WEBVTT

0:00:03.240000 --> 0:00:09.100000
 Hello and welcome to this video, which
 is part of our VPN overview series

0:00:09.100000 --> 0:00:11.400000
 of the CCNA bootcamp.

0:00:11.400000 --> 0:00:15.580000
 And this video is called MPLS terminology
 and concepts, who do just a

0:00:15.580000 --> 0:00:21.040000
 brief review and overview of how MPLS
 works and some of the terms of MPLS

0:00:21.040000 --> 0:00:22.480000
 you should know.

0:00:22.480000 --> 0:00:25.720000
 So first of all, what
 does MPLS stand for?

0:00:25.720000 --> 0:00:29.780000
 Multi-protocol label switching.

0:00:29.780000 --> 0:00:35.440000
 The idea here is that when data comes
 into an MPLS enabled router, that

0:00:35.440000 --> 0:00:39.340000
 MPLS router, after the layer two header,
 now that layer two header is

0:00:39.340000 --> 0:00:44.020000
 probably an ethernet header, although
 MPLS was designed to work with other

0:00:44.020000 --> 0:00:48.260000
 things besides ethernet, but the idea
 is after the ethernet header, it

0:00:48.260000 --> 0:00:52.500000
 will sort of move the layer three header
 over and after the ethernet header,

0:00:52.500000 --> 0:00:56.080000
 it will put in this thing called an
 MPLS label and we'll look at what

0:00:56.080000 --> 0:00:58.940000
 that label structure looks
 like in just a second.

0:00:58.940000 --> 0:01:04.020000
 And then subsequent MPLS routers, we'll
 look at that label to figure out

0:01:04.020000 --> 0:01:06.980000
 which interface this thing
 needs to be forwarded on.

0:01:06.980000 --> 0:01:10.960000
 So instead of forwarding data or switching
 data based on your layer three

0:01:10.960000 --> 0:01:15.560000
 header like your IPv4 or your IPv6 address,
 we can now forward it based

0:01:15.560000 --> 0:01:19.900000
 on this label. And because we're using
 this label, it freezes up to have

0:01:19.900000 --> 0:01:21.600000
 anything after the label.

0:01:21.600000 --> 0:01:27.580000
 So MPLS is typically used for IPv4,
 IPv6 packets, but it can carry all

0:01:27.580000 --> 0:01:28.420000
 sorts of other things.

0:01:28.420000 --> 0:01:32.160000
 For example, there's types of MPLS where
 you could take an ethernet frame

0:01:32.160000 --> 0:01:39.940000
 and you could have MPLS, that ethernet
 frame that does not have any IPv4,

0:01:39.940000 --> 0:01:46.300000
 IPv6 in it, and we could use MPLS to transport
 that frame across the cloud.

0:01:46.300000 --> 0:01:50.340000
 So that's why we call it multi
-protocol label switching.

0:01:50.340000 --> 0:01:53.340000
 Anything that could have a label put
 in front of it, any protocol, we

0:01:53.340000 --> 0:01:56.440000
 could switch it with
 these MPLS labels.

0:01:56.440000 --> 0:02:00.060000
 The labels like it says here are dynamically
 generated and propagated

0:02:00.060000 --> 0:02:01.700000
 throughout an MPLS network.

0:02:01.700000 --> 0:02:05.320000
 So you don't have to configure the label,
 you don't have to say what the

0:02:05.320000 --> 0:02:11.100000
 label number is, MPLS devices will
 figure that out for themselves.

0:02:11.100000 --> 0:02:14.480000
 Now MPLS is kind of weird here because
 it says this is considered a layer

0:02:14.480000 --> 0:02:17.040000
 2.5 technology. Why?

0:02:17.040000 --> 0:02:18.840000
 Because of where that
 label is put.

0:02:18.840000 --> 0:02:24.100000
 It's put after the layer 2 header,
 but before the layer 3 header.

0:02:24.100000 --> 0:02:27.440000
 Hence why we call it a
 layer 2.5 technology.

0:02:27.440000 --> 0:02:32.240000
 Now in Cisco routers, MPLS leverages
 both IP routing and Cisco Express

0:02:32.240000 --> 0:02:34.020000
 forwarding switching.

0:02:34.020000 --> 0:02:37.840000
 So you actually have to have Cisco
 Express forwarding turned on, which

0:02:37.840000 --> 0:02:41.120000
 is on by default and pretty much all
 Cisco routers and switches these

0:02:41.120000 --> 0:02:46.320000
 days in order to enable
 this feature.

