WEBVTT

0:00:08.040000 --> 0:00:14.440000
 One of the very popular services are
 provided by WAN provider is MPLS,

0:00:14.440000 --> 0:00:17.420000
 which is multi-protocol
 label switching.

0:00:17.420000 --> 0:00:20.580000
 And at the CCNA level, they want you
 to have a real high understanding

0:00:20.580000 --> 0:00:24.400000
 of MPLS and some of the terminology
 that goes along with it.

0:00:24.400000 --> 0:00:29.660000
 So let's talk about that.

0:00:29.660000 --> 0:00:35.280000
 So, with traditional routing, a pack,
 a layer 3 pack, it will come into

0:00:35.280000 --> 0:00:39.300000
 a router, and that router will
 strip off the layer 2 header.

0:00:39.300000 --> 0:00:43.280000
 It might come in with an Ethernet
 header or a PPP header.

0:00:43.280000 --> 0:00:46.540000
 The router will strip off that layer
 2 header, and then right behind it

0:00:46.540000 --> 0:00:48.480000
 will be the layer 3 header.

0:00:48.480000 --> 0:00:51.980000
 And then it will look at the destination
 layer 3 address, like an IP version

0:00:51.980000 --> 0:00:54.620000
 4 or an IP version 6 address.

0:00:54.620000 --> 0:00:58.780000
 And then that router will go into some
 table, like a routing table, or

0:00:58.780000 --> 0:01:02.780000
 if that router is running Cisco Express
 forwarding, a copy of the routing

0:01:02.780000 --> 0:01:06.420000
 table, so the essential elements from
 the routing table will be placed

0:01:06.420000 --> 0:01:08.280000
 into a CEP table.

0:01:08.280000 --> 0:01:11.720000
 But either way, the router will look for
 some sort of match of that destination

0:01:11.720000 --> 0:01:16.880000
 IP address, and then create a new layer
 2 header and forward the packet

0:01:16.880000 --> 0:01:22.940000
 on out. With MPLS, which stands for
 multi-protocol label switching, I

0:01:22.940000 --> 0:01:26.960000
 want you to pay attention to those
 last two letters, label switching.

0:01:26.960000 --> 0:01:29.220000
 MPLS actually does something
 different.

0:01:29.220000 --> 0:01:34.700000
 So when a frame comes in, a router
 in the WAN network, so the service

0:01:34.700000 --> 0:01:39.500000
 providers router, will actually take
 that layer 2 header and right behind

0:01:39.500000 --> 0:01:45.080000
 it, just before the layer 3 header,
 it will insert a new field called

0:01:45.080000 --> 0:01:47.460000
 an MPLS label, called a label.

0:01:47.460000 --> 0:01:50.620000
 This might sound kind of familiar to
 you, once again, drawing an analogy

0:01:50.620000 --> 0:01:54.120000
 to switching, which hopefully you've
 already studied up to this point.

0:01:54.120000 --> 0:01:57.700000
 We know that in the world of switching,
 when Ethernet frame is going across

0:01:57.700000 --> 0:02:02.200000
 an 802.1Q trunk, same
 type of thing happens.

0:02:02.200000 --> 0:02:06.280000
 When that frame gets to the switch,
 the switch takes and puts an 802.1Q

0:02:06.280000 --> 0:02:13.160000
 tag after the source MAC address, but
 before the layer 3 packet, it inserts

0:02:13.160000 --> 0:02:16.540000
 that tag there. So the same type of thing
 is happening here, except instead

0:02:16.540000 --> 0:02:20.680000
 of a .1Q tag being applied, or applying
 something else called an MPLS

0:02:20.680000 --> 0:02:24.760000
 label. And then that label is actually
 what's used by all these routers

0:02:24.760000 --> 0:02:27.540000
 to forward the packet.

0:02:27.540000 --> 0:02:30.940000
 And actually, the reason why it's called
 multi-protocol label switching

0:02:30.940000 --> 0:02:35.780000
 is because MPLS can apply that label
 to lots of different things.

0:02:35.780000 --> 0:02:39.860000
 We typically think of MPLS as being used
 against IP packets, like IP version

0:02:39.860000 --> 0:02:42.040000
 4 or IP version 6.

0:02:42.040000 --> 0:02:45.380000
 But MPLS can actually be used against
 a wide variety of things, not even

0:02:45.380000 --> 0:02:48.380000
 necessarily layer 3 packets.

0:02:48.380000 --> 0:02:51.120000
 That's why it's called
 multi-protocol.

