
 all right, so here's my router, here's
 my switch, and let's say I've got

0:03:24.440000 --> 0:03:28.780000
 a bunch of interfaces here on some
 VLAN, VLAN 12 is what all these are

0:03:28.780000 --> 0:03:33.240000
 on, and let's say I've decided that
 all these devices are going to be

0:03:33.240000 --> 0:03:40.920000
 in the 12.1.1 slash 24 network.

0:03:40.920000 --> 0:03:44.520000
 Well, if I may have a router here connecting
 to the switch who's also

0:03:44.520000 --> 0:03:50.540000
 going to be in VLAN 12, his interface
 also needs to have an IP address

0:03:50.540000 --> 0:03:58.580000
 that is in that same network,
 12.1.1, something.

0:03:58.580000 --> 0:04:02.780000
 That way he not only understands the
 internet protocol and can send and

0:04:02.780000 --> 0:04:07.980000
 receive packets, he also recognizes
 that he is in that same group.

0:04:07.980000 --> 0:04:11.620000
 So if we have a router that's going to
 be routing between multiple VLANs,

0:04:11.620000 --> 0:04:14.340000
 we're talking about a router that's
 going to have to have multiple IP

0:04:14.340000 --> 0:04:17.980000
 addresses, each one in
 a different subnet.

0:04:17.980000 --> 0:04:25.520000
 The router may be required to support
 802.1Q VLAN tagging, we'll look

0:04:25.520000 --> 0:04:29.640000
 at that here in just one moment, and
 then lastly, hosts must be aware

0:04:29.640000 --> 0:04:31.380000
 that that router exists.

0:04:31.380000 --> 0:04:34.760000
 Once I give him an IP address and I connect
 to his interface to that VLAN,

0:04:34.760000 --> 0:04:38.720000
 we have to do something to make sure the
 hosts know that he is their default

0:04:38.720000 --> 0:04:41.960000
 gateway, and he's the one who's going
 to be routing their traffic.

0:04:41.960000 --> 0:04:44.540000
 All right, so let's take a look at what's
 this whole thing about router

0:04:44.540000 --> 0:04:46.720000
 on a stick refer to.

0:04:46.720000 --> 0:04:53.860000
 So here, I have a switch, switch one on
 my left here, who has two different

0:04:53.860000 --> 0:04:57.180000
 departments, and let's say I've decided
 to break out these two different

0:04:57.180000 --> 0:04:59.420000
 departments into two
 different VLANs.

0:04:59.420000 --> 0:05:05.820000
 So I've got some devices on VLAN
 2 and some devices on VLAN 3.

0:05:05.820000 --> 0:05:08.140000
 Let's just say that each one of these
 circles here represents maybe a

0:05:08.140000 --> 0:05:11.660000
 laptop or a PC or a server,
 something like that.

0:05:11.660000 --> 0:05:15.960000
 So we previously learned that step
 number one is to create my VLAN.

0:05:15.960000 --> 0:05:21.780000
 So the process of doing that is here in
 the switch from the global configuration

0:05:21.780000 --> 0:05:27.940000
 level just typing VLAN 2, and then
 repeating that for VLAN 3.

0:05:27.940000 --> 0:05:32.840000
 We also know from our switching technology
 lessons that if you've watched

0:05:32.840000 --> 0:05:36.060000
 any switching videos or anything,
 that that's just part one.

0:05:36.060000 --> 0:05:38.320000
 Part two is you have to go
 to the physical interface.

0:05:38.320000 --> 0:05:44.600000
 Maybe that's interface fast ethernet
 zero slash four right there.

0:05:44.600000 --> 0:05:49.040000
 And you have to tell that interface,
 you're in that particular VLAN, and

0:05:49.040000 --> 0:05:54.480000
 that's the switch port access command,
 switch port access VLAN 2, and

0:05:54.480000 --> 0:05:58.340000
 down here, I would do that under
 switch port access VLAN 3.

0:05:58.340000 --> 0:06:02.540000
 Okay, so I've created my VLANs, I've
 assigned my ports to my VLANs.

0:06:02.540000 --> 0:06:06.880000
 Now I got to get a router
 who is in those VLANs.

0:06:06.880000 --> 0:06:12.100000
 Well, one method, because I have two VLANs,
 which are two different broadcast

0:06:12.100000 --> 0:06:19.860000
 domains, I could go to this interface
 right here on my switch, and I could

0:06:19.860000 --> 0:06:25.880000
 do the same thing.

0:06:25.880000 --> 0:06:31.400000
 And switch port access VLAN 2, and then
 down here on fast ethernet zero

0:06:31.400000 --> 0:06:41.140000
 two, switch port mode access
 switch port access VLAN 3.

