WEBVTT

0:00:03.180000 --> 0:00:07.120000
 Hello and welcome this video
 titled routing IP packets.

0:00:07.120000 --> 0:00:11.400000
 We're going to do a review of what is
 the process of routing, what happens

0:00:11.400000 --> 0:00:16.020000
 during routing and how do routers find
 the best route to a destination

0:00:16.020000 --> 0:00:19.780000
 if they know of more than one
 route to a destination.

0:00:19.780000 --> 0:00:20.900000
 So what is routing?

0:00:20.900000 --> 0:00:25.380000
 Well, the process of routing is really
 forwarding packets, a device that

0:00:25.380000 --> 0:00:29.580000
 has at least two knit cards, at least
 two interfaces that receives a packet

0:00:29.580000 --> 0:00:34.480000
 on one knit card that belongs to one
 network, one IP network and forwards

0:00:34.480000 --> 0:00:39.700000
 it out another knit card because it's
 destined to a different IP network.

0:00:39.700000 --> 0:00:42.480000
 Where the source and destination networks
 in the IP packet are different

0:00:42.480000 --> 0:00:46.200000
 and the router knows how to get
 the packet to where it's going.

0:00:46.200000 --> 0:00:49.220000
 So in order to do routing, a router
 needs some basic things.

0:00:49.220000 --> 0:00:52.640000
 The data that comes in
 needs to be routable.

0:00:52.640000 --> 0:00:56.560000
 It has to contain a layer three header
 like an IP before, IPV6 header

0:00:56.560000 --> 0:00:59.940000
 that has source and destination
 IP addresses in it.

0:00:59.940000 --> 0:01:05.080000
 The router itself has to have a routing
 table that contains routes that

0:01:05.080000 --> 0:01:09.700000
 contains network addresses, subnet masks
 and next hops as well as metrics

0:01:09.700000 --> 0:01:11.400000
 that go along with those.

0:01:11.400000 --> 0:01:15.280000
 If all of that is true, the
 router can do routing.

0:01:15.280000 --> 0:01:16.860000
 So some general rules of routing.

0:01:16.860000 --> 0:01:20.740000
 Number one, routers will only use
 routes with reachable next hops.

0:01:20.740000 --> 0:01:26.480000
 In the case of your interior gateway
 protocols like RIP, OSPF and EIGRP,

0:01:26.480000 --> 0:01:28.280000
 that's not going to be a problem.

0:01:28.280000 --> 0:01:34.140000
 Only when you're talking about using BGP
 might that be a significant issue.

0:01:34.140000 --> 0:01:36.460000
 ROUDERS will only use
 the best routes.

0:01:36.460000 --> 0:01:40.280000
 This comes into play if a router learns
 of a remote network like maybe

0:01:40.280000 --> 0:01:44.800000
 a router learns of the 10.10.10.0 network,
 but he learns about it via

0:01:44.800000 --> 0:01:48.380000
 two paths. He says, according to my
 routing table, I could go that way

0:01:48.380000 --> 0:01:49.700000
 or I could go that way.

0:01:49.700000 --> 0:01:53.040000
 In that case, the router is going to
 try to figure out which of these

0:01:53.040000 --> 0:01:57.300000
 routes is the best and only the best
 one will actually be installed into

0:01:57.300000 --> 0:01:59.060000
 the IP routing table.

0:01:59.060000 --> 0:02:02.340000
 Routes must be believable.

0:02:02.340000 --> 0:02:06.620000
 In other words, if you and I are running
 a routing protocol, take your

0:02:06.620000 --> 0:02:08.380000
 pick of whichever one it is.

0:02:08.380000 --> 0:02:12.380000
 If I send you a route saying, hey, neighbor,
 you can get to the 50 network

0:02:12.380000 --> 0:02:18.240000
 through me. Well, that's great, but
 if an hour passes, if a day passes,

0:02:18.240000 --> 0:02:22.860000
 if a week passes and you never hear
 about that route from me again, how

0:02:22.860000 --> 0:02:24.400000
 do you know it's still believable?


0:02:24.400000 --> 0:02:27.840000
 How do you know you can still
 use me to get to that network?