0:02:51.120000 --> 0:02:55.960000
 The key idea though is that the routers
 in the service provider's cloud,

0:02:55.960000 --> 0:02:59.280000
 when they're trying to figure out where
 does this thing go, they're no

0:02:59.280000 --> 0:03:02.600000
 longer looking at a layer 3 header
 to answer that question.

0:03:02.600000 --> 0:03:04.400000
 They're now looking at this label.


0:03:04.400000 --> 0:03:08.700000
 They have a table of labels that says,
 if I see this label, I'm going

0:03:08.700000 --> 0:03:10.320000
 to send it out this interface.

0:03:10.320000 --> 0:03:12.340000
 So it's label switching.

0:03:12.340000 --> 0:03:17.180000
 So this here shows an example
 of what that label looks like.

0:03:17.180000 --> 0:03:21.660000
 So the label is a 32-bit field, which
 should hopefully be easy to remember

0:03:21.660000 --> 0:03:28.040000
 because IP version 4 addresses are
 32 bits, MPLS labels are 32 bits.

0:03:28.040000 --> 0:03:31.040000
 And you can see here the
 breakdown of that label.

0:03:31.040000 --> 0:03:33.760000
 Now, I'm going to come back
 to this in just one second.

0:03:33.760000 --> 0:03:36.040000
 I just want to flash forward
 here for a second.

0:03:36.040000 --> 0:03:38.660000
 And just to give you an idea
 of where this label goes.

0:03:38.660000 --> 0:03:47.120000
 So here is our regular frame with our
 layer 2 header, frame relay, PPP,

0:03:47.120000 --> 0:03:50.480000
 Ethernet, whatever it is,
 followed by our IP header.

0:03:50.480000 --> 0:03:54.540000
 Well, in the world of MPLS, we've got
 our frame header, and then the label

0:03:54.540000 --> 0:03:57.480000
 is placed right here.

0:03:57.480000 --> 0:03:59.360000
 So that's where the label goes.

0:03:59.360000 --> 0:04:04.020000
 And sometimes you'll hear this called
 as a shim header, as a shim header.

0:04:04.020000 --> 0:04:05.760000
 But that's where the label goes.

0:04:05.760000 --> 0:04:09.420000
 So now let's go back to the previous
 slide and take a look at what does

0:04:09.420000 --> 0:04:13.240000
 this label look like if we break it
 down, what components are in that

0:04:13.240000 --> 0:04:17.480000
 label. OK, so here is the
 label and its entirety.

0:04:17.480000 --> 0:04:24.520000
 So most of the label, so 20 bits of
 it, 20 bits out of the 32, are the

0:04:24.520000 --> 0:04:27.860000
 label value. So you can see there's
 actually quite a few numbers you can

0:04:27.860000 --> 0:04:30.340000
 come up with with a 20-bit field.

0:04:30.340000 --> 0:04:34.740000
 Then after that, we have a three-bit field,
 which is called the experimental

0:04:34.740000 --> 0:04:39.800000
 field. So one of the things that MPLS
 Service Peraras can offer is something

0:04:39.800000 --> 0:04:41.580000
 called quality of service.

0:04:41.580000 --> 0:04:44.720000
 And this experimental field is
 what allows them to do that.

0:04:44.720000 --> 0:04:49.340000
 So you can send packets into the Service
 Provider's network, maybe some

0:04:49.340000 --> 0:04:53.300000
 of those packets carry data, maybe
 some of those packets carry voice.

0:04:53.300000 --> 0:04:59.500000
 And by modifying or changing the bits
 here in the experimental field of

0:04:59.500000 --> 0:05:04.440000
 the label, the voice packets can be differentiated
 and identified differently

0:05:04.440000 --> 0:05:07.120000
 from the data packets so they
 can get better service.

0:05:07.120000 --> 0:05:10.840000
 They can take faster paths,
 paths with less latency.

0:05:10.840000 --> 0:05:13.720000
 That's because of the experimental
 field right here, which carries our

0:05:13.720000 --> 0:05:16.200000
 quality of service markings.

0:05:16.200000 --> 0:05:20.700000
 And then there's a bit here called
 the bottom of the stack indicator,

0:05:20.700000 --> 0:05:23.080000
 bottom of the stack indicator.

0:05:23.080000 --> 0:05:29.260000
 And the reason why that's there is
 because in a lot of MPLS offerings,

0:05:29.260000 --> 0:05:33.680000
 I'm going to talk briefly about MPLS
 VPNs as an example, you might see

0:05:33.680000 --> 0:05:38.160000
 a situation where there's two or more
 labels applied to something, two

0:05:38.160000 --> 0:05:40.560000
 or sometimes even three labels.