0:06:41.140000 --> 0:06:48.760000
 And now I could run a cable on that
 port to my interface on my router,

0:06:48.760000 --> 0:06:53.460000
 and another cable on that other port
 to another free unused interface

0:06:53.460000 --> 0:07:00.440000
 on the router. Now this top interface
 fast ethernet zero zero, it needs

0:07:00.440000 --> 0:07:03.720000
 to know that it's going to be the default
 gateway for the two two two

0:07:03.720000 --> 0:07:09.920000
 network. So I'd have to give it some
 IP address IP address two two two

0:07:09.920000 --> 0:07:14.460000
 two maybe two two two four.

0:07:14.460000 --> 0:07:19.780000
 And then do the same thing for fast
 ethernet zero one IP address three

0:07:19.780000 --> 0:07:21.580000
 three three four.

0:07:21.580000 --> 0:07:24.500000
 Because he's in a
 different network.

0:07:24.500000 --> 0:07:31.320000
 Now what I've just done here is I
 have enabled inter VLAN routing.

0:07:31.320000 --> 0:07:35.300000
 Now this is not called router on a
 stick, but this is capable of inter

0:07:35.300000 --> 0:07:38.760000
 VLAN routing. As soon as I put these IP
 addresses on these routers interfaces,

0:07:38.760000 --> 0:07:43.000000
 those will create a new
 in his routing table.

0:07:43.000000 --> 0:07:46.840000
 So let's just say this is his
 IPV four routing table.

0:07:46.840000 --> 0:07:50.960000
 Those will create what we
 call connected routes.

0:07:50.960000 --> 0:07:54.880000
 They'll say, well, I'm directly connected
 to the two two two network via

0:07:54.880000 --> 0:07:56.560000
 fast ethernet zero zero.

0:07:56.560000 --> 0:08:01.300000
 And I'm directly connected to the three
 three three three network via

0:08:01.300000 --> 0:08:03.860000
 fast ethernet zero slash one.

0:08:03.860000 --> 0:08:09.740000
 And now if I go into my DHCP server,
 which we don't see here, but if I

0:08:09.740000 --> 0:08:13.440000
 tell my DHCP server, hey, if you ever
 allocate an IP address in the two

0:08:13.440000 --> 0:08:17.560000
 two two network, make sure that you
 give that host the default gateway

0:08:17.560000 --> 0:08:19.940000
 of two two two four.

0:08:19.940000 --> 0:08:23.980000
 And for these people over here, make sure
 you give them the default gateway

0:08:23.980000 --> 0:08:26.020000
 of three three three four.

0:08:26.020000 --> 0:08:30.420000
 So we change these numbers here because
 these aren't quite correct.

0:08:30.420000 --> 0:08:37.520000
 So the people in red here, they would
 now have a default gateway of two

0:08:37.520000 --> 0:08:43.020000
 two two four. And the people in
 blue who are in VLAN three.

0:08:43.020000 --> 0:08:48.380000
 They would have a different default
 gateway, which is three three three

0:08:48.380000 --> 0:08:53.580000
 four. People in VLAN two would ARP
 for their default gateway and they

0:08:53.580000 --> 0:08:55.340000
 would very quickly learn.

0:08:55.340000 --> 0:09:01.460000
 That the MAC address of two two two four
 was zero zero zero one zero zero

0:09:01.460000 --> 0:09:04.360000
 eleven. See that's his MAC address right
 here on this physical interface.

0:09:04.360000 --> 0:09:10.720000
 So that's what they would learn when
 they send their ARP request.

0:09:10.720000 --> 0:09:14.880000
 These people down here
 in VLAN three.

0:09:14.880000 --> 0:09:19.500000
 When they sent out an ARP, for example,
 this guy right here sends an ARP

0:09:19.500000 --> 0:09:26.440000
 request for his default gateway,
 which is three three three four.

0:09:26.440000 --> 0:09:27.540000
 That's a broadcast.

0:09:27.540000 --> 0:09:29.200000
 So it would be flooded here.

0:09:29.200000 --> 0:09:31.280000
 It would be flooded here.

0:09:31.280000 --> 0:09:33.040000
 This interface would get it.

0:09:33.040000 --> 0:09:40.820000
 And that interface would
 send back an ARP reply.

0:09:40.820000 --> 0:09:45.020000
 And in his ARP reply, he would say,
 hey, I am three three three four.

0:09:45.020000 --> 0:09:48.120000
 And let me give you
 my MAC address.

0:09:48.120000 --> 0:09:55.320000
 So that's how these devices would learn
 that their default gateway of

0:09:55.320000 --> 0:09:56.780000
 three three three four.

0:09:56.780000 --> 0:10:05.360000
 Had a MAC equal to zero
 zero dot two two.

0:10:05.360000 --> 0:10:11.740000
 Now, I could continue on with this, but
 there's a downside with this approach.