0:02:27.840000 --> 0:02:31.220000
 So every routing protocol has a different
 way of answering that question,

0:02:31.220000 --> 0:02:34.360000
 but ultimately it does
 have to be answered.

0:02:34.360000 --> 0:02:39.460000
 And routers will only accept routes that
 match their own active protocols.

0:02:39.460000 --> 0:02:45.000000
 So for example, if you and I on this
 particular wire that we share are

0:02:45.000000 --> 0:02:49.560000
 doing IPV4, I've got an IPV4 address,
 you have an IPV4 address.

0:02:49.560000 --> 0:02:55.440000
 If I enable RIP on my side and you enable
 RIP on your side, every 30 seconds,

0:02:55.440000 --> 0:02:59.400000
 I'm going to send you my routes,
 my IPV4 routes via RIP.

0:02:59.400000 --> 0:03:01.040000
 Everything's well and good.

0:03:01.040000 --> 0:03:05.900000
 Now, what if on this cable,
 I also have an IPV6 address?

0:03:05.900000 --> 0:03:08.280000
 You don't, but I do.

0:03:08.280000 --> 0:03:10.700000
 And I've enabled RIP for IPV6.

0:03:10.700000 --> 0:03:11.520000
 Well, guess what?

0:03:11.520000 --> 0:03:14.900000
 Every 30 seconds, I'm going
 to send you IPV6 RIP routes.

0:03:14.900000 --> 0:03:19.160000
 But in that case, those routes won't
 do you any good because you don't

0:03:19.160000 --> 0:03:22.860000
 speak IPV6. You don't have
 IPV6 on this interface.

0:03:22.860000 --> 0:03:25.180000
 Those routes are completely
 meaningless to you.

0:03:25.180000 --> 0:03:27.420000
 That's what that last point
 is talking about.

0:03:27.420000 --> 0:03:31.480000
 Now, the general way that routes get
 into the routing table is one of

0:03:31.480000 --> 0:03:33.000000
 three ways connected.

0:03:33.000000 --> 0:03:38.480000
 So any time you configure either an
 IPV4 or an IPV6 address and subnet

0:03:38.480000 --> 0:03:43.040000
 mask on an interface, whether it be
 a physical interface like gig 01,

0:03:43.040000 --> 0:03:51.720000
 a sub interface like gig 01.1, a logical
 interface like interface VLAN

0:03:51.720000 --> 0:03:55.840000
 one or something, any interface that
 can hold an IP address, the moment

0:03:55.840000 --> 0:04:00.820000
 you put that IP address on there and
 the interface comes up, it creates

0:04:00.820000 --> 0:04:03.580000
 a connected route in
 the IP routing table.

0:04:03.580000 --> 0:04:07.820000
 You can also statically program a route
 into the routing table with static

0:04:07.820000 --> 0:04:11.000000
 routes. And we'll get around to doing
 a refresher on how to do that.

0:04:11.000000 --> 0:04:15.220000
 And certainly, you can have dynamic
 routing protocols exchanging routes

0:04:15.220000 --> 0:04:16.520000
 among each other.

0:04:16.520000 --> 0:04:21.600000
 And so your routing table
 can be populated that way.

0:04:21.600000 --> 0:04:25.340000
 All right, so like I mentioned, when
 two or more routes are learned to

0:04:25.340000 --> 0:04:29.100000
 the same network, a router will attempt
 to identify which is best for

0:04:29.100000 --> 0:04:31.840000
 installation into the
 IP routing table.

0:04:31.840000 --> 0:04:36.260000
 And there's basically two
 ways it will do that.

0:04:36.260000 --> 0:04:40.560000
 Two factors, number one, it will check
 the administrative distance of

0:04:40.560000 --> 0:04:44.680000
 both of those routes to see if they're
 different or if they're the same.

0:04:44.680000 --> 0:04:48.640000
 If the administrative distance is different,
 so if I learn about the 1010

0:04:48.640000 --> 0:04:52.300000
 network that way, and I also learned
 about the 1010 network going that

0:04:52.300000 --> 0:04:57.120000
 way, if the methods that I learned
 them by, if the protocols that gave

0:04:57.120000 --> 0:05:01.620000
 me those routes have different administrative
 distances, now I can compare

0:05:01.620000 --> 0:05:04.500000
 the administrative distance values
 and select one as being better than

0:05:04.500000 --> 0:05:07.620000
 the other. And we're going to look
 at that in about two seconds.