0:05:40.560000 --> 0:05:46.960000
 And so an MPLS device, an MPLS router,
 only looks at the top label, what's

0:05:46.960000 --> 0:05:50.440000
 the top label, the label that's right
 next to the layer two header.

0:05:50.440000 --> 0:05:55.820000
 So if I go forward for a moment, so for
 example, right here, here we have

0:05:55.820000 --> 0:06:00.540000
 an example of a frame that has
 three labels inside of it.

0:06:00.540000 --> 0:06:05.940000
 So we can see that the label that's
 right after the layer two header,

0:06:05.940000 --> 0:06:08.320000
 that's what we consider
 the top label.

0:06:08.320000 --> 0:06:12.780000
 And that's the label that an MPLS router
 will look at to determine where

0:06:12.780000 --> 0:06:13.860000
 do I send this thing?

0:06:13.860000 --> 0:06:17.360000
 What's the egress interface where
 I send this data structure?

0:06:17.360000 --> 0:06:18.960000
 It's the top label.

0:06:18.960000 --> 0:06:21.980000
 But then there can be, and that's called
 the top, then there can be labels

0:06:21.980000 --> 0:06:26.980000
 after that. And then the very last
 label, the very last one is what we

0:06:26.980000 --> 0:06:28.900000
 call the bottom label.

0:06:28.900000 --> 0:06:32.740000
 And the bottom label has that one little
 bit set, that bottom of the stack

0:06:32.740000 --> 0:06:34.220000
 bit set to a one.

0:06:34.220000 --> 0:06:35.940000
 Why is that important?

0:06:35.940000 --> 0:06:42.800000
 Because when a router receives an incoming
 MPLS labeled structure, and

0:06:42.800000 --> 0:06:45.920000
 it sees that label there, and it sees
 that the bottom of the stack bit

0:06:45.920000 --> 0:06:49.020000
 is set to a one, that
 router knows aha.

0:06:49.020000 --> 0:06:53.100000
 Okay, I need to pop that label off and
 revert this thing back to looking

0:06:53.100000 --> 0:06:57.060000
 like a normal IP packet without
 any label at all.

0:06:57.060000 --> 0:07:00.300000
 So that bottom of the stack bit indicates
 that this is the very last label,

0:07:00.300000 --> 0:07:04.140000
 and behind it, you're going to see
 your regular layer three header.

0:07:04.140000 --> 0:07:08.620000
 So that's what that bottom
 of stack bit is used for.

0:07:08.620000 --> 0:07:11.800000
 So let me go back.

0:07:11.800000 --> 0:07:15.720000
 And then also we have at the very
 end, we have a TTL field.

0:07:15.720000 --> 0:07:20.460000
 So just like an IP packet has a time
 to live field, an MPLS label has

0:07:20.460000 --> 0:07:21.400000
 a time to live field.

0:07:21.400000 --> 0:07:27.520000
 Typically speaking, when an IP packet
 comes into an MPLS router, the router

0:07:27.520000 --> 0:07:31.000000
 will take a look at the TTL field in the
 IP packet, and it will take whatever

0:07:31.000000 --> 0:07:38.080000
 that number is, 18, 64, 3, and it will
 copy it into the TTL field right

0:07:38.080000 --> 0:07:41.900000
 here. And now as each router gets this,
 as each label router gets it and

0:07:41.900000 --> 0:07:45.940000
 switches it based on labels, just like
 we would normally take a TTL and

0:07:45.940000 --> 0:07:49.860000
 IP packet and decrement it by one, while
 our label switching routers will

0:07:49.860000 --> 0:07:54.560000
 also decrement the TTL value here
 in the label itself by one.

0:07:54.560000 --> 0:07:56.220000
 Same rules apply.

0:07:56.220000 --> 0:08:03.160000
 Okay, so I think we've seen
 everything on here.

0:08:03.160000 --> 0:08:07.220000
 There's nothing we already talked about
 where the labels apply, and this

0:08:07.220000 --> 0:08:11.260000
 is just showing you that's right
 after the layer 2 header.

0:08:11.260000 --> 0:08:16.660000
 And sometimes you'll see that MPLS
 is called a layer 2.5 protocol.

0:08:16.660000 --> 0:08:19.620000
 In other words, when talking about
 the OSI model, and people are all,

0:08:19.620000 --> 0:08:20.500000
 where is this reside?

0:08:20.500000 --> 0:08:24.220000
 It reside at the data link layer,
 at the network layer.

0:08:24.220000 --> 0:08:29.260000
 They say it's a layer 2.5 protocol because
 it's inserting the label between

0:08:29.260000 --> 0:08:32.720000
 the layer 2 header and
 the layer 3 header.

