- Introduction
In the last section, you learned about the fundamentals of routing. We covered our connected routes and our static routes, and we spoke about some of them as well. In this section, we’re going to start learning about our dynamic routing protocol. And before we get into each individual protocol, like Rep, EIGRP, OSPF, and BGP, we’re going to look at some of the characteristics that are common across all of the different protocols. That’s going to make it easier to understand when we get into each individual protocol later on. So that’s what we’re going to do.
In this section, we’ll talk about the different types of routing protocols are. So we’ve got our interior gateway protocols, we’ve got our exterior gateway protocols, and for our IGPs, they can be either distance vector routing protocols or linked state routing protocols. We’ll discuss what all the different types are and which protocols fit into each type. After we’ve done that, we’ll talk about how different routes make it into the routing table and how we can learn multiple routes and multiple possible paths to get to the same destination. And when that happens, it has to decide which is the best path, and that’s the one that’s going to be used. So it needs a method to decide that. You’ll learn about that in this section as well.
2. Dynamic Routing Protocols vs Static Routes
In this lecture, you’ll learn about the basics of dynamic routing protocols and how they compare to using static routes. When a routing protocol is used, routers will automatically advertise their best paths to known networks to each other. Routers will then use this information to determine their own best paths to those known destinations. And if the state of the network changes, like if a link goes up or down or if a new network is added to the topology, then the routers will automatically update each other with that information. On routers, we can use that information to automatically calculate the new best path and update their routing table in accordance with those network changes.
So let’s look at an example here. I’ve got three routers. R 1, R 2, and R 3. To the right of R 1, I’ve got the 10 124 and the 100 224 networks. R Two and R Three are not directly connected to those networks, so they’re going to find a way to find out about them. I could either use my traditional static routing, which would require me as an administrator to configure static routes everywhere, or we can have them learn it automatically through the use of dynamic routing protocols. So what we’ll do is configure a routing protocol on R1, R2, and R3, and they will then share information about their networks with each other. So we do that. We do the configuration, and R1 and R2 form a peering relationship with each other, and R2 and R3 also form an adjacency. R One will then advertise its routes to R Two.
So it informs R 2. Here, you can get to these networks via Me 100. Information will come in on the Fasteners Zero Interface on RTwo, and it will see that it came from Ronen with the IP address of 100 124, and it will then use that information to update its routing table. So the routing table will now show that it’s directly connected to the 10000:24 network that’s on Fast Ethernet Zero. It’s also directly connected to Ten100:24 on Fast Ethernet 10. And the two routes that it learned about from R 1100:10 and 100 200:24 both have a next hop of 100 One, which is on R1 under Reachable Interface Fast Ethernet Zero. R Two also has a relationship with R Three, so it will advertise information there as well. So R Two will inform R Three that you can connect to these networks using Me 100 10, 100 20, and a Slash 24 mask.
So R Two doesn’t just advertise the route it’s directly connected to; it also advertises the routes that it learned from R One as well. Again, R Three can now update its routing table. It’s got routes to Ten One, Ten Ten, and Ten 100:24, which are connected on Fast Ethernet. And the routes that I learned about are 10, 00:10, and 100, 00:24. They’re all reachable at its interface, fast Ethernet 10, and the next top address is R [email protected] dot O dot two. So again, just like with our static routes, Route 3 does not see ours as the next hop because it’s not directly connected to it. The next hop is always going to be reachable via a directly connected interface. In this example, we have two. So that was how our routes got propagated from right to left, from R1 to R2, and then on to R Three.
Obviously, the same thing will happen in the opposite direction, with R3 advertising its routes to R2, who will then advertise them on ours. So with this set up, all of the routes everywhere will be advertised everywhere, and the routers will update their routing tables with that information. Just like we could wave static routes, we can use summary routes with our dynamic routing protocols as well. So the same example, but this time two, will teach you about tendon dot dot 24 networks. Rather than advertise the fact that we can have this configured, we will send a summary route to R Three. Instead of two individual slash24s, it advertises 100 + 16.
We would do this summary routing to save memory and routers because our routing tables contain fewer routes. Obviously, in our small example, it wouldn’t make much difference, but this can make a big difference in really large networks. They also lead to less CPU usage, as changes in the network only affect other routers in the same area. To explain this, I’ll go back a slide. Assume the 100-1-1 link on R-1 goes down. When that happens, R2 will be notified that the link has failed and that an alternate path may exist. Maybe up on the top we’ve got R4, which has also got a route to get to the 100/10 network. So when a link goes down, routers that have a route to it will converge. This means that they will recalculate the routing table. They’ll try to find an alternate path that takes up fewer CPU cycles on the router. So R2 does have a route to 100100:24, so it’s going to have to do that. Our three, on the other hand, only has a route to ten 00:16. It doesn’t have that individual route to the 100-one network. So the routing table doesn’t actually change.