0:02:46.320000 --> 0:02:51.100000
 And typically, in just a basic MPLS
 environment, MPLS labels correspond

0:02:51.100000 --> 0:02:53.420000
 to destination networks.

0:02:53.420000 --> 0:02:54.520000
 What does that mean?

0:02:54.520000 --> 0:02:59.020000
 That means that I as a router, as I
 start learning about networks via

0:02:59.020000 --> 0:03:04.340000
 RIP or EIGRP or OSPF, for every network
 I learn about, I will come up

0:03:04.340000 --> 0:03:09.620000
 with an that I can use for
 forwarding of that network.

0:03:09.620000 --> 0:03:12.120000
 And we'll see a little bit
 about how that works.

0:03:12.120000 --> 0:03:17.060000
 All right, so some MPLS
 terminology here.

0:03:17.060000 --> 0:03:24.180000
 So any router that deals with an MPLS
 label, whether it be adding a label

0:03:24.180000 --> 0:03:28.620000
 to a packet that currently does not
 have one, stripping a label off of

0:03:28.620000 --> 0:03:33.380000
 a packet and then sending it unlabeled
 or receiving something that's already

0:03:33.380000 --> 0:03:37.560000
 labeled and then switching it out and
 maintaining a label, anything that

0:03:37.560000 --> 0:03:41.540000
 deals with a label is called a
 label switch router or an LSR.

0:03:41.540000 --> 0:03:46.940000
 Now, depending on where that router
 is in the network, it can be called

0:03:46.940000 --> 0:03:53.180000
 a ingress LSR, which is also called a
 label edge router or an intermediate

0:03:53.180000 --> 0:03:56.160000
 LSR or an egress LSR.

0:03:56.160000 --> 0:04:02.800000
 Take a look at what that's
 talking about right there.

0:04:02.800000 --> 0:04:13.060000
 Okay, so here I have some network, composed
 of routers and switches and

0:04:13.060000 --> 0:04:20.060000
 such. Now MPLS, although it can be used
 within your own company's network,

0:04:20.060000 --> 0:04:22.280000
 that's not typically where
 you're going to find it.

0:04:22.280000 --> 0:04:28.280000
 MPLS is usually a service that MPLS service
 providers like ISPs will offer

0:04:28.280000 --> 0:04:32.500000
 you to connect one remote office to
 a central office or remote offices

0:04:32.500000 --> 0:04:37.620000
 together. So imagine this cloud
 here is one remote office.

0:04:37.620000 --> 0:04:39.680000
 All right, and here's
 the edge router.

0:04:39.680000 --> 0:04:43.240000
 So this is the customer
 edge router.

0:04:43.240000 --> 0:04:49.660000
 Here's another network of that same customer
 with their own customer edge

0:04:49.660000 --> 0:04:55.080000
 router. Okay. Okay.

0:04:55.080000 --> 0:05:02.500000
 And here is the first MPLS router
 that's talking to it.

0:05:02.500000 --> 0:05:11.660000
 So as packets come in this way, they
 are regular old IPv4 or IPv6 packets.

0:05:11.660000 --> 0:05:17.320000
 This router here is
 a label edge router.

0:05:17.320000 --> 0:05:21.400000
 Also, sometimes like the slide said,
 sometimes called an ingress LSR.

0:05:21.400000 --> 0:05:25.620000
 So what he's going to do is he's going
 to take that packet and apply a

0:05:25.620000 --> 0:05:32.160000
 label to it, apply an
 MPLS label to it.

0:05:32.160000 --> 0:05:39.040000
 And then that now that labeled thing
 is going to go through here.

0:05:39.040000 --> 0:05:50.800000
 So these devices in here, these are called
 intermediate intermediate LSRs.

0:05:50.800000 --> 0:05:55.160000
 A lot of times they're
 called core LSRs.

0:05:55.160000 --> 0:06:06.700000
 Because these guys have no visibility
 to the actual IPv4 or IPv6 header

0:06:06.700000 --> 0:06:13.060000
 at all. What they are doing is they
 are making their switching decisions

0:06:13.060000 --> 0:06:17.340000
 purely based on the MPLS
 labels that they see.