0:08:32.720000 --> 0:08:34.620000
 So it's a layer 2.5 protocol.

0:08:34.620000 --> 0:08:37.840000
 So you might hear that term.

0:08:37.840000 --> 0:08:42.180000
 And so we've already talked about this,
 and just this grouping of labels

0:08:42.180000 --> 0:08:46.520000
 goes by the term label
 stack, a label stack.

0:08:46.520000 --> 0:08:51.280000
 So you could have one or more
 labels in your label stack.

0:08:51.280000 --> 0:08:55.480000
 And just remember the key term of the
 top label, that's the one right

0:08:55.480000 --> 0:08:59.980000
 next to the layer 2 header and the bottom
 label, which is one right next

0:08:59.980000 --> 0:09:01.360000
 to the layer 3 header.

0:09:01.360000 --> 0:09:03.880000
 Of course if you've only got one
 label, they're both the same.

0:09:03.880000 --> 0:09:09.260000
 The top and the bottom are exactly the
 same, so you've only got one label.

0:09:09.260000 --> 0:09:14.140000
 And some terminology that
 we need to know for MPLS.

0:09:14.140000 --> 0:09:21.920000
 So a router that has the capability of
 looking at a label, something that

0:09:21.920000 --> 0:09:25.460000
 comes in that already has a label on
 it and saying, okay, based on this

0:09:25.460000 --> 0:09:29.980000
 label value, not the routing table,
 but my label table, I'm going to be

0:09:29.980000 --> 0:09:34.020000
 able to figure out which interface
 I need to forward this packet on or

0:09:34.020000 --> 0:09:35.940000
 forward this thing on.

0:09:35.940000 --> 0:09:40.900000
 We technically call that
 a label switch router.

0:09:40.900000 --> 0:09:45.920000
 And label switch routers fall into sort
 of three categories, ingress label

0:09:45.920000 --> 0:09:51.260000
 switch router, intermediate label switch
 router, and egress label switch

0:09:51.260000 --> 0:09:55.020000
 router. So let me show you an
 example, a picture of this.

0:09:55.020000 --> 0:09:58.720000
 So take a look at this
 as an example.

0:09:58.720000 --> 0:10:03.340000
 So let's say that each of these boxes
 right here represents a customer

0:10:03.340000 --> 0:10:13.060000
 site. And so this router here, the
 brown routers, these belong to the

0:10:13.060000 --> 0:10:16.600000
 customer. So these are the customer
 routers and these routers are at the

0:10:16.600000 --> 0:10:19.120000
 edge of the customer's network.

0:10:19.120000 --> 0:10:21.240000
 In other words, these are at the edge
 of the network that connect to the

0:10:21.240000 --> 0:10:26.820000
 ISP. So from an MPLS terminology, these
 routers are called customer edge

0:10:26.820000 --> 0:10:30.260000
 routers or CE routers.

0:10:30.260000 --> 0:10:34.720000
 And yes, you will want to be familiar
 with this terminology, customer

0:10:34.720000 --> 0:10:40.000000
 edge routers. Okay.

0:10:40.000000 --> 0:10:45.160000
 Now, all the routers here in the MPLS
 providers network are called label

0:10:45.160000 --> 0:10:47.960000
 switching routers, every
 single one of them.

0:10:47.960000 --> 0:10:51.320000
 Because they either have the ability
 of taking in a regular IP packet

0:10:51.320000 --> 0:10:55.420000
 and slapping a new label on top
 of it that did not exist before.

0:10:55.420000 --> 0:11:00.120000
 They have the capability of taking in
 a labeled packet, looking at that

0:11:00.120000 --> 0:11:03.820000
 label and figuring out which egress
 interface to switch it out of.

0:11:03.820000 --> 0:11:07.940000
 Or if it's on the other edge, it has
 the ability of taking the label,

0:11:07.940000 --> 0:11:11.580000
 stripping it off completely, revealing
 just a regular IP packet and then

0:11:11.580000 --> 0:11:18.660000
 sending it out. So if we're talking
 about customer A here, customer A

0:11:18.660000 --> 0:11:30.660000
 site one and customer A site two, in
 terminology we have the customer

0:11:30.660000 --> 0:11:34.820000
 edge router. So let's say that our
 packet comes in this way, I'll just

0:11:34.820000 --> 0:11:39.760000
 say P for packet and
 it gets right here.

0:11:39.760000 --> 0:11:50.580000
 So this here would be called
 an ingress label edge router.

