8. Equal Cost Multi Path
In this lecture, you’ll learn about equal-cost multi-path ECMP. If multiple paths to a destination network have an equal metric, the router will enter all of those paths into the routing table, and it will load balance the outbound traffic to that destination network over those available multiple equal-cost paths and all routing protocols. All your IGPs will perform the equal-cost melted path by default. So Rip will do it. ISIS, EIGRP, and OSPF will all do equal-cost multipath by default. Meaning, if they learn multiple paths to the same destination, and all of them have the same cost and are the best path, it will put them all in the routing table and load balance across all of them. So they can all do equal-cost multi path.
EIGRP is the only routing protocol that can do unequal-cost multipath, so they’ll all do equal-cost multipath by default, including EIGRP. EIGRP will not do unequal cost load balancing by default, but it can do it if you manually configure it to do so. Looking at our example here, we have R 1 on the right and R 4 on the left. An example is going to take a look at how R-4 is going to get to the 10, 10 slash, 24 network behind R one.And if you look at our example topology, all of the links have the same bandwidth. They are all 100 megabits per second, and they must pass through two 3100-megabyte links on the top path and two 3100-megabyte links on the bottom path. So all of our different routing protocols would see these two paths, top and bottom, as being equal-cost paths.
So in that case, our four will install two routes into the routing table for the 100-one network. The same will happen for the 1002 network in this example as well. Half of the traffic will take the R4 to R3 to R2 to R1 route, while the other half will take the R4 to R5 to R6 to R1 route. Load balancing can also be accomplished with static routes. To do that, just configure two different static routes, both for the exact same subnet and the same subnet mask, but put in two different next hops. Again, that will install two different routes into the routing table, and traffic will be load balanced over both of them. You can have more than two routes if you want to. If you put in three routes, the load would be balanced over all three. Again, this is something that we covered in the earlier election when we looked at load balancing for static routing.
This is the same for our IGP routing protocols as well. The traffic is not going to be load-balanced round-robin for the same flow. Meaning, if you’ve got one particular source host that’s talking to a web server, for example, its traffic for that one flow is not going to go through the first packet R 1, the second packet R 2, the third packet R 3, and so on. If it gets loaded down to R 1, all traffic for that single flow will go through the same router, R 1, for example.
However, if you have a different source host communicating with a different destination, that will be load balanced onto a different link. As a result, the same flow is always routed through the same router, but different flows are load balanced through different routers. The reason it does this is because we don’t want packets for the same flow going over different paths because that could cause the packets to arrive out of order if there’s a longer delay on one path than there is on another. and that can cause some applications to fail. As a result, a single flow always follows the same path. Okay, so that was our equal-cost multi path. In the next lecture, we’ll have a look at the lab example.
9. Equal Cost Multi Path Lab Demo
In this lab, we’ll have a look at equal-cost multipath with a lab demo. Now, to start off, I’ll use the same lab again where there are unequal cost paths. So we have the topology with our four descending to R 1. And R1 is trailed by the 10-1/24 network. And R 4 can get there either via R 3 or via R 5. And what I’ve done in the lab already is configure all the IP addresses and configure RIP on all the routers as well.
So to get to the route behind R1, R4 is going to go via R5 because it’s got a lower hop count. It’s only two hops via R five. It’s three hops via Route 3. So let’s have a look at that in the lab. And if I do a show IP route on our one, it will appear on my roots all the way back to R 4. Let’s have a look at our four going to R One. So your shown IP route, and I can seemly route for 100 1, will go through ten one three two, which is on R five.
And I’ve only got one route to get to 100-1 because it’s the best path; it’s better than the route going through R-3. Okay? So that’s what happens when we’ve got a visual app topology with unequal cost paths. What happens if I change your lap topology at equal cost? So I’m going to reorganize the laboratory. I’m going to put in R six in here, so that way it’s going to be three hops whether we go via R three or go via R five. For the time being, I’m going to leave the links on R-5 alone because they are only 10 meg in comparison to the 100 meg of fast Ethernet on the other links. And you’ll see the difference it makes whether we use rip or OSPF. So I’ll go and reconfigure the lab now. I’ll see you back here in a second. Okay, I’m back again.