0:10:11.740000 --> 0:10:16.760000
 You see, if I am expecting that every
 VLAN on my switch is going to have

0:10:16.760000 --> 0:10:21.800000
 a physical cable leading to a router's
 interface, that means my router

0:10:21.800000 --> 0:10:26.600000
 is going to have to have one physical
 interface that's available for each

0:10:26.600000 --> 0:10:29.260000
 VLAN that's configured
 on my switch.

0:10:29.260000 --> 0:10:33.540000
 Now, if I just got two VLANs like this,
 not a big deal, I can very easily

0:10:33.540000 --> 0:10:37.400000
 find routers that have two physical
 fast ethernet ports.

0:10:37.400000 --> 0:10:39.380000
 But this is not a very
 scalable design.

0:10:39.380000 --> 0:10:43.180000
 What if I had 20 or 200 VLANs?

0:10:43.180000 --> 0:10:47.520000
 You're going to be hard pressed to find
 a router with 20 or 200 physical

0:10:47.520000 --> 0:10:51.900000
 interfaces. And even if you did, you
 probably would not want to use them

0:10:51.900000 --> 0:10:55.240000
 in this way. It's just
 not a good design.

0:10:55.240000 --> 0:11:01.320000
 Wouldn't it be better if we could just
 have one cable going from the switch

0:11:01.320000 --> 0:11:06.860000
 to one physical interface on the router
 and have that one interface on

0:11:06.860000 --> 0:11:10.540000
 the router capable of routing
 between all these packets?

0:11:10.540000 --> 0:11:14.280000
 And this is what router
 on a stick is all about.

0:11:14.280000 --> 0:11:19.640000
 So with router on a stick,
 we have one physical cable.

0:11:19.640000 --> 0:11:23.240000
 Now, this means that on the router's
 interface, fast ethernet 00, we're

0:11:23.240000 --> 0:11:27.280000
 still going to have to support two
 subnets, just like in the previous

0:11:27.280000 --> 0:11:31.380000
 picture. In the previous picture, when
 we had two physical interfaces,

0:11:31.380000 --> 0:11:35.760000
 we had two subnets, the 2.2.2
 network and the 3.3.3 network.

0:11:35.760000 --> 0:11:38.980000
 And this is how the router was able to
 create two directly connected routes

0:11:38.980000 --> 0:11:40.880000
 in his routing table.

0:11:40.880000 --> 0:11:44.780000
 Now, with router on a stick, the same
 exact thing is going to be true.

0:11:44.780000 --> 0:11:49.500000
 Somehow, I have to have this one physical
 interface capable of supporting

0:11:49.500000 --> 0:11:53.160000
 two subnets. Now,
 here's my problem.

0:11:53.160000 --> 0:11:54.660000
 Here's my dilemma.

0:11:54.660000 --> 0:11:59.360000
 If I go on to my physical interface,
 if I say interface fast ethernet

0:11:59.360000 --> 0:12:05.340000
 00, and then from there, I say IP address
 and give him an IP address that's

0:12:05.340000 --> 0:12:14.460000
 appropriate for VLAN2, like
 2.2.2.3, let's say.

0:12:14.460000 --> 0:12:16.740000
 And I said, okay, now I'm going to
 give him a second IP address that's

0:12:16.740000 --> 0:12:18.640000
 appropriate for VLAN3.

0:12:18.640000 --> 0:12:25.220000
 3.3.3.3. Well, here's my problem.

0:12:25.220000 --> 0:12:30.700000
 Router interfaces, by default,
 only support one IPv4 address.

0:12:30.700000 --> 0:12:34.500000
 So the moment I hit enter on that second
 address, my first address will

0:12:34.500000 --> 0:12:37.040000
 be gone. It'll be overwritten.

0:12:37.040000 --> 0:12:41.080000
 So my first challenge is I need this
 one physical interface on the router

0:12:41.080000 --> 0:12:46.780000
 to support multiple subnets without
 overriding themselves.

0:12:46.780000 --> 0:12:51.240000
 Well, the way we accomplished that is
 not by doing any configuration on

0:12:51.240000 --> 0:12:53.200000
 the fast ethernet
 interface itself.

0:12:53.200000 --> 0:12:56.860000
 On the physical interface, all
 we need to do is no shutdown.

0:12:56.860000 --> 0:13:00.140000
 We just need to make sure it's
 electrically up and active.

0:13:00.140000 --> 0:13:03.600000
 Now, in order to get this fast ethernet
 interface to support multiple

0:13:03.600000 --> 0:13:07.360000
 subnets, now in this case 2, but in
 reality, we could be talking about

0:13:07.360000 --> 0:13:12.340000
 dozens of subnets, what I have to do
 is divide this multiple interface

0:13:12.340000 --> 0:13:16.340000
 into multiple what's
 called subinterfaces.