0:06:17.340000 --> 0:06:26.240000
 And then as it goes this way, here is
 the last router who's going to receive

0:06:26.240000 --> 0:06:30.620000
 the packet that is pre-labeled,
 that has a label on it.

0:06:30.620000 --> 0:06:36.740000
 He's going to pop that label off and send
 it as a regular unlabeled packet.

0:06:36.740000 --> 0:06:40.720000
 He is also a label edge router
 because he's on the edge.

0:06:40.720000 --> 0:06:42.700000
 But we don't call him
 an ingress, right?

0:06:42.700000 --> 0:06:44.620000
 The ingress was over here.

0:06:44.620000 --> 0:06:53.600000
 We call him an egress label edge router
 or an ingress label switching

0:06:53.600000 --> 0:07:05.720000
 router. Okay. So this, based on the
 label, and let's just put some other

0:07:05.720000 --> 0:07:07.620000
 routers in here.

0:07:07.620000 --> 0:07:20.360000
 Okay, there might be some other
 core routers over here.

0:07:20.360000 --> 0:07:27.380000
 But as this packet came in, as it was
 labeled, notice it took this path

0:07:27.380000 --> 0:07:32.060000
 from here to here, from here to here,
 from here to here, and here to here.

0:07:32.060000 --> 0:07:39.520000
 Just like routers have, which are running
 SIF, let's just do a quick review

0:07:39.520000 --> 0:07:43.860000
 of Cisco Express forwarding.

0:07:43.860000 --> 0:07:47.080000
 So in Cisco you start with
 your routing table.

0:07:47.080000 --> 0:07:53.120000
 You're all familiar with what that
 is, which contains your routes, and

0:07:53.120000 --> 0:07:55.560000
 probably an ARP table.

0:07:55.560000 --> 0:08:00.400000
 Right? Your ARP table will contain
 the MAC address of the next top.

0:08:00.400000 --> 0:08:03.100000
 So you can create an Ethernet header
 and send an Ethernet frame to the

0:08:03.100000 --> 0:08:10.200000
 next top. Well, when you enable SIF,
 the basic components of this are

0:08:10.200000 --> 0:08:13.860000
 downloaded into two different
 SIF tables.

0:08:13.860000 --> 0:08:19.500000
 We have the SIF FIB table, the
 forwarding information base.

0:08:19.500000 --> 0:08:22.960000
 So that's going to contain just the
 basic stuff from the routing table,

0:08:22.960000 --> 0:08:27.020000
 like the routes and the next
 top, and that's basically it.

0:08:27.020000 --> 0:08:29.680000
 Because there's stuff in the routing
 table that we don't really need to

0:08:29.680000 --> 0:08:30.980000
 actually forward a packet.

0:08:30.980000 --> 0:08:34.640000
 For example, the routing table for every
 single route shows us how long

0:08:34.640000 --> 0:08:36.320000
 that route has been learned.

0:08:36.320000 --> 0:08:39.920000
 Well, we don't care about how long
 it's been in there to actually use

0:08:39.920000 --> 0:08:41.280000
 it and forward a packet.

0:08:41.280000 --> 0:08:43.620000
 A route will also tell
 us how it was learned.

0:08:43.620000 --> 0:08:46.020000
 Was it learned via OSPF, EIGRP?

0:08:46.020000 --> 0:08:48.260000
 It'll be a little D or
 an O or an R there.

0:08:48.260000 --> 0:08:52.060000
 Once again, do we really need that
 information to forward a packet?

0:08:52.060000 --> 0:08:56.560000
 No, we don't. We just need the route,
 the subnet mask, and the next top.

0:08:56.560000 --> 0:08:59.700000
 That's what's going to be downloaded
 right here to the SIF FIB table.

0:08:59.700000 --> 0:09:03.540000
 And then the ARP table, the contents
 of that will be downloaded to an

0:09:03.540000 --> 0:09:07.440000
 adjacency table.

0:09:07.440000 --> 0:09:12.440000
 So these are the two structures that
 Cisco express forwarding utilize.

0:09:12.440000 --> 0:09:16.400000
 It builds and maintains and switches
 packets based on the contents of

0:09:16.400000 --> 0:09:19.740000
 the SIF FIB and the
 adjacency table.

0:09:19.740000 --> 0:09:26.260000
 Well, these routers here that are running
 MPLS, they have something called

0:09:26.260000 --> 0:09:34.400000
 an L-FIB, which is
 also built by SIF.