And if we have a look at the old configuration, you’ll see that R Five was directly attached to R One with its fast-free-zero interface. What I’ve done is put an “R” in there. So quickly free slash zero, and R5 is now linked to R6. And I’ve changed the subnet; it’s now a 10 1 4 subnet with R 6. Its fast Ethernet 30 interface now has the IP address 100 3 2 that was previously assigned to R 5. I still have the links on Rive at 10 megabits per second. The links everywhere else are 100 megabytes. So let’s just check the differences that I made in the configuration. So if we have a look at R-5, you’ll see I’ve just changed the IP address on that link to R six. I assigned it to the subnet ten one four. Then there’s Rogue Six.
I have configured its IP address on fast zero going back to R 5, and I’ve configured rip on there. I can actually see that I forgot to do something. I also need to go to interface fastfreezero, which is connected to R 1, and it’s going to get IP address 100 three two with a 24. That’s the IP address; it used to be on five, and we used to do a no-shoot on here. Let’s just make sure there aren’t any rips running around. So I’ll do a show IP route on R Six. It may take a few moments for rip to run on here. Okay, there we go. So our rip route has appeared. So, if we go back and look at R four, it used to only have that one route to the 100-1 network behind R one. Also for the 100-2 network, which is located behind R-1. But I can see it there.
So 100-1 and 100-2 versus one route, which goes out the fast Ethernet-2 zero-internes, because that was the best route going down that bottom route via R-5. But because I’ve made those changes, now both the top and bottom paths are equal in cost. They’ve both got the same hop count, which is going to be three hops. So if I do a show IP route now, the routing protocol should have converged by now, but it hasn’t yet. I can see that I still have one route. Okay, you know, what I’ll do to speed things up is do a no router disable rip and then enable it again. So I’ll say “router rip,” and it was for network 100. Let’s see if it speeded things up a bit. I’ll do an IP route, and there’s the difference. Now I can see that for the 101 network, it’s now going to go via two parts: ten one three two, which is on R five, and ten one two, which is on R three. So it will load balance through both of those routers.
It’s also going to do it for the 100-2 network, which is also behind R-1 on the right. Okay, so that’s just configured as an equal-cost multipath for rip. Let’s see what happens if we enable OSPF on here now. So I’ll go to my text file again, and I’ll paste in a basic configuration for OSPF. I’ll do that on all of my routers. Actually, let me just put in a config t in front of here to save me typing that every time. So, I’ll copy and paste this into every router. So, R 1, R 2, R 3, R 4, R 5, and R 6, and we’ll see if it’s converged yet. So, back on R-4, I’ll do a show IP route, and I should see the rip routes replaced with OSPFroute, which has recently appeared on our force. You see, it does take the routing protocol a little bit of time to converge. And there’s the other neighbor. So if I do a Show IP Route on here now, I can see that my RIP routes have been replaced with OSPF routes.
You can see that I had two equal costpaths through R 3 and R 5 for Rip. But OSPF just puts one route in the routing table for the 1001 and 1002 networks. So why is that happening? Well, if I go back to the topology diagram, you’ll see that the links on either side of R Five are 10 meg, and the links along the top path are 100 meg. So these were equal costs when it was with Rip because it just looked at hop count. But OSPF takes bandwidth into account, so the top path is the better path here. That’s why it’s the one that was installed in the routing table. So if I want to get an equal-cost multipath here, what I can do is go on to Route 5 and make those links 100 megabits per second as well. So let’s just check which interfaces they were on.
Other displays IP interfaces briefly, but only on Fast 20 and 30. So I’ll go to Global Configuration Interface fast 20 and, just to be sure, I’ll do a do show run interface fast two slashes zero and a no. And then in my client here, I can just select the bandwidth with a right-click, and it will paste it in. So that’s going to remove that manually configured bandwidth and set it back to the default bandwidth of 100 megabits per second. So I need to do that on Fast 20 and also on Interface Fast 30, and I’ll put the command in there as well. I also need to do that on the other sides of the link as well. So I’ll go to R-4, run aconfig, and it’s set to interface-fast 20 on this site. Let’s double-check it. I’ll do a show run interface at fast 20, and I can see that I’ve also configured the lower bandwidth there. So I’ll remove that command, restore it to its default state, and also on R-6, this is on interface 0 0.