0:13:16.340000 --> 0:13:19.560000
 So how do we do that?

0:13:19.560000 --> 0:13:22.560000
 Well, in Cisco iOS, we
 simply type interface.

0:13:22.560000 --> 0:13:28.360000
 We type our main interface number like
 fast ethernet 00, and then we type

0:13:28.360000 --> 0:13:31.280000
 a dot followed by some number.

0:13:31.280000 --> 0:13:34.900000
 Now, the number after the dot
 could be anything you want.

0:13:34.900000 --> 0:13:39.440000
 It really doesn't make a difference,
 but it probably makes a lot of sense

0:13:39.440000 --> 0:13:43.580000
 to select a number that's meaningful
 to you as a network administrator.

0:13:43.580000 --> 0:13:47.320000
 Like in my particular case, I want
 this subinterface to be the default

0:13:47.320000 --> 0:13:48.860000
 gateway for people in VLAN 2.

0:13:48.860000 --> 0:13:52.500000
 So it would probably make a lot more
 sense for me to say fast ethernet

0:13:52.500000 --> 0:14:02.720000
 00.2. And then create another one,
 interface fast ethernet 0 slash 0.3

0:14:02.720000 --> 0:14:09.920000
 for VLAN 3. Now, just by putting a dot
 and a number after your fast ethernet,

0:14:09.920000 --> 0:14:15.240000
 that is creating a new logical interface
 called a sub interface.

0:14:15.240000 --> 0:14:17.720000
 It's a sub component.

0:14:17.720000 --> 0:14:19.160000
 It's associated with.

0:14:19.160000 --> 0:14:21.860000
 It's tied to the physical
 interface.

0:14:21.860000 --> 0:14:26.980000
 But Cisco iOS, the operating system now
 sort of views these as two completely

0:14:26.980000 --> 0:14:29.100000
 independent interfaces.

0:14:29.100000 --> 0:14:33.520000
 And because they are independent interfaces,
 I can put IP addresses on

0:14:33.520000 --> 0:14:39.460000
 them that are in different
 subnets like this.

0:14:39.460000 --> 0:14:44.200000
 I can put an IP address on there in
 fast ethernet 00 slash dot 2 that's

0:14:44.200000 --> 0:14:45.820000
 appropriate for VLAN 2.

0:14:45.820000 --> 0:14:51.860000
 And I can put another IP address on
 this one that's appropriate for the

0:14:51.860000 --> 0:14:55.700000
 people in VLAN 3 for their
 default gateway.

0:14:55.700000 --> 0:15:00.000000
 Now, you might be wondering, Keith, why
 did you leave just such huge spaces

0:15:00.000000 --> 0:15:02.460000
 here? There's actually a reason for
 this, and I'll get to that in just

0:15:02.460000 --> 0:15:07.460000
 one moment. But the main takeaway I
 want you to have right now is that

0:15:07.460000 --> 0:15:14.660000
 a physical fast ethernet interface by
 default only supports one subnet,

0:15:14.660000 --> 0:15:16.120000
 one subnet only.

0:15:16.120000 --> 0:15:21.580000
 So if I want to have this physical fast
 ethernet supporting multiple IPV

0:15:21.580000 --> 0:15:26.760000
 for subnets, then I break it out into
 what's called sub interfaces.

0:15:26.760000 --> 0:15:32.460000
 If I had 300 VLANs on here, I would
 have to create 300 sub interfaces.

0:15:32.460000 --> 0:15:34.820000
 Yes, it would take a
 little bit of time.

0:15:34.820000 --> 0:15:39.220000
 But the benefit is each sub interface
 now is capable of controlling one

0:15:39.220000 --> 0:15:43.360000
 unique subnet. And I can do that.

0:15:43.360000 --> 0:15:47.800000
 So that takes care of my first problem,
 which is I need one physical interface

0:15:47.800000 --> 0:15:50.480000
 carrying multiple subnets.

0:15:50.480000 --> 0:15:52.580000
 Now, here's my second problem.

0:15:52.580000 --> 0:15:55.460000
 Let's take a look at the switch
 interface for a second.

0:15:55.460000 --> 0:15:59.520000
 By default, this switch interface
 is going to be an access port.

0:15:59.520000 --> 0:16:02.820000
 And it's going to be an
 access port in VLAN 1.

0:16:02.820000 --> 0:16:06.560000
 I need to go on to the switch and tell
 the switch, hey, this interface

0:16:06.560000 --> 0:16:11.980000
 actually needs to be associated
 with both VLAN 2 and VLAN 3.

0:16:11.980000 --> 0:16:15.340000
 And any other VLANs I
 create in the future.