Because it doesn’t change, it doesn’t need to recalculate anything. So, because we used summarization, Route 3 will use less memory, have fewer routes, and because we’ve compartmentalised our network, any outages or changes will only affect that part of the network. We’re not going to be propagated everywhere, which is going to take less CPU cycles on our other routers, not in that part of the network. OK, so let’s compare our dynamic routing protocols with static routes. It should be pretty obvious that dynamic routing protocols are more scalable than administrator-defined static routes. Using only static routes everywhere is really only feasible in very small environments. With our dynamic routing protocols, routers automatically advertise available subnets to each other without the administrator having to manually enter every route on every router.
With static routing, the administrator does have to manually enter routes everywhere, which obviously is going to be very tedious and time-consuming. Also, with our dynamic routing protocols, if a subnet is added or removed, the routers will automatically discover that and update their routing tables. Static routing doesn’t do that because it’s all configured manually, and if the best path for a subnet goes down, dynamic routing protocol routers will automatically discover that and calculate a new best path if one is available. With static routes, everything is manually configured by the administrator. It’s a lot of administrative work, and it doesn’t really recover very well from any failures that are going to require additional manual administrative work. Again, so you’re probably thinking that in real-world environments we’re going to use a dynamic routing protocol. And yes, for sure, in all but the smallest or test environments, we’re always going to use a dynamic routing protocol.
But that doesn’t mean that we don’t use static routes. It is quite common to use a combination of a dynamic routing protocol and static routes. In that case, the bulk of the information is going to be carried by your dynamic routing protocol. You’re only going to use static routing for special use cases. For example, if you want to configure a backup route or a static route to the internet, Now that the default static route to the internet has been established, you’re only going to need to do that on the router that is actually connected to the service provider and all the other routers inside that. What you can do is use the edge router. You can propagate that default static route into the routing protocol, and then you can have a routing, as we call it, “carry” it through the rest of your network. So you’re not going to need to configure a default stack and all of your hours, just the one that is on the edge. OK, that’s everything I needed to tell you in this first introductory lecture. Let’s have a look at that information in the lab. We’ll do that in the next lecture.
3. Dynamic Routing Protocols Lab Demo
In this lecture, I’ll do a lab demo showing the basics of how a routing protocol works. So the lab typology: we’ve got those five routers again, R 1, R 2, and R 5, and I’ve already configured IP addresses on all of the interfaces. So the first thing that I’m going to do is configure a routing protocol. Now, the focus of this section is not on the routing protocol configuration. We’ll get into that when we cover each of the routing protocols individually in their own separate sections later on. The point of this section is to show you the commonalities—the things that work the same on all of the different routing protocols. So I’m going to use Rip, the routing information protocol, for this demo, and I’ll show you the configuration now.
So I’ve got that configured in WordPad here. You’ll see that you’ll do this a lot in the real world as well, where you have common configurations that you’re going to use on multiple routers. It’s easiest to write them down in a text editor, and then you can copy and paste from there. In the real world, you’ll also have templates that you’ll use for common router and switch configurations. And whenever you provision a new router or switch, all you have to do is take that template, change the IP addresses in there, and then any other specific settings, and then you can copy and paste that into the router or switch. It makes it a lot easier that way. It also stops you from making errors as well. So that’s what I’m going to do here. I’ve got my basic rip configuration typed out in any text editor, so I’ve got router rip enabled to enable the protocol. I’m going to run version two rather than version one. I’m not going to do automatic summary at the classical boundary, and I’m going to enable it.
And don’t worry about the configuration of any of my ten network interfaces for the time being; we’ll go over it in more detail later. So I’m going to copy that from my text editor and then go to the lab, and I’m going to paste this in on each of the routers. So I’ll do it on R One. I’ll do it on R 2 as well. On line three, I should have put “config” in the first line of a text editor, and that would have made it a little bit easier for me. But that’s okay; I can just type that in each time. So I have exactly the same configuration on all of my routers. Now I’ll go to one of the routers that was in the middle, that was R 3, and I’m going to do a debug on this one. Debugging is similar to using show commands, but unlike show commands, which provide snapshot information at a specific point in time, debug information is updated in real time. You’ll see what I mean now. So just like show commands, Debug commands get entered at the enable prompt, and I’m going to debug IP rip, so I should start seeing updates being sent and received on here. So there you go. I noticed a rip update arriving at 10:02. Let’s just check with our apologies.