So I’ll do a downhaul run interface fast zero, and I’ll see that the lower bandwidth is also configured there. I’ll remove it there too; it’s the wrong prompt there. So I need to go to Interface Faster and I’ll try that again. And there we go; we’re done. Okay, so again, we might need to give ourselves a little bit of time to converge. But if I go into R-4 now, before it just had that one route going to the routes behind R one. If we try a Show IP route now, yeah, we can see it’s already converged already. So now, because there are two equal-cost paths to get to 10010, it can go via ten one three two, which is on five, or ten one two, which is on three, and also for ten two, which is also on one. It will also load balance traffic over those two different paths. Okay, so that was a demonstration of our equal-cost multi pathing. See you in the next class.
10. Administrative Distance
In this lecture, you’ll learn about administrative distance. Ad is used in conjunction with the metric to determine which of the available paths will be added to the routing table. A router will typically only learn routes to a particular destination from a single routing protocol. It’s not normal for an organisation to be running multiple different routing protocols. When multiple routes to a destination are learned through a routing protocol, the router will install the path or paths with the lowest metric into the routing table. And different routing protocols use different methods to calculate that metric. For example, let’s say we’re using Rep as our routing protocol and we want to get to a particular destination, and we’ve got two different paths to get there.
The first path goes from router A to B to C to D. So A to B to C to D would be a hop count of three. And we’ve got another path, which is A to B to D, which would have a hop count of two. Rip uses hop count as its metric. It’s going to put the shortest-hop count into the routing table. So in this example, AB would be preferred. Now we might have exactly the same network topology, but we’re using OSPF. And with OSPF, maybe path ABCD has a cost of 60 and path Abd has a cost of 100 because the Abd path has lower bandwidth links. Remember, OSPF takes the bandwidth into account by default, and with our metric, it’s always the lowest value that is preferred. So with OSPF, it would be the other path, ABCD, that would be used, making it into the routing table. If paths to the same destination are received from different routing protocols,
if for some reason your organisation is running multiple routing protocols and a router receives routes to the same destination from those different routing protocols, then it can’t compare their metrics to each other. For example, a Rip Hopper of five cannot be compared to an OSPF of 60. That comparison would be meaningless because the routing protocols calculate the metric in completely different ways. It’s like we’re talking foreign languages. So you can’t compare one routing protocol metric to another. The router needs to use a different method to choose when routes to the same destination are received from different routing protocols. And that’s what we use the administrative distance for. The administrative distance is a measure of how trusted a particular routing protocol is. If routes to the same destination are received via different routing protocols, the protocol with the “best,” which is the lowest ad value, wins. So with metrics, the lowest is best.
With Ad, it’s the same; the lowest number is best. And this slide here shows the default values of our popular routing protocols. Rip, with a value of 120, is the worst. Then we’ve got ISIS at one. One of the five OSPFs has an AD of 110. EIGRP is the most preferred of our IGPs. It’s got an AD of 90, and the external BGP has 20. A static route has a default address, and connected interfaces will always be preferred. They have an administrative distance of zero. So administrative distance is used to choose between multiple paths and learn via different routing protocols. Metric is used to choose between multiple paths and is learned with the same routing protocol. The administrative distance is considered first to narrow the choice down to the single best routing protocol. And then the metric is considered to choose the best paths, which will make it into the routing table. We can see what’s happening with the Show IP route command. So showing IP right now will show us what routes did make it into the routing table.
You can see from here that connected routes have an administrative distance of zero. They’re always going to be the most preferred. We’ve also got some routes in here, but we’ve learned from Rip as well. and you see the digits in the square brackets. That shows us the administrative distance and the metric distance. The first value is the administrative distance. We already know that Rip has a target of 120. The second value is the metric. So here, the first route has a hop count of one. The second route has a hop count of two. So let’s talk through an example. Let’s see, we’ve got a router, and it receives multiple routes to the 1010 Org 24 network. It receives those routes from both OSPF and Rip. So when passed to the same destination, information is received from multiple routing protocols. The administrative distance is considered first before the metric, where SPF has a better administrative distance than Rip. So the rip routes are going to be discarded. Then the router will compare the routes we’ve received via OSPF and install the one with the lowest cost into the routing table. If we receive multiple equal-cost paths, then they’ll all go into the routing table, and the router will load balance between them.