0:16:15.340000 --> 0:16:18.320000
 In other words, if you remember your
 days of switching, that means I need

0:16:18.320000 --> 0:16:20.960000
 to make this a VLAN trunk.

0:16:20.960000 --> 0:16:25.040000
 So the command for that on the switching
 side would simply be switch port

0:16:25.040000 --> 0:16:34.560000
 mode trunk. Okay, so this is now I'm going
 to put some colored boxes around

0:16:34.560000 --> 0:16:39.180000
 this. So that means that one physical
 interface now is capable of carrying

0:16:39.180000 --> 0:16:42.360000
 traffic on the red VLAN 2.

0:16:42.360000 --> 0:16:47.640000
 It's also capable of carrying traffic
 on the blue VLAN, which is VLAN

0:16:47.640000 --> 0:16:52.960000
 3. That's what a VLAN trunk is.

0:16:52.960000 --> 0:16:56.200000
 Okay, so let's say that
 I've just done this.

0:16:56.200000 --> 0:16:58.840000
 Now, this is not enough yet, but let's
 say that this is where I've stopped

0:16:58.840000 --> 0:17:03.160000
 so far. Here's a problem
 I'm going to encounter.

0:17:03.160000 --> 0:17:10.940000
 Once, let's say 2-2-1, once he does
 DHCP in the morning and he discovers

0:17:10.940000 --> 0:17:15.880000
 that his default gateway is 2-2-3,
 very quickly he's going to arpe for

0:17:15.880000 --> 0:17:17.060000
 his default gateway.

0:17:17.060000 --> 0:17:21.080000
 Because he wants to find out what his
 default gateway's MAC address is.

0:17:21.080000 --> 0:17:28.260000
 So he's going to send
 an arpe for 2-2-3.

0:17:28.260000 --> 0:17:31.960000
 That arpe is a broadcast, so it's going
 to be flooded by the switch out

0:17:31.960000 --> 0:17:35.660000
 all other ports carrying VLAN 2.

0:17:35.660000 --> 0:17:38.660000
 So it'll go out to that other
 PC, which is 2-2-2-2.

0:17:38.660000 --> 0:17:40.460000
 He doesn't really care
 about the arpe.

0:17:40.460000 --> 0:17:42.980000
 It'll also go out this trunk.

0:17:42.980000 --> 0:17:47.000000
 Right here. Now, here's the thing.


0:17:47.000000 --> 0:17:53.300000
 Your trunk is most likely going to
 be implementing trunking using the

0:17:53.300000 --> 0:17:56.180000
 IEEE protocol of 802.1Q.

0:17:56.180000 --> 0:18:04.920000
 And if you remember from your switching
 days, the way 802.1Q works is

0:18:04.920000 --> 0:18:09.040000
 when a frame goes across a
 VLAN trunk doing 802.1Q.

0:18:09.040000 --> 0:18:15.740000
 If that frame does not belong to the
 native VLAN, a VLAN tag gets applied

0:18:15.740000 --> 0:18:19.800000
 to it. Well, the native VLAN is VLAN
 1, and let's just assume I haven't

0:18:19.800000 --> 0:18:21.560000
 done anything to that.

0:18:21.560000 --> 0:18:23.300000
 I haven't changed that.

0:18:23.300000 --> 0:18:29.820000
 So if this is my arpe request right
 here, 802.1Q is going to put a VLAN

0:18:29.820000 --> 0:18:32.680000
 tag into my Ethernet header.

0:18:32.680000 --> 0:18:37.260000
 And it's going to tag that frame saying
 this frame belongs to VLAN 2.

0:18:37.260000 --> 0:18:41.520000
 So let's say this right
 here is my .1Q tag.

0:18:41.520000 --> 0:18:47.640000
 Well, here's my second problem.

0:18:47.640000 --> 0:18:52.360000
 Router interfaces don't natively
 understand tags.

0:18:52.360000 --> 0:18:56.440000
 When this Ethernet frame comes into
 this interface, and he looks at the

0:18:56.440000 --> 0:18:59.740000
 Ethernet type code and the Ethernet
 type code says, hey, there's a .1Q

0:18:59.740000 --> 0:19:03.660000
 tag in here. The router's going to
 say, I have no idea what that is.

0:19:03.660000 --> 0:19:05.860000
 I understand ARP.

0:19:05.860000 --> 0:19:07.680000
 I understand IPv4.

0:19:07.680000 --> 0:19:10.980000
 I don't understand this
 thing called tags.

0:19:10.980000 --> 0:19:14.040000
 And so the router will
 delete this frame.

0:19:14.040000 --> 0:19:17.760000
 So this is where we get into the second
 requirement where the router not

0:19:17.760000 --> 0:19:21.980000
 only needs to have separate sub-interfaces,
 separate logical interfaces

0:19:21.980000 --> 0:19:26.420000
 representing different VLANs
 or different subnets.