So I’m currently on R3, and 10102 was an update from R2. I’m expecting to see updates from R 4 on 10 12 as well. So, because I enabled the routing protocol on all my routers, they will form adjacencies with each of their directly connected neighbors, and they will start sending updates to each other, not just about their own directly connected routes but about all the networks they’ve learned from the other routers as well. So our three should receive information about the ten-o-24 network, the ten-four network, and the tend to one-slash-24 network from R-2, as well as the tendo-324 network as well. It’s also going to be sending out updates to R2 and R4 as well. So if I jump back onto the command line, you’ll see that, yes, it is sending and receiving updates. So, if I scroll back a little bit, you’ll see that it’s been receiving updates from ten ten two, which was R two, and it’s also been receiving updates from ten one, which is R, so it’s also sending updates to those routers.
So it’s getting this information. If I now do a Show IP Route, you’ll see that the routing table is updated. Actually, let me just turn the debugging off as well to save it from updating anymore. The command to turn off all debugging is undebug all. After you’ve entered this, you might still see a few debug outputs coming in, but it will turn them off. So let’s just scroll down a bit and do a Show IP Route, and you’ll see that my routing table has now been updated. So you see the codes up at the top. Here it tells you the way that that particular network was learned, whether it’s directly connected to a local route, whether it’s a manually configured static route, or whether it was learned from a routing protocol, and which routing protocol it was learned from. So you can see the code R here.
There is rip if we look up a little higher up here, and if I can find it. So it discovered the Rip routes to the 100 o maternal one, maternal two, and 100 o maternal three networks. Also on the left are the numbers ten one two and ten one three. It is directly connected to all of the other networks. With the example here, I didn’t have any static routes set up beforehand, so the routers only knew about their directly connected interfaces. But when I enabled my routing protocol, they shared information with each other. And the routing tables are going to be updated, as are all the routers, if I jump onto ours as well and do a show IP route. Here. As you can see, it is also aware of all of the networks in our topology, whether they are directly connected or learned about them through RIP. So, that’s the fundamentals of how a routing protocol works. We’ll get a bit deeper into routing protocols, starting with the network lecture.
4. Routing Protocol Types
In this lecture, you’ll learn what the different types of routing protocols are. Our routing protocols can be split into two main types: IgPE, which are our interior gateway protocols, and GP, which is an exterior gateway protocol. Interior gateway protocols are used for routing within an organization. An exterior gateway protocol is used for routing between organizations over the internet.
There are what used to be called “legacy EGP,” but the only EGP that is in use today is BGP, the border gateway protocol. There are several different IGPs, though they can be split into two main types: distance vector routing protocols or link states. Routing protocols will explain what the different ones are and which group they fall into in a couple of slides. So in distance vector protocols, each router sends its directly connected neighbours a list of all its known networks along with its own distance to each of those networks. Distance-vector routing protocols do not advertise the entire network topology.
A router only knows its directly connected neighbors and the list of networks those neighbors have advertised. Beyond those directly connected neighbors, it lacks detailed topology information. So because of this, distance vector routing protocols are often called “routing by rumor.” We can compare this with our link-state routing protocols. In link-state routing protocols, each router describes itself and its interfaces with its directly connected neighbors. That information is then passed unchanged from one router to another. Every router thus learns the full picture of the network, including every router it interfaces with and what they connect to. Okay, so I know it’s a subtle difference and it can be a bit confusing, and I’ve seen lots of people that are actually confused with this on the internet and spreading misinformation. So let’s break it down and see what the difference is between our distance vector and our links to routing protocols. Both of them only form adjacencies with directly connected neighbors. So if you think back to our lab topology before, we had R 1, R 2, and R 3 in a row.
Our two are going to talk to each other, and R 2 and R 3 are going to talk to each other and share information with each other. But R1 does not talk to R3 directly with the routing protocol. That’s for both distance vector and linked routing protocols as well. So both of them only share information with our directly connected neighbors. The difference is that with distance vector routing protocols, those updates are from the point of view of the neighbor. The neighbour says I know about these networks, and this is what my distance is. To each of those with link-state routing protocols, the routers still talk to their neighbors, saying, “Hey, these are all of the routers and their links that are in the network,” and the information passes along unchanged. It does not get updated from the point of view of that router like it does with distance vector routing protocols. So with link-state routing protocols, the routers have a full picture of the topology, and they have a reputation for being able to make better routing decisions because of this. But like state routing protocols, they do put a bit more load on the router because there’s more information there. Okay, so let’s take a look at what the different routing protocols are and what type they fit into.