0:05:07.620000 --> 0:05:11.280000
 If the administrative distance values
 are the same, then I look at the

0:05:11.280000 --> 0:05:15.300000
 routing metric, and the routing metric
 will help me decide which is best.

0:05:15.300000 --> 0:05:17.880000
 So as a refresher, what is
 administrative distance?

0:05:17.880000 --> 0:05:25.040000
 Every possible way that a router or
 a switch, any Cisco iOS device that

0:05:25.040000 --> 0:05:30.080000
 can do routing. Any possible way it
 can learn of a route, whether it be

0:05:30.080000 --> 0:05:33.640000
 connected route, static routes, or some
 sort of dynamic route, all those

0:05:33.640000 --> 0:05:37.280000
 particular methods have administrative
 distance values.

0:05:37.280000 --> 0:05:41.620000
 They are numerical values between
 the ranges of zero and 255.

0:05:41.620000 --> 0:05:46.060000
 The lower the administrative distance,
 the better or more preferable that

0:05:46.060000 --> 0:05:48.600000
 particular way is.

0:05:48.600000 --> 0:05:51.800000
 So here are the general administrative
 distance values that you will want

0:05:51.800000 --> 0:05:55.100000
 to memorize. You can see nothing
 beats a connected route.

0:05:55.100000 --> 0:05:58.340000
 If I'm sitting on this network because
 I have a wire connected to it,

0:05:58.340000 --> 0:06:02.320000
 and all of a sudden somebody advertises
 that exact same network to me,

0:06:02.320000 --> 0:06:04.920000
 they're not going to beat the
 route I have that's connected.

0:06:04.920000 --> 0:06:07.000000
 That has administrative
 distance of zero.

0:06:07.000000 --> 0:06:08.660000
 That one you cannot change.

0:06:08.660000 --> 0:06:13.220000
 There's no configuration commands
 to make that less preferable.

0:06:13.220000 --> 0:06:17.040000
 Static routes on up, all of those are
 configurable, but these are the

0:06:17.040000 --> 0:06:19.460000
 default values that
 you see right here.

0:06:19.460000 --> 0:06:25.180000
 Now, let's say I learned about a route
 to the 1010 network via this interface,

0:06:25.180000 --> 0:06:29.500000
 and a route to the 1010 network via
 this other interface, and they both

0:06:29.500000 --> 0:06:31.840000
 have the exact same administrative
 distance.

0:06:31.840000 --> 0:06:37.340000
 Maybe I learned about both of them
 via RIP or both of them via OSPF.

0:06:37.340000 --> 0:06:39.840000
 Well, if the administrative distance
 value is the same and I can't use

0:06:39.840000 --> 0:06:44.620000
 that, then I'm going to look
 at the routing metric.

0:06:44.620000 --> 0:06:50.180000
 And interior gateway routing protocols
 like RIP, OSPF, EIGRP, even ISIS

0:06:50.180000 --> 0:06:55.160000
 use metrics. And the lower the metric,
 the better route happens to be.

0:06:55.160000 --> 0:06:57.140000
 Lower is preferred.

0:06:57.140000 --> 0:07:01.240000
 So the main ones you need to know
 about are what you see right here.

0:07:01.240000 --> 0:07:03.480000
 EIGRP uses a composite metric.

0:07:03.480000 --> 0:07:06.780000
 It actually is probably the most complicated
 of the metric values of the

0:07:06.780000 --> 0:07:07.960000
 routing protocols.

0:07:07.960000 --> 0:07:11.860000
 It factors in about four or five different
 things to spit out this one

0:07:11.860000 --> 0:07:14.860000
 big number, which
 we call distance.

0:07:14.860000 --> 0:07:18.440000
 So the metric the EIGRP uses, the technical
 term for its metric is called

0:07:18.440000 --> 0:07:24.220000
 distance, but several things factor
 into the formula to make distance.

0:07:24.220000 --> 0:07:26.220000
 RIP metric is hop count.