0:19:26.420000 --> 0:19:30.500000
 The router's sub-interface also needs
 to be capable of understanding and

0:19:30.500000 --> 0:19:33.700000
 processing these .1Q tags.

0:19:33.700000 --> 0:19:38.780000
 So for that, we need to add some
 additional commands in here.

0:19:38.780000 --> 0:19:42.820000
 So under the router's sub-interfaces,
 we have to use the encapsulation

0:19:42.820000 --> 0:19:48.940000
 command. Encapsulation.1Q, in
 this case it's going to be 2.

0:19:48.940000 --> 0:19:54.520000
 Meaning, okay, anytime an Ethernet
 frame comes in with a .1Q tag, and

0:19:54.520000 --> 0:19:59.100000
 that .1Q tag says this frame belongs
 to VLAN2, it'll be redirected to

0:19:59.100000 --> 0:20:00.540000
 this sub-interface.

0:20:00.540000 --> 0:20:03.060000
 This sub-interface
 will process it.

0:20:03.060000 --> 0:20:07.540000
 Similarly, we'll do the same thing
 on the other sub-interface.

0:20:07.540000 --> 0:20:12.900000
 Encapsulation.1Q.3.

0:20:12.900000 --> 0:20:17.960000
 Now in the actual configuration of things,
 you have to do the encapsulation

0:20:17.960000 --> 0:20:22.160000
 command before you do your
 IP addressing command.

0:20:22.160000 --> 0:20:26.140000
 If I actually went into a router and
 I created a sub-interface, and then

0:20:26.140000 --> 0:20:29.900000
 the very next thing I did in that sub
-interface was to add an IP address,

0:20:29.900000 --> 0:20:31.760000
 I would actually get
 an error message.

0:20:31.760000 --> 0:20:34.720000
 I would get some sort of error message
 saying, you know, IP addressing

0:20:34.720000 --> 0:20:37.980000
 not supported without .1Q tagging
 or something like that.

0:20:37.980000 --> 0:20:41.000000
 In the next video, I'll show you exactly
 what that error message looks

0:20:41.000000 --> 0:20:44.920000
 like. So the order of processing is
 on the physical interface, just make

0:20:44.920000 --> 0:20:47.320000
 sure it's no shut.

0:20:47.320000 --> 0:20:51.700000
 Then start creating sub-interfaces,
 one sub-interface for each VLAN that

0:20:51.700000 --> 0:20:54.780000
 this router is going to need
 to be able to route for.

0:20:54.780000 --> 0:20:59.920000
 After you create your sub-interface,
 use the encapsulation .1Q command,

0:20:59.920000 --> 0:21:03.480000
 and the number after .1Q
 is your VLAN number.

0:21:03.480000 --> 0:21:06.820000
 This is the number that the router
 is going to see in incoming frames

0:21:06.820000 --> 0:21:09.380000
 that have 802.1Q tags.

0:21:09.380000 --> 0:21:14.020000
 Then put an IP address under that sub
-interface that is an unused host

0:21:14.020000 --> 0:21:19.200000
 address that is in the same sub-net
 as all the devices that are going

0:21:19.200000 --> 0:21:20.640000
 to be in VLAN 2.

0:21:20.640000 --> 0:21:27.080000
 Now this router will be
 capable of routing.

0:21:27.080000 --> 0:21:31.480000
 Because now when the ARP request comes
 in, the router will say, oh, here

0:21:31.480000 --> 0:21:34.180000
 I see an ARP request
 with a .1Q tag.

0:21:34.180000 --> 0:21:39.180000
 That ARP request will be forwarded
 to this sub-interface.

0:21:39.180000 --> 0:21:44.500000
 He will then send out
 an ARP response.

0:21:44.500000 --> 0:21:49.300000
 When he sends it out, he will tag it
 on the outbound direction as well.

0:21:49.300000 --> 0:21:54.440000
 That is how all these devices here
 will learn what the MAC address is

0:21:54.440000 --> 0:21:57.400000
 of their default gateway.

0:21:57.400000 --> 0:21:59.400000
 Now how is the routing
 actually happening?

0:21:59.400000 --> 0:22:03.200000
 Remember, in this router,
 he has a routing table.

0:22:03.200000 --> 0:22:10.640000
 Anytime that you add an IP address
 to a router's interface, whether it

0:22:10.640000 --> 0:22:13.720000
 be a physical interface or whether,
 like in this case, it be a logical

0:22:13.720000 --> 0:22:18.400000
 interface, that creates a route
 in the routing table.

0:22:18.400000 --> 0:22:24.600000
 So now he will say, I am directly connected
 to the .222 network via sub

0:22:24.600000 --> 0:22:27.220000
-interface fast-ethanet00.2.