So our IGPs are using the routing protocol EIGRP, the enhanced interior gateway routing protocol OSPF, which is open source, and ISIS, which is an intermediate system. Rip and EIGRP are distance vector routing protocols, and OSPF and ISIS are linked routing protocols. And that’s all the IGPs that are in use today. EIGRP is a Cisco proprietary distance vector routing protocol, and it’s got some really good advantages that we’ll speak about later. And because it’s proprietary, Cisco would like you to use it because then it locks you into using purely Cisco devices. Now, like I said, it is a really good routing protocol, but distance vector routing protocols have a bit of an inferior reputation as compared to link-state routing protocols for the reasons that I just described.
So Cisco doesn’t want you to think of EIGRP as an inferior routing protocol, so they call it an “advanced distance vector routing protocol.” However, to be honest, it’s a distance vector routing protocol. It works the same way as Rip does. Cisco calls it “advanced” because it has more advanced metrics than Ripe. So it is a much better routing protocol than Rip. Again, as we get more into detail about the different routing protocols, I’ll explain why that is later on. Okay, so that’s our different IGPs, and as I said earlier, our one and only EGP that is used today is BGP, the border gateway protocol. In the next several sections, I’ll be covering the IGP in depth. We’ll get to BGP in a later section with our IGPs. They all do the same job, which is to advertise routes within an organisation and determine the best path or paths to the different networks.
So an organisation will typically pick one of the IGPs. They all do the same job, so there isn’t really any need to use more than one of them. Actually, using more than one of them is a bad idea. That does sometimes happen, though, usually because of a merger or maybe just historical reasons. If that does happen, then information can be redistributed between the different routing protocols. But because they work in different ways internally, this can get a bit messy. So it’s generally best to avoid that if possible. Okay, so those were our different routing protocol types. Let’s have a quick look at this in the lab in the next lecture.
5. Routing Protocol Types Lab Demo
This lecture is a lab demo where I’ll compare our distance vector and our link-state routing protocols. So it’s the same lab topology we’ve been using for the last few demos of the routers one, two, R, and five, and this is continuing on from the last demo where we already had Rip set up and configured. So let’s double-check that. So I’ll move on to ours here. Now the command to check which routing protocols are running is Show IP protocols. So I do that.
You can have multiple routing protocols running on the same router, but it’s usually not a good idea to do that here. I see that the routing protocol is RIP, and I’m currently using it. To check the configuration, I could do a show run and then scroll all the way through my configuration until I get to the rip section. But you see here, even on a router without much configuration like the one I have here, it can be a bit inconvenient to find the section. So here are a couple of shortcuts: I can do a show run, then use the pipe symbol and say Section Rip, and it will just show me that part of the running configuration. So it’s a lot more convenient. Another way you can do it is to show run, then pipe it and say, “Begin Rip.” And it will start running configuration output where it sees that rip string. So that will take me to the correct place as well. So you can see that I’ve got Rip set up here. And to see the information that was received from Rip, I can do a Show IP Rip database, and that will show me all the routes that were learned from the neighbors. So with our routing protocols, there are really three things that happen.
The routers will form an adjacency with each other. I’ll talk about that more later. They will then exchange routes with each other, which will go into the routing protocol database, and then the best routes will make it into the routing table. And I can see the database by viewing the IP Rip database. To see the routes that are best that made it into the routing table, I can do a show ip route, and I can see my rip routes in there. Okay, so if I scroll back a little, you can see the Show IP rip database. Remember, RIP was a distance vector routing protocol. It uses routing by rumor. So I just get information from neighbors as far as their point of view is concerned. Okay, next thing I’m going to do is configure OSPF here. So I’ll go to global configuration again. Don’t worry too much about the configuration yet, because we’ll cover this in more detail. When we get to the OSPF section, I’m going to say “routeropspf one,” and then “network 100.” It uses a wildcard mask, which is the inverse of a subnet mask. So that is 025-25-5255. That’s equivalent to a subnet mask of 2550, meaning I want to enable OSPF on all my interfaces, each of which has an IP address that begins with 10. I don’t care what the other octets are.
And you see, it’s giving me the error message “incomplete command” because I forgot to specify the area or see area 0. Don’t worry about the areas yet. We’ll cover that when we get to OSPF. Okay, so that was on our end. I need to do it on all of my other routers as well. So I’ll configure R 2 for router west PF 1, network 10, and 025-525-5255, and R 4 for PF 1, net 10, dot, dot, dot. I can just put in the same command in every router because we’ve all got networks that begin with “aten” and remember to put the area in as well. and just one more to do. So R five router network ten configuration. Remember in the last lab demo, I used a text editor to copy and paste? You can see the benefit of it there. It was quicker when I used the text editor. It’s more convenient. And I made a couple of typos. I did it again. I keep forgetting to put the area in. If you use a text editor, that’s another benefit—you’re not going to make any errors. You’re not going to have any typos there. Okay, so that is how OSPF is configured. I can see the OSPF adjacencies coming up. So now, if on my router here I do show IPOs in the PF database, I need to do this at the enable prompt or put the do in front. So I’ll do that. The difference is that I am seeing link information here.