0:09:34.400000 --> 0:09:38.420000
 This can guess it contains the label
 values and the egress interface you

0:09:38.420000 --> 0:09:42.200000
 use if you receive in something
 that's got that label on it.

0:09:42.200000 --> 0:09:45.200000
 Oh, I just got in something that's
 got a label of 15 on it.

0:09:45.200000 --> 0:09:49.060000
 That tells me I need to go out
 that interface right there.

0:09:49.060000 --> 0:09:53.480000
 So, and then they also reference
 an adjacency table as well.

0:09:53.480000 --> 0:10:04.220000
 Because after all, we need to know what
 the MAC address is of the label

0:10:04.220000 --> 0:10:07.320000
 edge router here was sending
 it along this way.

0:10:07.320000 --> 0:10:11.900000
 This entire path that was determined
 because of the L-FIB, there's a name

0:10:11.900000 --> 0:10:18.680000
 for this. This path was determined not
 because of IPv4, IPv6 information.

0:10:18.680000 --> 0:10:22.300000
 This path was determined because
 of label information.

0:10:22.300000 --> 0:10:35.420000
 And so we actually
 call this an L-FIB.

0:10:35.420000 --> 0:10:39.220000
 All right, don't have to worry too much
 about direction upstream and downstream.

0:10:39.220000 --> 0:10:43.840000
 All right, so let's dig into it real quickly
 here what these labels actually

0:10:43.840000 --> 0:10:45.680000
 look like we've been
 talking about them.

0:10:45.680000 --> 0:10:49.660000
 What are they? Well, for one thing,
 they're four bytes long.

0:10:49.660000 --> 0:10:53.920000
 So, easy to remember, right?

0:10:53.920000 --> 0:10:57.920000
 IP address, four bytes long,
 MPLS label, four bytes long.

0:10:57.920000 --> 0:11:00.500000
 So hopefully that'll
 make it easy for you.

0:11:00.500000 --> 0:11:04.120000
 They define the destination
 and services for a packet.

0:11:04.120000 --> 0:11:07.360000
 They identify a forwarding
 equivalence class.

0:11:07.360000 --> 0:11:12.420000
 So, what the heck is a forwarding
 equivalence class?

0:11:12.420000 --> 0:11:25.280000
 Okay, let's say that behind me
 is the 50.50 IPv4 network.

0:11:25.280000 --> 0:11:29.140000
 Okay, and let's say that you and I are
 routers and we're doing some routing

0:11:29.140000 --> 0:11:33.860000
 protocol. Take your pick, OSPF,
 EIGRP, doesn't really matter.

0:11:33.860000 --> 0:11:37.680000
 But we use, I use the routing protocol
 and I advertise to you the 50.50

0:11:37.680000 --> 0:11:42.220000
 network. So, you know, to get to the
 50.50 network, you have to go through

0:11:42.220000 --> 0:11:47.440000
 me. Now, if we're doing MPLS, this
 is what we're going to do.

0:11:47.440000 --> 0:11:52.420000
 As soon as I learn the 50.50 network,
 after I advertise it to you, I'm

0:11:52.420000 --> 0:11:58.120000
 going to derive a label for this, a four
 byte label, part of the component,

0:11:58.120000 --> 0:12:01.080000
 the main part of the component is going
 to be a label number, a label

0:12:01.080000 --> 0:12:05.700000
 value. So, let's say I decide, and there's
 really no way to predict what

0:12:05.700000 --> 0:12:07.980000
 that's going to be in advance.

0:12:07.980000 --> 0:12:09.880000
 But let's say, let's
 just make it simple.

0:12:09.880000 --> 0:12:13.220000
 Okay, let's say I come up
 with a label number five.

0:12:13.220000 --> 0:12:19.160000
 All right, so I say to myself, okay,
 so in my L-Fib, I say label five

0:12:19.160000 --> 0:12:23.660000
 equals the 50.50 network,
 which is back behind me.

0:12:23.660000 --> 0:12:28.360000
 And my L-Fib says, hey, okay, and then
 I'm going to advertise the label

0:12:28.360000 --> 0:12:32.760000
 five to you. So, just like there's routing
 protocols that advertise routes

0:12:32.760000 --> 0:12:38.500000
 like EIGRP and OSPF, there's a protocol
 called LDP, the label distribution

0:12:38.500000 --> 0:12:40.380000
 protocol that you and I have.