0:07:26.220000 --> 0:07:29.540000
 How many routers do you have to hop through
 to get to that remote destination?

0:07:29.540000 --> 0:07:31.140000
 Very simple with RIP.

0:07:31.140000 --> 0:07:36.480000
 OSPF's metric is link bandwidth, which
 we call cost, very similar to spanning

0:07:36.480000 --> 0:07:40.820000
 tree. Just like in spanning tree, how
 the faster an interface is, the

0:07:40.820000 --> 0:07:44.720000
 lower the spanning tree cost will be,
 the more preferable it will be.

0:07:44.720000 --> 0:07:46.560000
 Exactly the same thing with OSPF.

0:07:46.560000 --> 0:07:50.380000
 No different. It's just that OSPF uses
 a different formula, but the same

0:07:50.380000 --> 0:07:54.180000
 holds true. The faster the bandwidth
 of an interface, the lower the OSPF

0:07:54.180000 --> 0:07:57.620000
 cost will be, which is the
 metric that OSPF uses.

0:07:57.620000 --> 0:08:03.800000
 Now more than one routing entry may
 match the destination address of a

0:08:03.800000 --> 0:08:06.460000
 packet, longest match wins.

0:08:06.460000 --> 0:08:07.540000
 What's that referring to?

0:08:07.540000 --> 0:08:11.880000
 So now we're not talking anymore about,
 oh, I've learned about the exact

0:08:11.880000 --> 0:08:15.700000
 same prefix, the exact same network
 from two or three different sources.

0:08:15.700000 --> 0:08:16.960000
 That's not what we're looking at.

0:08:16.960000 --> 0:08:22.380000
 Now we're looking at something like
 this, where here's my routing table.

0:08:22.380000 --> 0:08:33.140000
 And maybe I have a connected route
 on interface that is 10.0.0.0.8 via

0:08:33.140000 --> 0:08:37.260000
 fast Ethernet 0.1.

0:08:37.260000 --> 0:08:47.880000
 Then I learned via RIP about the 10.0
.0.0.16 via the router that advertised

0:08:47.880000 --> 0:08:50.900000
 that to me was the 1111 router.

0:08:50.900000 --> 0:08:52.780000
 Here's my RIP neighbor.

0:08:52.780000 --> 0:08:58.580000
 And I also learned via EIGRP,
 the 101010 network.

0:08:58.580000 --> 0:09:05.300000
 And that's via my EIGRP
 neighbor of 2222.

0:09:05.300000 --> 0:09:08.000000
 Okay, now these routes
 are different, right?

0:09:08.000000 --> 0:09:08.760000
 They're not all the same.

0:09:08.760000 --> 0:09:11.220000
 They have different masks in them.


0:09:11.220000 --> 0:09:14.620000
 So now here comes in
 a bunch of packets.

0:09:14.620000 --> 0:09:19.260000
 And these packets are
 all going to 10.10.0.

0:09:19.260000 --> 0:09:25.300000
 Let's say 20.1. That's
 their destination.

0:09:25.300000 --> 0:09:29.540000
 All right, of my routes in my routing
 table, well, this one here is a

0:09:29.540000 --> 0:09:34.260000
 match, because this one says, hey,
 you can match me if the first octet

0:09:34.260000 --> 0:09:36.100000
 in your destination is 10.

0:09:36.100000 --> 0:09:38.320000
 Yep, that matches these packets.

0:09:38.320000 --> 0:09:39.780000
 Oh, but hold on a second.

0:09:39.780000 --> 0:09:43.920000
 My RIP route is also a match because
 according to its subnet mask, it

0:09:43.920000 --> 0:09:48.600000
 says, hey, if the first two octets
 match 10.10, that's a match.

0:09:48.600000 --> 0:09:50.100000
 Well, that also matches this.

0:09:50.100000 --> 0:09:53.000000
 My third route, it does
 not match, right?

0:09:53.000000 --> 0:09:57.580000
 This one says, if the first three octets
 are 10.10.10, that does not match

0:09:57.580000 --> 0:10:01.060000
 this packet. So that third route will
 not be considered, but the first

0:10:01.060000 --> 0:10:04.760000
 two will. Well, the rules of routing
 say, hey, if you've got two routes

0:10:04.760000 --> 0:10:08.540000
 that could match, the
 longest one matches.