So the router here, R-5, learns about all of the other routers in the area, and it learns about all of the links on those routers as well. And if I went on to R-5 and did the same command on here as well as on your IPOsPF database, it’s going to have exactly the same information. So it does not get updated from the neighbourhood point of view. All of the states of all of the links of all of the routers in the network get shared the same way between all of the routers in the area. So the routers have a more complete view of the network with our links to routing protocols. Okay, that was what I wanted to show you there. See you in the next class.
6. Routing Protocol Metrics
In this lecture, you’ll learn about routing protocol metrics. A router may receive multiple possible paths to a destination network because it might have multiple different ways to get to that network. It might have multiple neighboring routers with paths available through all of them.
Only the best path, however, out of all the ones that the router knows about, is going to make it into the routing table and be used. So the different interior gateway protocols need some way of determining which is the best path. and they all use different methods to calculate that. Each possible path will be assigned a metric value by the routing protocol, which indicates how preferred the path is, and the path with the lowest metric value is preferred. So high numbers are bad; lower numbers are better. SPF is an easy way to remember this. The name of the metric is cost.
And just like when you go shopping, the lower the cost of something, the better it is. So remember, the lower the cost or the lower the metric, the more preferred the path is going to be. Our distance-vector routers advertise to each other the networks that they know about and their metric to get to each of them. Link-state routers advertise all the links in their area to each other. Each router will take that information and then use it to make an independent calculation of its own best path to get to each destination. If the best path to a destination is lost, for example, because a link went down, it will be removed from the routing table and replaced with the next best route, the route that has the next best metric, if one is available.
So this is a big advantage of our dynamic routing protocols: their self-healing. If anything changes on the network, the routing tables will be updated to reflect that. So let’s take a look at the different methods with different metrics that are used by our different routing protocols. Starting off with Rip, the routing information protocol always uses the hop account as the metric, meaning it goes through another router. So how many routers does this router have to go through to get to the destination network? By default, the maximum hop count for Rip is 15.
You can change that, though the paths that are more than 15 hops away are marked as unreachable by default. So there’s a scalability limitation with Rip. In the example you see on this slide, we’ve got network 100. 100:24 is connected behind R1, and from R4, all the links in the top half of our 100 megabits per second links are fast ethernet. The links going down via R-5 are old, 10-megabit per second links. But with Rip, because it always uses hop count, the bottom path is going to be preferred because it’s only two hops to go from R 4 to R 5 to R 1 rather than three hops going by R 3, R 2, and R 1. So you can see a problem with Rip here.
It’s always going to use the shortest hop count, even if those links are low-bandwidth links, which would really not be the best path. So because of the scalability limitation and because it also doesn’t take the bandwidth of links into account, rip is not normally used in production networks, only in really small networks or in test environments. So let’s go through an example of how the metric is going to work in Rip. So R One has formed an adjacency with R Two, and it will say, “Hey, R Two, you can get to these networks via ME 10 100:24, which is going to cost you one hop, as are 100 2:24 and 100 300:24, which are also one hop away.” Ten one 300:24 is two hops away, and ten one two four is three hops away because the first three networks there are all directly connected on R 110. One three is behind R Five from R One’s point of view, and ten, one two, is behind R Four. So there are one or two more hops away.
So R-2 will get that information, update its routing table, and then pass the information on to our three.So R2 will say, “Hey, R3, you can get to these networks via me.” Ten to dot 24 is one hop because our two are directly connected to it: 100, 1100, two, and 1024 r. Two hops away R three will then pass the information on to R four, and it will tell R four that you can get to ten 00:24 through me. It will cost you two hops. 100 and 224 are three hops away, and ten is one hop away. That’s the information getting propagated. So, looking at the information that reached R-4 from our three, R-3 told it that you can connect to the 100:24 network through me, but it will take three hops. So that was along the top route. R One is also directly connected to R Five, and we’re running Rip everywhere.
So it will form an adjacency with R 5 as well. And it will also send updates to R Five. So it will tell R Five that you can get to 100 and 224; it’s one hop because we’re directly connected to R One and Ten, and 100:24 is two hops. So it sends that information to R Fever Five will update its routing table, and then it will pass the information on to R Four. It will inform you that R is four and that you can reach ten 00:10 and one and ten 200:24 via me. It will be two hops, with one hop also attending.