Another thing we can do is use floating static routes. What I just covered is your standard core information for an ad. This is some extra information here.If the best path to a destination is lost, for example, because a link went down, it will be removed from the routing table when the router detects that and replaced with the next-best route. Now, we might want to configure a static route as a backup for the route learned via a routing protocol. For example, maybe we’re worried the entire routing protocol is going to go down. A problem if we want to do this is that static routes have a default administrative distance of one. So they’re always going to be preferred over routes learned via an IGP. So if we’re running an IGP, we’ve got IGP routes in there, and we want to add a static route as a backup. That won’t work by default because it’s going to be preferred over the IGP rather than being the second choice. So how can we make that static route the second choice? The way we can do that is by using a static floating route. When we do this, we change the administrative distance of a static route from the default of one to having a higher ad value added—an ad value that is higher than the administrative distance of our routing protocol.
Assume we’re using OSPF and we’ve established our network topology. As you can see in the diagram, we’re running OSPF from R-4 to R-3 to R-2 to R-1. And R5 does not use OSPF. Let’s say it’s a router that does not support OSPF. So we’ll have OSPF being used with the first-choice path going along the top from R Four to R Three to R Two to R One. And we want to configure a backup static route on our four that is going to send the traffic through R5, but only if the top path is not available. So we can’t add a normal static route because then it would go through R-5 as the first choice. So we enter IP route ten01025: five dots over ten dots over one dot over three dots over two, with R Five selected as the next hop.
And then we put in an extra keyword at the end. Here we say one one five, which sets an ad of one one five, which is higher than the West PF ID of 110. So this will be used as a backup route. Now, obviously, we’d also need to add static routes coming back in the other direction along the bottom path. And we’d also need a static route going from R Five to R One. Another example is that you can also use floating static routes when you’re using static routes only. So here again, on R 4, I could have my first choice go to ten one one two, which is on R 3, or I could have my second choice for the ten one network go to ten one three two, which is on R 5.
I would do that by giving it an administrative distance of five. If I just entered both of these routes without an administrative distance, it would load balance between the two of them, but by adding the second route with an added distance of five, only the first route will make it into the routing table. But if the link from R 4 to R 3 goes down, it will be removed from the routing table, and we will use the second route instead. Now, a word of warning with this: If you do a floating static route as a backup for us PF, if any of the links from R Four to R One fail, our PF will detect it and the route will be removed from the routing table. So that will work just fine.
But with our second example, this will work if the link from R 4 to R 3 goes down, because R 4 will detect that. But if the link from our three to r-2 or r-2 to r-1 goes down, r-1 won’t know, and it will continue sending traffic along the top path. And it only gets as far as our third, where the broken link is, and then the traffic is going to fail. So this will work fine if you just have that one router and nothing else along the path. But if you do have other routers along the path, be careful with this one. You could end up causing problems. Okay, so that was our administrative distance.
11. Administrative Distance Lab Demo
In this lecture, we’ll take a look at administrative distance in the lab. So I have the standard topology here, R 1 through 2, and R 5. The IP addresses are already configured, and right now, I’m running Rip along the top path between routes 1 and R four.R Five is not configured with any routing protocols. So let’s have a look at this. So I’ll go into the command line on R One.I can do a Show IP Protocols, and you’ll see that the only protocol I’ve got running right now is Rip. And if I do a Show IP route, you can see I’ve got my rip routes in the routing protocol and in the table, and they’ve got an administrative distance of 120, which is the ad for rip.
So next up, I will configure ISIS, which is the next most preferred administrative distance. So I’ve got my config ready here, and I will paste it on the routers. Actually, you know what? You’ve seen me do this before. So I’ll just pause the video, and I will paste the ISIS configuration on each of the routers. So, I’m back, and I’ve pasted in the configuration. I’ll issue IP protocols on R1 again, and you can see I’m running Rip. And if you scroll through, you’ll see I’m also running ISIS as well. When I do a Show IP route, the ISIS routes are the ones that made it into the routing table. I can also do a show IP rip database, and you can see I’m learning rip routes as well. So, if you look at the routing table, my IS routes have an ad of 1, which is preferable to Rip’s ad of 120. So that’s why my ISIS routes are making it into the routing table. The next post-preferred IGP is OSPF. So let’s configure that on our routers.