0:12:40.380000 --> 0:12:44.420000
 So, you and I are going to form an LDP
 neighbor relationship just alongside

0:12:44.420000 --> 0:12:48.940000
 our OSPF neighbor relationship or
 our EIGRP neighbor relationship.

0:12:48.940000 --> 0:12:52.760000
 And once this LDP neighbor relationship
 is up, I can use that to advertise

0:12:52.760000 --> 0:13:00.100000
 labels to you. So, I can say, hey, the
 50.50 network equals label number

0:13:00.100000 --> 0:13:05.200000
 five. And so, now you put that in
 your routing to in your L-Fib.

0:13:05.200000 --> 0:13:10.040000
 So, let's just draw this here real quick
 to make this a little bit more

0:13:10.040000 --> 0:13:18.860000
 meaningful. So, here comes
 in the 50.50 network to me.

0:13:18.860000 --> 0:13:21.260000
 I'll just say I'm router one.

0:13:21.260000 --> 0:13:28.380000
 So, what I come up with, and let's
 say here we have router two, router

0:13:28.380000 --> 0:13:32.100000
 three, and router four.

0:13:32.100000 --> 0:13:39.520000
 Okay, so, I am router one,
 I am a label edge router.

0:13:39.520000 --> 0:13:44.360000
 So, anything I send out this way is going
 to be unlabeled, just a regular

0:13:44.360000 --> 0:13:53.360000
 old IP packet. I form an
 LDP connection with you.

0:13:53.360000 --> 0:13:57.540000
 You form an LDP connection
 with somebody else.

0:13:57.540000 --> 0:14:00.560000
 E forms an LDP connection
 with somebody else.

0:14:00.560000 --> 0:14:10.440000
 And then maybe there's other also a label
 edge router, just like me, router

0:14:10.440000 --> 0:14:14.940000
 four is. Okay, so, once I learned about
 that network, so let's assume

0:14:14.940000 --> 0:14:17.920000
 the network has been propagated
 via some routing protocol.

0:14:17.920000 --> 0:14:20.080000
 We don't really care which it is.

0:14:20.080000 --> 0:14:34.200000
 So, I create in my L-Fib this mapping
 saying that 50.50.0.0 equals label

0:14:34.200000 --> 0:14:42.340000
 five. And I say this is not the technical
 term, but I'm just going to

0:14:42.340000 --> 0:14:48.200000
 say remove it. You'll see why
 I say this in just a second.

0:14:48.200000 --> 0:14:53.740000
 Okay, now I advertise
 via LDP to you.

0:14:53.740000 --> 0:15:02.340000
 I say 50.50 equals five.

0:15:02.340000 --> 0:15:08.980000
 So, in your table,
 you create this.

0:15:08.980000 --> 0:15:13.860000
 You say okay, I just
 learned that five.

0:15:13.860000 --> 0:15:22.500000
 And let's say here this
 is your interface one.

0:15:22.500000 --> 0:15:30.300000
 So, five equals the 50.50.0.0.

0:15:30.300000 --> 0:15:36.000000
 And I learned it on interface one.


0:15:36.000000 --> 0:15:42.280000
 And that is my egress.

0:15:42.280000 --> 0:15:46.220000
 So, let's put right here.

0:15:46.220000 --> 0:15:50.320000
 I could spell egress correctly.

0:15:50.320000 --> 0:16:01.800000
 Now what you're going to do is you're
 going to come up with this kind

0:16:01.800000 --> 0:16:04.640000
 of weird, but you're going to come up
 with your own label number for this

0:16:04.640000 --> 0:16:07.540000
 exact same network.

0:16:07.540000 --> 0:16:12.220000
 You're going to say okay,
 let's maybe have 20.

0:16:12.220000 --> 0:16:25.040000
 Ingress. Ingress.

0:16:25.040000 --> 0:16:32.140000
 And now you're going to advertise
 20 to this guy.

0:16:32.140000 --> 0:16:35.640000
 And so I'm just going to abbreviate
 some of this stuff right here.

0:16:35.640000 --> 0:16:39.400000
 Three, four, five, six.

0:16:39.400000 --> 0:16:45.300000
 And then this guy right here, his table
 is going to look something like

0:16:45.300000 --> 0:16:56.020000
 this. I have just learned
 of an egress label of 20.