0:22:27.220000 --> 0:22:43.820000
 I am also directly connected to the
 .333 network via fast-ethanet00.3,

0:22:43.820000 --> 0:22:48.180000
 that sub-interface.

0:22:48.180000 --> 0:22:50.340000
 So how are the frames
 going to flow?

0:22:50.340000 --> 0:22:56.020000
 Well, if .2221 right here is trying
 to send a packet to a destination

0:22:56.020000 --> 0:23:06.600000
 of, let's say, .3331, .221 is very
 quickly going to realize that, oh,

0:23:06.600000 --> 0:23:08.000000
 that is not on my network.

0:23:08.000000 --> 0:23:11.820000
 My network has the first
 24 bits of .222.

0:23:11.820000 --> 0:23:14.260000
 This is a different pattern.

0:23:14.260000 --> 0:23:19.240000
 Therefore, when I create my Ethernet
 header, so I will put that in the

0:23:19.240000 --> 0:23:24.860000
 packet, I will create Ethernet header,
 I use my own MAC address as the

0:23:24.860000 --> 0:23:31.680000
 source MAC address, which is .22.21,
 and for the destination MAC address,

0:23:31.680000 --> 0:23:43.260000
 I will send it to my router,
 which is .00.11.

0:23:43.260000 --> 0:23:47.860000
 When that frame gets to the switch,
 the switch will know exactly what

0:23:47.860000 --> 0:23:53.640000
 interface, in this case, fast-ethanet01,
 it will know that .0011 has been

0:23:53.640000 --> 0:23:56.040000
 learned on fast-ethanet01.

0:23:56.040000 --> 0:23:57.400000
 Now, how did it know that?

0:23:57.400000 --> 0:24:02.860000
 It knew that because when the router
 responded to the ARP request, when

0:24:02.860000 --> 0:24:07.940000
 the router sent an ARP reply, that
 ARP reply had a source MAC address

0:24:07.940000 --> 0:24:12.200000
 of whatever is on this physical
 interface right here.

0:24:12.200000 --> 0:24:16.460000
 .0011. That's what allowed the switch
 to learn in its MAC address table,

0:24:16.460000 --> 0:24:23.600000
 that .00.11 is associated to or learned
 on this trunk, fast-ethanet01.

0:24:23.600000 --> 0:24:29.560000
 So let's just put that right here.


0:24:29.560000 --> 0:24:39.820000
 So the switch, in his MAC address-table,
 learned that .00.11 is on VLAN2

0:24:39.820000 --> 0:24:43.140000
 on fast-ethanet01.

0:24:43.140000 --> 0:24:47.720000
 Similarly, when the router responded
 to ARP request on behalf of host

0:24:47.720000 --> 0:24:52.280000
 in VLAN3, he said, hey, people, you're
 looking for me, 3333, let me tell

0:24:52.280000 --> 0:24:55.180000
 you about myself, I'm .00.11.

0:24:55.180000 --> 0:24:59.260000
 The switch learned that that same MAC
 address was also learned on VLAN3

0:24:59.260000 --> 0:25:09.420000
 on fast-ethanet01.11
 or fast-ethanet01.1.

0:25:09.420000 --> 0:25:19.120000
 So now, when this frame here, from
 .222.1, going to .333.1, now that's

0:25:19.120000 --> 0:25:24.780000
 layer 3. At layer 2, the destination
 MAC address is default gateway, which

0:25:24.780000 --> 0:25:31.480000
 is .00.11. That gets
 to the switch.

0:25:31.480000 --> 0:25:34.940000
 The switch is an incoming
 frame on VLAN2.

0:25:34.940000 --> 0:25:38.660000
 Remember, it came in on a port in VLAN2,
 going to a destination MAC address

0:25:38.660000 --> 0:25:44.460000
 of .0011. Sure enough, in his MAC address
 table, he has that destination.

0:25:44.460000 --> 0:25:48.840000
 He knows that .00.11 on VLAN2 is
 reachable via fast-ethanet01.

0:25:48.840000 --> 0:25:53.080000
 So that frame will be switched
 out, fast-ethanet01.

0:25:53.080000 --> 0:25:57.380000
 It'll hit the router on his physical
 interface, fast-ethanet00, going

0:25:57.380000 --> 0:26:02.500000
 to his MAC. But when he sees the .1Q
 tag in here, and I didn't draw that,

0:26:02.500000 --> 0:26:07.060000
 but there's going to be a .1Q tag saying,
 this frame belongs to VLAN2,

0:26:07.060000 --> 0:26:12.160000
 the router will route that frame to
 his own internal sub-interface, .00

0:26:12.160000 --> 0:26:15.580000
.2, because of this
 command right here.