So if I just skip back a few slides here, you see along the top path, r three, r four, you can get to the 100/124 network, and it’s three hops away. R Five promotes the 100:24 network by claiming that it is only two hops away. So R Four learns two potential paths it can use to get to that 100-person network: either through R Three or R Five. Three hops will get you to R Three. R-5 is only two hops away. So it is the route via R-5 that is going to make it into the routing table. because it’s got the best metric. Both routes will be in the database, but only the best one actually makes it into the routing table, which is the one that’s got the lowest metric.
So, as I previously explained, our four were two possible routes to the 100:24 network. Three hops through a ten one two out interface for fast Ethernet zero two hops through a ten one three two out interface Fast Ethernet 20 ranks first in the routing table, followed by R5. So when we do the show IP route, we see just that one best route via R Five has made it into the routing table. Okay, so that was how Rip works with the hop count as its metric. Let’s look at the metrics for the other routing protocols as well. I don’t need to go through it step by step for these other ones. So you saw a problem there with Rep. If I go back a couple of slides, the top path in our network topology is all 100 me links, but the bottom path is only ten megabits per second links. So, while we would prefer that traffic take the top path, because Rip uses Hopkins, traffic will always take the worst bottom path. So let’s see how it works with OSPF as compared with that.
So OSPF does take bandwidth into account. It uses cost as the metric, which is automatically derived from interface bandwidth by default. You can also manually configure the cost of links if you want to manipulate the path. But OSPF is typically going to take the best path, and in our example, it is going to prefer the path from R Four to R Three to Two to R One for the 100:24 network. So unlike Rip, which used hop count and went along the bottom path, which we didn’t want, OSPF is going to use cost, which takes bandwidth into account, and it’s going to go along the top path because those links have much higher bandwidth.
So, if we have the same topology and enable OSPF on our interfaces, and then perform a show out, you’ll see that traffic will be routed along the top path. Traffic is moving at a faster interface, zero from route 4, which is the fastest path. ISIS also uses cost as the metric, but unlike with OSPF, it’s not automatically derived from the interface bandwidth. All links have an equal cost by default. So if you want to force a particular path to be used in ISIS based on which path has the best bandwidth, you’re going to have to manually configure it; it’s not going to do it automatically like OSPF does. If you don’t manually set the link cost in ISIS in our example, then the path with the lowest hop count will be used, which was the bottom path in our example again. So, once again, ISIS will use path R 4 to R 5 to R 1 by default.
Now in ISIS, you can manipulate it to go along the top path, but in Rip, you’re not really able to do that. The last IGP we have is EIGRP, and it uses the bandwidth and delay of links to calculate the metric. Load and reliability can also be configured, but they are ignored by default. So bandwidth is still being used, and there is also delay. However, the EIGRP protocol does not send probes over the links to determine the delay. It uses a fixed delay that is based on the bandwidth is.So basically, it’s based on the bandwidth. Again, you can manually configure the delay on links if you want to manipulate the path. So with OSPF and EIGRP, because it is going to use the best bandwidth links anyway, typically it’s going to choose the path that you would want it to, but if for some reason it’s not, you can override that with manual configuration. And, once again, in our example, it will use the top path because those are the higher bandwidth links; it will use that path by default.
So, last slide, let’s consider how we would choose a routing protocol. Like I said before, you do not want to be running multiple routing protocols inside your organisation because it’s going to get messy. It makes things hard to work with each other. Organizations will almost always standardise on one protocol. Really, the only reason they would have multiple different protocols is if there were a merger or some kind of historical or political reason for it. Okay, so we’re going to choose one routing protocol by comparing them. Rip uses Hopkins and has a default maximum metric of 15. As a result, it is rarely used in production networks. Because of its scalability and limitations, EIGRP is very simple to maintain, calculates changes very quickly, and its metric calculation will normally choose the best path by default. It is, however, typically only supported on Cisco routers. It was originally completely proprietary to Cisco proprietary. Cisco made moves to open it up a while ago, but it’s still mostly only supported on Cisco routers.