0:16:56.020000 --> 0:17:03.820000
 Use the same color scheme
 here on interface three.

0:17:03.820000 --> 0:17:10.080000
 He says now I'm going to
 come up with an egress.

0:17:10.080000 --> 0:17:19.340000
 Oh no. Yeah, an ingress
 label of 42.

0:17:19.340000 --> 0:17:23.700000
 He sends out 42.

0:17:23.700000 --> 0:17:26.000000
 Almost done with an example.

0:17:26.000000 --> 0:17:27.960000
 Just bear with me.

0:17:27.960000 --> 0:17:41.580000
 This guy now says I have an egress of
 42, which was learned on interface

0:17:41.580000 --> 0:17:52.880000
 five. And we'll just
 say no ingress.

0:17:52.880000 --> 0:17:55.940000
 Okay, so here's how this
 is going to work.

0:17:55.940000 --> 0:18:04.400000
 Now here comes a bunch of just
 regular old IP packets.

0:18:04.400000 --> 0:18:08.100000
 And their destination address is actually
 the 5050 network way out there

0:18:08.100000 --> 0:18:14.500000
 to the left. They hit router four,
 who is our label edge router.

0:18:14.500000 --> 0:18:21.240000
 So these are going, so I'll
 just say here destination.

0:18:21.240000 --> 0:18:24.260000
 5050 1.1, let's say.

0:18:24.260000 --> 0:18:26.860000
 Okay, he looks it up via Ceph.

0:18:26.860000 --> 0:18:31.380000
 And according to a Ceph table, he says,
 oh, okay, I need to apply an egress

0:18:31.380000 --> 0:18:38.080000
 label of 42 to that and send those
 packets out interface number five.

0:18:38.080000 --> 0:18:39.700000
 So here they go.

0:18:39.700000 --> 0:18:45.940000
 So now we have applied an MPLS label
 to them all with the number 42.

0:18:45.940000 --> 0:18:46.420000
 And now we have a label of 42.

0:18:46.420000 --> 0:18:51.420000
 Router three gets it ingress.

0:18:51.420000 --> 0:18:53.060000
 He's receiving packets now.

0:18:53.060000 --> 0:18:58.220000
 Now he's receiving ingress something
 with a label of 42.

0:18:58.220000 --> 0:19:01.700000
 He says, oh, I'm receiving something
 ingress with a label 42.

0:19:01.700000 --> 0:19:08.380000
 I need to swap that label for 20 and send
 it out interface egress interface

0:19:08.380000 --> 0:19:14.040000
 number three. So he sends those
 packets out this way.

0:19:14.040000 --> 0:19:18.800000
 And now they have 20 on them, which is
 what he learned from his downstream

0:19:18.800000 --> 0:19:24.300000
 neighbor. So you see right here router
 three just forwarded these packets,

0:19:24.300000 --> 0:19:26.300000
 not based on their IP headers.

0:19:26.300000 --> 0:19:31.140000
 He forwarded them because when they
 came in, they had label 42 on them,

0:19:31.140000 --> 0:19:33.100000
 which is a label he created.

0:19:33.100000 --> 0:19:35.300000
 They came in as 42.

0:19:35.300000 --> 0:19:37.300000
 That map to label 20.

0:19:37.300000 --> 0:19:39.860000
 Outbound interface number three.

0:19:39.860000 --> 0:19:42.220000
 Same thing's going to happen
 here on router two.

0:19:42.220000 --> 0:19:47.980000
 Router two receives ingress label 20,
 which is what he, a label number

0:19:47.980000 --> 0:19:49.260000
 he came up with.

0:19:49.260000 --> 0:19:53.980000
 He says, oh, I need to swap that for
 label five and send it egress on

0:19:53.980000 --> 0:19:55.500000
 interface number one.

0:19:55.500000 --> 0:19:57.760000
 So here they go.

0:19:57.760000 --> 0:20:01.780000
 This way with the
 number five on it.

0:20:01.780000 --> 0:20:06.960000
 And now when they reach router one,
 router one says, oh, ingress label

0:20:06.960000 --> 0:20:09.660000
 number five is something
 I need to remove.

0:20:09.660000 --> 0:20:14.460000
 I need to pop it off and just send
 the packets in their native format.

0:20:14.460000 --> 0:20:19.380000
 So that's how labels are propagated
 and how label switching occurs.