0:10:08.540000 --> 0:10:11.380000
 So this one has a 16-bit match.

0:10:11.380000 --> 0:10:13.580000
 This only has an 8-bit match.

0:10:13.580000 --> 0:10:17.480000
 So we're going to route those packets
 via the RIP route to our neighbor

0:10:17.480000 --> 0:10:29.820000
 at 1111. As a quick review, routing
 protocols, a lot of them have this

0:10:29.820000 --> 0:10:31.600000
 concept of autonomous system.

0:10:31.600000 --> 0:10:36.560000
 An autonomous system is a collection of networks
 all managed by one administrative

0:10:36.560000 --> 0:10:40.220000
 authority. So that collection of networks
 gives no indication as to its

0:10:40.220000 --> 0:10:44.420000
 size. So as I've mentioned in some
 other videos, if you have a network

0:10:44.420000 --> 0:10:49.100000
 at home consisting of a cable modem or
 DSL modem, and you've got a couple

0:10:49.100000 --> 0:10:53.260000
 laptops and a printer there and maybe
 a TV that's connected via Wi-Fi,

0:10:53.260000 --> 0:10:55.520000
 that's all under your control.

0:10:55.520000 --> 0:10:57.140000
 You control your home network.

0:10:57.140000 --> 0:10:59.900000
 So technically, that's
 an autonomous system.

0:10:59.900000 --> 0:11:02.460000
 But usually when we use this term autonomous
 system, we're talking about

0:11:02.460000 --> 0:11:08.000000
 really large networks that have like
 a central IT staff that manages what

0:11:08.000000 --> 0:11:08.960000
 that network is doing.

0:11:08.960000 --> 0:11:11.660000
 And that's what we refer to
 as an autonomous system.

0:11:11.660000 --> 0:11:16.640000
 Some routing protocols are designed to
 operate purely within an autonomous

0:11:16.640000 --> 0:11:20.580000
 system. So let's think about autonomous
 system as being like a company's

0:11:20.580000 --> 0:11:24.840000
 network. So let's think of a large
 company that has like 20 buildings,

0:11:24.840000 --> 0:11:28.260000
 hundreds of routes, hundreds
 of routers and switches.

0:11:28.260000 --> 0:11:32.520000
 That one big network in those 20 buildings
 is considered an autonomous

0:11:32.520000 --> 0:11:36.480000
 system. Well, there's some routing protocols
 that were designed to operate

0:11:36.480000 --> 0:11:40.320000
 within that space, within
 the autonomous system.

0:11:40.320000 --> 0:11:43.240000
 And those are called interior
 gateway protocols.

0:11:43.240000 --> 0:11:47.360000
 Vast majority of your routing protocols
 fall into that category.

0:11:47.360000 --> 0:11:53.780000
 RIP, EIGRP, OSPF, even ISIS, they are
 all examples of interior gateway

0:11:53.780000 --> 0:11:59.640000
 protocols. Sometimes a router will sit
 on the edge of an autonomous system

0:11:59.640000 --> 0:12:03.880000
 and it will speak to another router who's
 on the edge of a different autonomous

0:12:03.880000 --> 0:12:07.560000
 system. And those routers will need
 to exchange routes between them.

0:12:07.560000 --> 0:12:11.040000
 So we'll have one autonomous system exchanging
 routes into another autonomous

0:12:11.040000 --> 0:12:16.160000
 system. Those types of routing protocols
 are called exterior gateway protocols.

0:12:16.160000 --> 0:12:21.000000
 And as of right now and for the last 30
 years, there's only been one routing

0:12:21.000000 --> 0:12:23.160000
 protocol that fell
 into that category.

0:12:23.160000 --> 0:12:26.360000
 That was BGP, the border
 gateway protocol.

0:12:26.360000 --> 0:12:30.100000
 That's the only one that is an
 exterior gateway protocol.

0:12:30.100000 --> 0:12:36.220000
 So that pretty much concludes this
 refresher on an introduction to IP

0:12:36.220000 --> 0:12:37.800000
 routing. And its concepts.