0:26:15.580000 --> 0:26:18.780000
 Initially, he'll think, oh, this frame
 was for me, because after all,

0:26:18.780000 --> 0:26:22.120000
 it was destined to my
 MAC address, .00.11.

0:26:22.120000 --> 0:26:29.500000
 But when he strips off the layer 2 header,
 when he gets rid of this stuff,

0:26:29.500000 --> 0:26:35.160000
 and sees the destination IP address inside,
 he'll realize, oh, this actually

0:26:35.160000 --> 0:26:36.380000
 isn't meant for me.

0:26:36.380000 --> 0:26:38.600000
 This is meant for .333.1.

0:26:38.600000 --> 0:26:41.040000
 Do I have any idea
 where that lives?

0:26:41.040000 --> 0:26:42.960000
 Oh, yeah, right there.

0:26:42.960000 --> 0:26:45.340000
 I know exactly where the
 .33 network lives.

0:26:45.340000 --> 0:26:48.720000
 It's on my sub-interface, .00.3.

0:26:48.720000 --> 0:26:53.800000
 So that frame will be routed from sub
-interface .00.2, where it came in

0:26:53.800000 --> 0:26:58.480000
 on to .00.3, where it needs to go.


0:26:58.480000 --> 0:27:02.540000
 And so now he'll create a new layer 2
 header, so as the frame gets routed

0:27:02.540000 --> 0:27:05.480000
 out, so now it's going to go right back
 out the same physical interface

0:27:05.480000 --> 0:27:06.240000
 where it came in.

0:27:06.240000 --> 0:27:08.720000
 It came in fast-eathed at .00.

0:27:08.720000 --> 0:27:13.300000
 It was redirected on
 the ingress to .00.2.

0:27:13.300000 --> 0:27:18.160000
 Now it was routed to .00.3.

0:27:18.160000 --> 0:27:25.340000
 So now it gets a new layer
 2 header applied to it.

0:27:25.340000 --> 0:27:32.940000
 And as it goes out here, the new destination
 MAC address is .333.1.

0:27:32.940000 --> 0:27:44.480000
 So whatever his MAC address is,
 let's just say it's XX.YY.

0:27:44.480000 --> 0:27:49.340000
 So our router here will most likely,
 well, he will already know about

0:27:49.340000 --> 0:27:51.000000
 XX.YY.Y. Why is that?

0:27:51.000000 --> 0:27:54.140000
 Well, because .333
.1 arped for him.

0:27:54.140000 --> 0:27:57.720000
 Way back when .333.1 got his IP address
 and learned that his default gateway

0:27:57.720000 --> 0:28:14.700000
 was .333. He arped for .333.

0:28:14.700000 --> 0:28:21.240000
 So the router knows that .333.1 is
 associated to the MAC address of XX

0:28:21.240000 --> 0:28:27.260000
.YY. So when this packet comes in that
 needs to be routed out, the new

0:28:27.260000 --> 0:28:33.780000
 destination MAC address as it leaves
 the router will be XX.YY.

0:28:33.780000 --> 0:28:39.900000
 And the source MAC address, well, will
 be the router's MAC address .00

0:28:39.900000 --> 0:28:51.060000
.11. And this is how router
 on a stick works.

0:28:51.060000 --> 0:28:54.080000
 They call it router on a stick because
 it's a single router connected

0:28:54.080000 --> 0:29:00.240000
 to one stick, one link, one cable
 connected to our switch.

0:29:00.240000 --> 0:29:05.220000
 But this one router, his one physical
 interface is supporting multiple

0:29:05.220000 --> 0:29:12.540000
 subnets by the use of creating
 sub interfaces.

0:29:12.540000 --> 0:29:16.580000
 And so now frames can come in
 on that physical interface.

0:29:16.580000 --> 0:29:22.780000
 They're redirected to the sub interface
 that's designed to handle them.

0:29:22.780000 --> 0:29:26.180000
 Once that sub interface strips off the
 layer two ethernet header and looks

0:29:26.180000 --> 0:29:29.860000
 at the destination IP address, it realizes
 that that frame needs to be

0:29:29.860000 --> 0:29:33.480000
 rerouted to a different
 sub interface.

0:29:33.480000 --> 0:29:39.080000
 That different sub interface that gets
 it, then creates a new layer two

0:29:39.080000 --> 0:29:45.460000
 header with its own MAC address as the
 source, the destination MAC address

0:29:45.460000 --> 0:29:49.800000
 as the destination
 and sends it out.

0:29:49.800000 --> 0:29:54.340000
 So that is one way of doing intervLAN
 routing using an external router

0:29:54.340000 --> 0:29:56.600000
 connected to your switch.

0:29:56.600000 --> 0:29:59.100000
 And this method is called
 router on a stick.