0:20:19.380000 --> 0:20:24.020000
 It was the label distribution protocol
 that propagated label hop by hop

0:20:24.020000 --> 0:20:28.180000
 by hop. And then as things come in,
 they're labeled, they come in with

0:20:28.180000 --> 0:20:32.600000
 labels. So if you send something to
 me, you're going to use the label

0:20:32.600000 --> 0:20:35.200000
 that I just advertised to you.

0:20:35.200000 --> 0:20:39.760000
 And when I get my own label, I will,
 I will do one of two things.

0:20:39.760000 --> 0:20:45.440000
 I will either swap it with a label
 that he advertised to me, or if I'm

0:20:45.440000 --> 0:20:48.820000
 the last guy in the chain, if I'm the
 label edge router, I'll pop that

0:20:48.820000 --> 0:20:53.320000
 label off, revealing my IP packet,
 and then forward it as a regular IP

0:20:53.320000 --> 0:21:06.900000
 packet. So that's where it says
 labels have local significance.

0:21:06.900000 --> 0:21:11.660000
 Each label switch router independently
 maps a label to a forwarding equivalent

0:21:11.660000 --> 0:21:15.260000
 class. A forwarding equivalent
 class is a route.

0:21:15.260000 --> 0:21:16.480000
 Let's just, it's just simple.

0:21:16.480000 --> 0:21:18.040000
 Let's just think of it as a route.


0:21:18.040000 --> 0:21:23.200000
 In this case, 50.50, that was my
 forwarding equivalent class.

0:21:23.200000 --> 0:21:26.020000
 And label bindings are
 exchanged between LSRs.

0:21:26.020000 --> 0:21:26.780000
 We just looked at that.

0:21:26.780000 --> 0:21:30.100000
 All right, finally, what does
 the label actually look like?

0:21:30.100000 --> 0:21:32.820000
 Here it is. Here's all 32 bits.

0:21:32.820000 --> 0:21:37.680000
 You can see the first 20 bits from bit
 zero to bit 19, since we're starting

0:21:37.680000 --> 0:21:40.140000
 at zero, that averaged out to 20.

0:21:40.140000 --> 0:21:41.620000
 That's our actual label number.

0:21:41.620000 --> 0:21:44.500000
 That's where that number five
 or 42 or 20 would go in there.

0:21:44.500000 --> 0:21:46.440000
 So we have a 20-bit label.

0:21:46.440000 --> 0:21:52.120000
 We have a three-bit experimental field
 where we can carry IP precedence.

0:21:52.120000 --> 0:21:54.980000
 We have a one-bit bottom
 of stack indicator.

0:21:54.980000 --> 0:21:58.640000
 I'll just show you what that looks
 like in just a second and an eight

0:21:58.640000 --> 0:22:12.720000
-bit TTL. So how the TTL works is that
 going back to this situation right

0:22:12.720000 --> 0:22:19.620000
 here, when this packet came in, this
 regular IP packet, it had some TTL

0:22:19.620000 --> 0:22:26.720000
 in it, right? Maybe the TTL, when it
 was received by router 4, was 50.

0:22:26.720000 --> 0:22:31.000000
 So router 4 would have
 decremented that to 49.

0:22:31.000000 --> 0:22:34.480000
 And then when he applied the label
 on it, the MPLS label, they had the

0:22:34.480000 --> 0:22:36.420000
 number of 42 in it.

0:22:36.420000 --> 0:22:40.980000
 In the TTL field, he would
 have copied 49 into that.

0:22:40.980000 --> 0:22:47.260000
 So whatever the TTL is, let's go back
 to here, on the IP packet, that's

0:22:47.260000 --> 0:22:51.640000
 what's copied into the TTL
 field in the MPLS label.

0:22:51.640000 --> 0:22:57.060000
 And like I said, the MPLS label is
 imposed, that's a technical term.

0:22:57.060000 --> 0:23:00.940000
 In other words, it's put, it's imposed
 in between the layer 2 frame header

0:23:00.940000 --> 0:23:08.160000
 and the layer 3, presumably
 IPv4, IPv6 header.

0:23:08.160000 --> 0:23:12.100000
 And the last thing I wanted to talk about
 is that bottom of stack indicator.

0:23:12.100000 --> 0:23:17.180000
 In my example, in my whiteboard example,
 as that packet was going through

0:23:17.180000 --> 0:23:20.760000
 our label switching routers,
 it only had one label.