0:11:50.580000 --> 0:11:54.700000
 Because it's ingress, it's receiving
 the ingress packet and it's on the

0:11:54.700000 --> 0:11:58.560000
 edge of the labeling network,
 the label edge router.

0:11:58.560000 --> 0:12:06.100000
 So now this router here would take that
 packet and it would apply a label

0:12:06.100000 --> 0:12:09.520000
 to it, like we just
 took a look at.

0:12:09.520000 --> 0:12:15.240000
 And then as it went into here, these
 routers in the middle would just

0:12:15.240000 --> 0:12:17.400000
 be called intermediate.

0:12:17.400000 --> 0:12:20.520000
 See if I can put that right here.

0:12:20.520000 --> 0:12:32.520000
 Intermediate label
 switching routers.

0:12:32.520000 --> 0:12:35.800000
 So an intermediate label switching
 router, all it does is it receives

0:12:35.800000 --> 0:12:39.660000
 something that already has a label on
 it and then just switches it through

0:12:39.660000 --> 0:12:42.180000
 based on that label.

0:12:42.180000 --> 0:12:46.880000
 Now once it gets to this guy right
 here, so here it comes in, it's got

0:12:46.880000 --> 0:12:52.540000
 a label, we also have
 the packet behind it.

0:12:52.540000 --> 0:12:57.040000
 So it comes in. So this guy here is
 going to take that label, he's going

0:12:57.040000 --> 0:12:59.640000
 to recognize that okay this is the
 last label, there's no other label

0:12:59.640000 --> 0:13:08.720000
 left, he's going to strip it off and
 he's going so he will be called an

0:13:08.720000 --> 0:13:18.400000
 egress label edge router.

0:13:18.400000 --> 0:13:23.280000
 So all of these are label switching routers,
 all of them are LSRs because

0:13:23.280000 --> 0:13:28.240000
 they have the capability of processing
 and understanding labels.

0:13:28.240000 --> 0:13:33.220000
 A label edge router like it sounds is
 at the edge of the MPLS providers

0:13:33.220000 --> 0:13:38.520000
 network. And ingress or egress just refers
 to the direction of the traffic.

0:13:38.520000 --> 0:13:42.860000
 Is the traffic coming in or is it
 leaving the providers network?

0:13:42.860000 --> 0:13:45.620000
 So you should be familiar
 with those.

0:13:45.620000 --> 0:13:50.340000
 Now a lot of times also you'll see in
 a lot of books and documents that

0:13:50.340000 --> 0:13:54.760000
 in addition to the routers here on the
 edge being called label edge routers,

0:13:54.760000 --> 0:13:59.780000
 there's another term for those you'll
 see which is called provider edge

0:13:59.780000 --> 0:14:06.260000
 routers. Because router one and router
 two, they're owned by the service

0:14:06.260000 --> 0:14:10.780000
 provider. And just like a one and a
 two are owned by the customer but

0:14:10.780000 --> 0:14:15.500000
 they are at the edge of the customer's
 network, these two routers here

0:14:15.500000 --> 0:14:19.040000
 one and two, well they're owned by the
 provider and they are at the edge

0:14:19.040000 --> 0:14:21.180000
 of the provider's network.

0:14:21.180000 --> 0:14:27.640000
 So we have customer edge routers which
 connect to provider edge routers.

0:14:27.640000 --> 0:14:30.680000
 So CEs connect to PEs.

0:14:30.680000 --> 0:14:33.540000
 Now there's a couple of additional things
 about this I want to talk about.

0:14:33.540000 --> 0:14:37.280000
 One of the, you might be wondering okay
 well but really what's the benefit

0:14:37.280000 --> 0:14:42.200000
 of this? I mean so we're slapping on
 this new field called a label and

0:14:42.200000 --> 0:14:45.620000
 all of these label switching routers
 are figuring out where to forward

0:14:45.620000 --> 0:14:49.580000
 this thing based on a label instead
 of based on an IP header.

0:14:49.580000 --> 0:14:51.540000
 Why do we want to do that?

0:14:51.540000 --> 0:14:53.080000
 Where's the benefit of that?

0:14:53.080000 --> 0:14:58.180000
 A big reason why people like MPLS services
 is because they can get something

0:14:58.180000 --> 0:15:05.920000
 called MPLS VPNs which stands
 for virtual private network.

0:15:05.920000 --> 0:15:08.540000
 Well what makes it private?

0:15:08.540000 --> 0:15:12.120000
 Here's one of the great
 things about MPLS VPNs.

0:15:12.120000 --> 0:15:17.220000
 I could have customer A and maybe customer
 A is using the 10.10 network.