So it kind of forces you into using all Cisco routers. If you’re going to use EIGRP, OSPF’s metric calculation will typically choose the best path by default. Like EIGRP, it is an open standard that is supported by all vendors’ routers. And because of this, it’s the most commonly deployed IGP today. It is, however, more complicated to maintain than EIGRP. Finally, ISIS links need to be manually configured, or it will use hop counts to determine the best path, which is usually not what we want to happen. It’s typically only used in service provider networks or in large organizations with their own NPL networks, which choose it because of its scalability. Okay, so because of this, it really comes down to either EIGRP or OSPF. For most organizations, EIGRP is the simplest one to use, and it works great, but you can really only use it if you’re using only Cisco routers. OSPF also works great, but it’s more complicated to maintain. It is harder, however, and supported on all vendors’ routers. Okay, so that’s everything I need to tell you about metrics. Let’s look at how it works in the end.
7. Routing Protocol Metrics Lab Demo
This lecture, we’re going to take a look at how our IGP metrics work in the lab. I’ve got the same usual set up for the section with my five routers: one, two, three, four, and five, and IP addresses are already configured. I’ve got no routing protocols and no static routes configured yet. Let’s just verify that. So you can see at the command line that I did a Show IP interface brief and there are the IP addresses. And I can see from a Show IP route that it’s just connected routes there.
So there was no routing configuration. First up, let’s see what happens with Rip. So I’ve got the confit ready to go here. So let’s just copy this basic RIP configuration, and I’ll paste that in on every router. So we have it on R1, R2, R3, R4, and R5. And if I look back at the topology again, what we’re going to check for is that on Route, it’s route is to 10 1 20 slash 24. So when we’re using Rip, you know already that the path that it’s going to take is along the bottom path.
The reason is that Rip uses Hopkins as its metric, and it’s only two hops to the ten-two network over the path through R five. It would be three hops over the top path. So Rip doesn’t care that the top path actually has higher bandwidth. Let’s verify that at the command line. So I’ll go to R1 and do a show IP route. We should see it in the routing table now, and we’re looking for the path to the 10 1 2 network. I can see it was learned by Rip. And the next top is 100 three two, which is the bottom path on R five, and it’s sending fast Ethernet 30 out on R one. Okay, let’s see what happens if we shut down that interface. So I’ll go configure, go to Interface Fast 30, and do a shutdown.
Now when I do this, what’s going to happen is that the routing protocol is going to recon verge, meaning it’s going to see that the path has gone down and recalculate the next-best path. But this is going to take a little bit of time. So I’ve just shut down the interface, and if I do a Show IP route again, you’ll see, actually, that was really quick. We were just unlucky there. It can often take a bit longer than that. As a result, the path for ten and one has been recalculated.
It’s now going to 100 zero two, which is the top path to R on our Interface Fast Ethernet, to show you the effect of convergence. Actually, it’s 30 if I just do config T and Interface Fast; do a no shutdown on here and end. And if I run a Show IP route again, you’ll see that traffic is still being routed to Route 2 via the 100-2 interface. It will take a little bit of time before Rich sees that the better path is back up again, and then it will move back over to going through R Five.Okay, so that was our rip-hop count in action. Next up, let’s have a look at what happens with ISIS. So I’ll copy a basic ISIS config in here.
Now, ISIS isn’t really tested much on the CCNA exam because, as I previously stated, it’s not typically used in enterprise environments unless we’re very large. And they’ve got their own MPLS network. It’s more commonly used in service-provider networks. So if you do study for a service provider track later, you will get tested a lot more heavily in ISIS.
Anyway, let’s put that fact to use to save some time. I’ll put a config in here at the top to save me typing that every time. And I’m going to need to change the net address every time as well, so hopefully I’ll remember to do that. It’s fine for R-rated ones. There’s my ISIS configuration. Then I need to go back to the text file and just edit this because every router needs to have a unique net address. So R2 will be number two, R3 will be number three, and so on.
So I pace this one into our two, go back and edit the text file for our three, and copy, paste, and return to the text file. And we must complete R-4 by copying and pasting this into R-4. Then there’s just one more thing to do: configure our R-5 router. and that’s going to be number five. Okay, so I’m testing in an ISIS configuration here, and what’s going to happen is that ISIS routes are going to replace our RIP route in the routing table.
The reason for that is administrative distance, which is going to be covered in another lecture coming up really soon in this section. So I’ll do a show IP route on R1, and you can see now that my rip routes have been removed and replaced with ISIS routes. Again, we’re looking for the route to the eleven-two network, which is behind R four. And right now this is going via route 1002, which, if we had a look at the topology table, is going along the top path. Okay, so that’s not what we were expecting, right? because I didn’t configure any costs on my links. When you don’t manually set costs on your links in ISIS, it’s going to act just like rip because all links have the same cost by default. So, in essence, it operates on the basis of hop count.