0:23:20.760000 --> 0:23:24.540000
 It would come in with a label, we'd
 pop that label off and we would swap

0:23:24.540000 --> 0:23:26.300000
 it for another one.

0:23:26.300000 --> 0:23:30.400000
 So it was always just one, if we capture
 that packet on a link in a sniffer

0:23:30.400000 --> 0:23:33.300000
 capture, there just would have
 been one label in there.

0:23:33.300000 --> 0:23:37.200000
 But in a lot of times, if we're talking
 about other things like MPLS VPNs

0:23:37.200000 --> 0:23:42.760000
 or MPLS traffic engineering, there
 could be two or even three labels,

0:23:42.760000 --> 0:23:47.540000
 daisy chain one after the other, inside
 your, you know, between your layer

0:23:47.540000 --> 0:23:51.160000
 2 header and your packet, you
 can have two or three of them.

0:23:51.160000 --> 0:23:57.620000
 So like in this case right here, so
 typically speaking, when I'm a label

0:23:57.620000 --> 0:24:02.160000
 switch router and I receive something
 that's already got a label in it,

0:24:02.160000 --> 0:24:08.640000
 I will look at the label that's
 right next to the frame header.

0:24:08.640000 --> 0:24:10.660000
 We call that the outer label.

0:24:10.660000 --> 0:24:15.940000
 So it's going this direction
 right here, this direction.

0:24:15.940000 --> 0:24:18.840000
 Here's my normal IP header
 and my payload.

0:24:18.840000 --> 0:24:20.720000
 Here's my Ethernet header.

0:24:20.720000 --> 0:24:25.580000
 Now the Ethernet header has
 an ether type of 0x8847.

0:24:25.580000 --> 0:24:27.360000
 Don't have to memorize that.

0:24:27.360000 --> 0:24:31.140000
 But that's the ether type value that says,
 hey, behind me, behind my ethernet

0:24:31.140000 --> 0:24:33.120000
 header, is not an IP header.

0:24:33.120000 --> 0:24:36.140000
 It's a label, it's an MPLS label.

0:24:36.140000 --> 0:24:40.520000
 And so as a router, when I get an ethernet
 frame and I see that ether

0:24:40.520000 --> 0:24:44.420000
 type value, that tells me, oh, I'm
 not going to look at my system.

0:24:44.420000 --> 0:24:47.860000
 Except fib, I would look at my
 fib if it was an IP header.

0:24:47.860000 --> 0:24:51.260000
 I'm going to look at my L fib, my label
 fib, because that's going to tell

0:24:51.260000 --> 0:24:54.860000
 me, oh, there's a label of five,
 what do I do with that?

0:24:54.860000 --> 0:24:57.780000
 My L fib will tell me
 what to do with that.

0:24:57.780000 --> 0:25:01.800000
 So when I look at the label, this is
 the label I'm looking at, the outer

0:25:01.800000 --> 0:25:07.540000
 label. Now each label tells you basically
 the bottom of stack indicator

0:25:07.540000 --> 0:25:12.180000
 is basically an indicator of am I the
 last label or are there other labels

0:25:12.180000 --> 0:25:16.000000
 behind me? That's what the
 bottom of stack bit does.

0:25:16.000000 --> 0:25:22.400000
 So you can see this label right here,
 his bottom of stack indicator is

0:25:22.400000 --> 0:25:26.980000
 a zero. If it's a zero, that means I
 am not at the bottom of the stack.

0:25:26.980000 --> 0:25:29.280000
 There's other labels behind me.

0:25:29.280000 --> 0:25:33.160000
 Next label, maybe this
 is a label for a VPN.

0:25:33.160000 --> 0:25:35.340000
 Once again, it's not
 the last label.

0:25:35.340000 --> 0:25:37.680000
 So it's not at the bottom
 of the stack.

0:25:37.680000 --> 0:25:40.040000
 This is the last label.

0:25:40.040000 --> 0:25:45.420000
 And so the bottom of stack
 indicator is a one.

0:25:45.420000 --> 0:25:50.780000
 So a lot of stuff to learn, but for
 the CCNA purposes, this pretty much

0:25:50.780000 --> 0:25:55.040000
 covers everything you need
 to know about MPLS.

0:25:55.040000 --> 0:25:58.640000
 So thank you for watching this video,
 and I hope it was useful.