0:15:17.220000 --> 0:15:22.680000
 So you got all sorts of subnets of 10
.10 over here and all sorts of subnets

0:15:22.680000 --> 0:15:26.560000
 of 10.20 over here.

0:15:26.560000 --> 0:15:29.500000
 Now you might be saying but Keith that's
 a private address that's not

0:15:29.500000 --> 0:15:32.100000
 a publicly routable address
 actually that's okay.

0:15:32.100000 --> 0:15:36.900000
 In this particular example the customer
 is not using the MPLS VPN to access

0:15:36.900000 --> 0:15:41.660000
 the internet. All this WAN provider
 is doing is giving them access so

0:15:41.660000 --> 0:15:45.220000
 that one customer site can talk
 to another customer site.

0:15:45.220000 --> 0:15:49.320000
 It's not for internet access it's just
 connecting one campus or branch

0:15:49.320000 --> 0:15:57.220000
 to another. And then guess what down
 here in customer B we could be using

0:15:57.220000 --> 0:16:02.080000
 the exact same addresses.

0:16:02.080000 --> 0:16:10.660000
 10.20.0.0.16. And the MPLS VPN allows
 us to have overlapping address space

0:16:10.660000 --> 0:16:16.220000
 like this. And it allows the
 packets to be kept separate.

0:16:16.220000 --> 0:16:20.700000
 In other words customer A does not have
 to worry that his or her packets

0:16:20.700000 --> 0:16:24.760000
 will ever show up in
 customer B's network.

0:16:24.760000 --> 0:16:29.860000
 The MPLS VPN makes sure that the customer's
 packets are always separate,

0:16:29.860000 --> 0:16:33.360000
 isolated, they don't go where
 they're not supposed to go.

0:16:33.360000 --> 0:16:37.200000
 And from the provider's perspective
 it's great because they can reuse

0:16:37.200000 --> 0:16:41.340000
 IP addresses on different
 customers.

0:16:41.340000 --> 0:16:43.640000
 Now you might be thinking
 well how is that possible?

0:16:43.640000 --> 0:16:44.620000
 How does it even work?

0:16:44.620000 --> 0:16:48.560000
 Well without going into too many great
 details about this here's a high

0:16:48.560000 --> 0:16:50.140000
 level overview of this.

0:16:50.140000 --> 0:16:53.300000
 Up until now we've been
 talking about routing.

0:16:53.300000 --> 0:16:59.300000
 You've had this idea that a router
 has one single IP routing table.

0:16:59.300000 --> 0:17:01.920000
 Let's just focus on IP
 version 4 for now.

0:17:01.920000 --> 0:17:07.280000
 It's got one IP version 4 routing table
 and EIGRP routes or OSPF routes

0:17:07.280000 --> 0:17:09.600000
 or RIP routes can feed into that.

0:17:09.600000 --> 0:17:13.860000
 But ultimately when a packet comes into
 a router it's that one table the

0:17:13.860000 --> 0:17:18.080000
 router looks at we call that the global
 routing table to figure out what

0:17:18.080000 --> 0:17:19.300000
 do I do with this packet?

0:17:19.300000 --> 0:17:20.820000
 Where do I forward it?

0:17:20.820000 --> 0:17:25.060000
 Well when we're dealing with MPLS VPNs
 the router still has a global routing

0:17:25.060000 --> 0:17:29.100000
 table. So now what we're going to focus
 on is just the provider's network.

0:17:29.100000 --> 0:17:32.600000
 So the customers don't
 do MPLS at all.

0:17:32.600000 --> 0:17:37.600000
 An MPLS VPN or even if you're not doing
 VPN the customer routers the CE

0:17:37.600000 --> 0:17:39.800000
 routers and let me just
 put that on here.

0:17:39.800000 --> 0:17:44.740000
 They have no knowledge of MPLS.

0:17:44.740000 --> 0:17:46.720000
 They're not doing any
 kind of labeling.

0:17:46.720000 --> 0:17:51.240000
 They're just sending and receiving to
 the provider just regular IP packets

0:17:51.240000 --> 0:17:56.300000
 packaged in you know ethernet frames
 or PPP frames or whatever they choose

0:17:56.300000 --> 0:17:59.980000
 to use to connect to the provider.


0:17:59.980000 --> 0:18:03.260000
 So it's all the providers equipment in
 here that's doing all the labeling

0:18:03.260000 --> 0:18:07.620000
 and the swapping and doing all this MPLS
 VPN stuff I'm about to describe.

0:18:07.620000 --> 0:18:10.520000
 So this provider's network.