But if we’re using hop count, we’d expect the route to go via R five.Right, well, what’s probably happening here is convergence again, where the R1 router will have formed an adjacency with R2 and learned the route via R Two.It hasn’t learned the route through R Five yet, but when it does, it’ll see that it’s a better cost and will go out the fast 30 interface instead. So let’s just see if it’s happened yet. So, once again, it’s going out fastzero zero with a next hop of 100 zero two. Let’s try this command again. And there we go. Ten one twenty is now outperforming free zero, as expected. So that’s a good lesson, actually. That’s just after you’ve configured a routing protocol. If you have a look at the routing table and you’re not seeing what you were expecting, just give it another minute or so and you’ll see it will probably converge, and then you’ll get what you were expecting then.
Okay, so that was the metric with ISIS. Actually, let’s just check if it fails over this routing protocol as well. So I’ll go to configuration, I’ll go to interface fast three slashes zero, and I will shut that down and then do a showiproute. And for ten, one, twenty, you can see it’s still going via 100, three, two. So it has not converged yet. Again, we need to give it a little bit of time for the routing protocol to detect that the path has gone down and then recalculate the next-best path. So I put the command in, and again, you can see it has now reconverged. And now, for 101, it is taking the second-best path, via 2 at 100. Let’s remember to bring that interface back up again. So I’ll configure tinterfacefastfree zero and perform a no shutdown.
Okay, so we’ve had a look at Rip and ISIS, which are going to take the really lowest top count for both, unless you manually set a cost on your links in ISIS. So Rip uses Hopkins as a metric, and ISIS uses Cost, but all links default to the same Cost. So next up, we’ll take a look at another routing protocol. We’ll have a look at OSPF next. So for this one, I can just copy and paste the same configuration on every router. So let’s do that on r 1 and r 2, r 3, r 4, and r 5. And again, because of the administrative distance, which we’re going to discuss in more detail in COVID later, the OSPF routes are going to replace the ISS route. So IP route, and I see that SPF has not yet converged. So let’s just try this command again. You might have to wait a second for this. Okay, and there is the message in the log that the OSPF adjacency has just come up now. So I hit the up arrow again, and it’s just to one of my routers. So I can see.
In the routing table, I have an OSPF route going out interface fast zero to R 2. So that adjacency has come up, but the adjacency R-5 has not yet come up yet.I can see that because I don’t have any OSPF routes going to 30 yet. And if I hit the upper, I think I still want to see it because I didn’t see the adjacency coming up yet. Oh, there it is. Okay, I see what the problem is. Actually, that last route, that last IS route, is for our Internet interface, which I did not include in OSPF, but I did include it in ISS. That’s why it’s showing up there. So don’t worry about that bottom row. Okay, so OSPF uses cost as its metric, and that does take bandwidth into account by default. I’m using Fast 20 Three for a quick look at our five, and I do Show IP interface brief just to check the interfaces. If I do show run fast 20, I missed the interface.
If I do show run fast 20, I can see that I set the bandwidth to ten megabits per second. The default bandwidth on Fast Ethernet is, of course, 100 megabits per second. So I’ve set my interfaces on RFive to be lower-bandwidth links. And that’s why OSPF is preferring to go along the top path rather than going through R 5. So let’s just verify that. And from the Show IP route on R 1, I can see the route for 10 1 20 is going through the top path 100 2. That’s a departure from what we saw with Rip and ISIS. The reason for this is that the OSPF does not consider bandwidth by default. If I shut the interface down, which is when it’s currently being used, then we will see that it will failover to going through our five.
So I saw that my adjacency between our two routers went down. If I now do a Show IP Route and check the traffic for the 10 1 2 network, I can see that it has failed over to go via R 5 at 100 3 2. Okay, let’s bring up the interface, which was fast zero slash zero. This time I’ll do a no-shut. And the last routing protocol to look at is EIGRP. So let’s bring our text file back up again. I will copy our basic EIGRP configuration and paste it on every router. Again, EIGRP is going to be preferred to OSPF because it’s got a better administrative distance. So it is going to replace those OSPF routes in the routing table. And all I have to do is make a copy of our five and paste it in. Okay, and that should be EIGRP running. It might just take a minute to converge.
I can see my adjacencies coming up at the command line here. Display the IP route on R1. An EIGRP like OSPF does take bandwidth into account by default, so I expect my EIGRP route to go via R two. If we look for 101, it goes through R2 at 100. And again, we can watch EIGRP failover as well. If I disable that fastzerointerface, And then just give it a second to converge so I can see that the EIGRP adjacency was updated. Although EIGRP converges quickly, I no longer show an IP E route for ten and twenty. I see that it has failed over 100 times, 302. Okay, so that was a look at our routing protocol metrics. See you in the next class.