0:18:10.520000 --> 0:18:15.460000
 Now this provider's network is probably
 also used for them for internal

0:18:15.460000 --> 0:18:16.600000
 stuff for example.

0:18:16.600000 --> 0:18:19.720000
 They've got their own email system
 for their own employees.

0:18:19.720000 --> 0:18:22.960000
 They have their own internal websites
 for like their payroll, their human

0:18:22.960000 --> 0:18:24.580000
 resources and marketing and stuff.


0:18:24.580000 --> 0:18:29.080000
 So imagine that this network here is
 not only being used to transport

0:18:29.080000 --> 0:18:32.300000
 packets from one customer
 site to another.

0:18:32.300000 --> 0:18:36.360000
 It's actually being used also internally
 by the service provider itself

0:18:36.360000 --> 0:18:39.520000
 as their own enterprise
 network for themselves.

0:18:39.520000 --> 0:18:44.540000
 So there's going to be some routing
 protocol running here within these

0:18:44.540000 --> 0:18:48.560000
 routers. It could be anything
 whatever they want to use.

0:18:48.560000 --> 0:18:52.920000
 Let's just say it's doing OSPF.

0:18:52.920000 --> 0:18:58.360000
 Although it could be EIGRP, it could
 be RIP, it could be anything.

0:18:58.360000 --> 0:19:03.300000
 So all of these routers learn about
 each other and they're able to reach

0:19:03.300000 --> 0:19:07.700000
 each other via whatever interior gateway
 protocol the service provider

0:19:07.700000 --> 0:19:12.920000
 is running for themselves within
 their own network right here.

0:19:12.920000 --> 0:19:18.020000
 Okay so now in addition to that so
 all of that all of those routes are

0:19:18.020000 --> 0:19:21.300000
 being learned are being learned
 in the global routing tables.

0:19:21.300000 --> 0:19:25.560000
 And let's just put some terminology back
 on here that I mentioned earlier.

0:19:25.560000 --> 0:19:33.440000
 So we'll just say this is PE1 and
 this is provider edge 2 PE2.

0:19:33.440000 --> 0:19:38.940000
 Okay so for example PE1 if you did
 a show IP route command and looked

0:19:38.940000 --> 0:19:44.240000
 at his global routing table he would
 know how to reach 33344444 and all

0:19:44.240000 --> 0:19:45.580000
 these are the routers
 in the middle.

0:19:45.580000 --> 0:19:50.060000
 But here's what makes an MPLS VPN so
 powerful able to separate customer

0:19:50.060000 --> 0:19:54.320000
 routes, keep them separate even to
 the extent that one customer can be

0:19:54.320000 --> 0:19:58.580000
 using the exact same subnets as another
 customer which is the fact that

0:19:58.580000 --> 0:20:05.020000
 on these PE routers in addition to the
 global routing table each interface

0:20:05.020000 --> 0:20:10.680000
 that connects to a customer is in
 its own separate routing table.

0:20:10.680000 --> 0:20:13.920000
 It's like these little
 private routing tables.

0:20:13.920000 --> 0:20:20.960000
 We actually call these VRFs virtual routing
 and forwarding instances VRFs

0:20:20.960000 --> 0:20:28.140000
 of VRF. So right now as I've drawn
 it if we just focus in right now on

0:20:28.140000 --> 0:20:32.640000
 provider edge 1 he actually has three
 routing table he's keeping track

0:20:32.640000 --> 0:20:36.120000
 of he's got the global routing table
 for all the routes he's learned about

0:20:36.120000 --> 0:20:41.300000
 about his own internal network and he's
 got a separate private VRF routing

0:20:41.300000 --> 0:20:48.200000
 table for customer A and another
 VRF for customer B.

0:20:48.200000 --> 0:20:50.960000
 So right there that's one thing that
 keeps the route separate because

0:20:50.960000 --> 0:20:55.360000
 the 10-10 network that's learned and
 by the way between the CE and the

0:20:55.360000 --> 0:20:58.480000
 PE there's going to be some
 routing protocol running.

0:20:58.480000 --> 0:21:01.860000
 Let's just put that right here and once
 again that's completely between

0:21:01.860000 --> 0:21:05.820000
 the customer and between
 the provider.

0:21:05.820000 --> 0:21:17.020000
 So for example customer B they might
 choose to run EIGRP between their

0:21:17.020000 --> 0:21:23.660000
 customer edge router and the provider
 edge router and customer A might

0:21:23.660000 --> 0:21:35.400000
 choose to run routing going on here so
 that the customer can say hey service

0:21:35.400000 --> 0:21:39.800000
 provider here's all the routes I'm
 using in my network here they are.

0:21:39.800000 --> 0:21:44.040000
 So you can see that the customer edge
 router is forming a neighborhood

0:21:44.040000 --> 0:21:50.120000
 if it's EIGRP or OSPF with the service
 provider with the service provider's

0:21:50.120000 --> 0:21:53.820000
 PE router there's a neighborhood
 being formed.

0:21:53.820000 --> 0:21:57.080000
 Okay so that's one thing so that's how
 we keep the route separate because

0:21:57.080000 --> 0:22:00.840000
 these interfaces are actually assigned
 to their own private routing tables

0:22:00.840000 --> 0:22:08.520000
 called VRFs and then in order to actually
 complete the MPLS VPN to keep

0:22:08.520000 --> 0:22:15.260000
 the route separate there is actually
 from these PE routers right here

0:22:15.260000 --> 0:22:23.160000
 they are talking to each other via the
 border gateway protocol via BGP.

0:22:23.160000 --> 0:22:26.740000
 So I'm not going to go into much more
 detail than this this is actually

0:22:26.740000 --> 0:22:30.260000
 already going into more detail that the
 CCNA books go into that I've looked

0:22:30.260000 --> 0:22:35.140000
 at but the main takeaway I want you to
 have is that an MPLS VPN has several

0:22:35.140000 --> 0:22:39.920000
 benefits. Number one the customer's
 routes will be kept separate from

0:22:39.920000 --> 0:22:43.500000
 any other customer's routes by having
 this combination of virtual routing

0:22:43.500000 --> 0:22:49.720000
 and forwarding and tables by doing routing
 with the provider and by having

0:22:49.720000 --> 0:22:54.440000
 BGP running in the provider's network
 the combination of all that along

0:22:54.440000 --> 0:22:58.560000
 with labels labels play a factor in
 this labels play a very important

0:22:58.560000 --> 0:23:03.700000
 factor is what helps keep the customer's
 traffic separate so customer

0:23:03.700000 --> 0:23:08.000000
 A's traffic will never show
 up in customer B's network.

0:23:08.000000 --> 0:23:12.460000
 Also this MPLS VPN is a very good thing
 from the provider's point of view

0:23:12.460000 --> 0:23:17.020000
 because it helps them to conserve IP
 addresses they can provide the same

0:23:17.020000 --> 0:23:22.060000
 subnets to totally different customers
 overlapping subnets but the MPLS

0:23:22.060000 --> 0:23:27.220000
 VPN this combination of BGP and labeling
 and VRF's the combination of

0:23:27.220000 --> 0:23:31.620000
 all this stuff working together helps these
 overlapping subnets stay separate

0:23:31.620000 --> 0:23:35.940000
 in other words the service provider can
 know okay this 10 10 network down

0:23:35.940000 --> 0:23:40.200000
 here belongs to customer B there's
 some coupling that takes place with

0:23:40.200000 --> 0:23:46.000000
 MPLS labels this label along with this
 network means this belongs to customer

0:23:46.000000 --> 0:23:51.580000
 B this other label completely different
 number with the same network belongs

0:23:51.580000 --> 0:23:56.980000
 to customer A and that's how an MPLS
 VPN sort of is pieced together and

0:23:56.980000 --> 0:24:00.600000
 provides this product the security now
 one last thing I want to mention

0:24:00.600000 --> 0:24:03.680000
 before I leave this a lot of times when
 people think of virtual private

0:24:03.680000 --> 0:24:07.360000
 networks the first thing that comes
 to their mind is oh we're talking

0:24:07.360000 --> 0:24:12.560000
 encryption we're talking security MPLS
 VPNs by themselves have nothing

0:24:12.560000 --> 0:24:18.980000
 to do with encryption the privacy in
 an MPLS VPN is just the fact that

0:24:18.980000 --> 0:24:23.460000
 you know that your your traffic customer
 A is only going to stay between

0:24:23.460000 --> 0:24:27.060000
 your customer A sites you don't have
 to worry about your traffic ever

0:24:27.060000 --> 0:24:32.020000
 going to another customer sites that's
 where you get the privacy now you

0:24:32.020000 --> 0:24:36.520000
 certainly can run some sort of encryption
 on top of MPLS VPNs to add an

0:24:36.520000 --> 0:24:41.760000
 even added layer of security to encrypt
 your traffic but an MPLS VPN just

0:24:41.760000 --> 0:24:45.960000
 by itself and how it works has nothing
 to do with encryption it will not

0:24:45.960000 --> 0:24:52.840000
 do encryption so that concludes this
 section on an overview of MPLS.