This one doesn’t take long to enable, so you can just bear with me as I paste these in. So I’ll install it on R1, R2, R3, and R4, and we won’t configure a routing protocol on R5. Let’s get onto our one again, and I can do my Show IP Protocols command now. You see, I’m running Rip, I’m running ISIS, and I’m running OSPF as well. Now. And if I do a Show IProute, my OSPF route hasn’t converged yet. We’ll give it a second for the adjacencies to come up. This shouldn’t take too long. Okay, I’ll pause the video again. There we go. All right, I just came up with it now. I might need to give it a second on the other routers as well. Let’s have a look at the Show IP route, and I can see that it hasn’t fully converged. So I’ve still got some ISIS routes in there. However, an OSPF route has been added to the routing table. Again, it’s preferred over the ISIS route because it has a better administrative distance of 110. Okay. And the final one is EIGRP.
So let’s copy and paste this from Notepad again. I can put the same configuration on all the routers, so this will just take a second. So there’s R 1, R 2, R 3, and R 4. And if I go back to R1, EIGRP does converge very quickly, so hopefully it’ll show up by the time I get to looking at the routing table. Let’s do a show on IP protocols again. And I can see I’m running Rip, ISIS, OSPF, and EIGRP now. And if I have a look at the routing table, if I show IP route A, D is for EIGRP. It’s my EIGRP routes that are making it into the routing table because they’ve got an administrative distance of 90, which is most preferred. I’ve still got an ISIS in here for the 203 network because I didn’t include that in any of the other routing protocols. Okay, so you saw administrative distance in action. Let’s go back to the topology diagram. And what I’m going to do now is configure a static floating route. My routing protocols only run from our one to R four for the first half of the hour.
I’ve not configured a routing protocol on R five.And on our one, I’m going to configure a backup route to the 10 1 2 network behind R 4 to go through R 5 instead. So I need to create an IP route to 101 200:24 with a next hop of 100 three two.I’d also need a static route from R4—sorry, R5—to R4, as well as static routes returning in the opposite direction. But I only need to do it on our one to demonstrate the floating static route in action. So let’s do that. Now, back on R 1, I’ll go to configure, and then I’ll create an IP route for 10 1 20 255 255-2550, with the next hop on R 5, which is 100 3 2. Now, I’m not going to start with a static floating route. I’ll just do it as a normal static route. And when I enter this, you can see that prior to this, my route to the 10 1 2 network was learned via EIGRP, and the next top was 100 2, which is on R 2. If I do show the IP route now, you’ll see that my route to 101 and 2 has been replaced with the static route because it’s got an administrative distance of 1, which is better.
But for our example, I only want this to be a backup. I don’t want it to be the preferred route. So what I need to do is remove that route. So I’ll hit the up arrow twice, then control A to go back to the start of the line, and then enter no to remove it. And then I’ll put the same command in again, but this time I will give it an administrative distance of 95, which is higher and therefore worse than the administrative distance of 90 in EIGRP. So when I enter that, the EIGRP route should be put back into the routing table again.
So do a show IP route to see if it’s in there yet. And there we go: the EIGRP route is back in the routing table again. I don’t see the static route because it is not the best route. It doesn’t make it into the routing table, but it is there as a backup though.So let’s look at the topology diagram again. And the interface to get out to R2 is fast Ethernet 0. So if I shut that down, it should fail over to my backup route. So let’s try that. So I’ll go configure and interface with Fast Zero and do a shutdown here, and EIGRP should detect that. You see, a JRP detected that the neighbour went down, and if I do a Show IP route now, I’ll see that my static route has made it into the routing table, so my backup worked. Okay, that was administrative distance. See you next time.
12. Loopback Interfaces
In this lecture, you’ll learn about loopback interfaces. The loopback interface is a logical interface, and it allows you to assign an IP address to a router or a layer 3 switch that is not tied to a physical interface. Because they don’t have any physical attributes, they can fail. Loopback interfaces never go down. Loopbacks are logical, so it’s impossible for them to physically be in the same subnet as other devices. So they’re usually assigned a 32-bit subnet mask to avoid wasting IP addresses. That’s the standard. It’s best practice to assign a loopback interface to all of your routers and all of your layer 3 switches. The loopback is commonly used for traffic that terminates on the router itself. That could be the most commonly managed traffic. You can also use it for other purposes, such as sending voice over IP traffic to the router. Also useful for BGP peering, etc. That provides redundancy if there are multiple paths to the router. You’ll see how that works in a second.
The loopback is also used to identify the router in OSPF because the loopback address is used as the router ID. So you’ll see this when we do the OSPF section. When you’re looking at the OSPF database, et cetera, you’ll see routers being identified by the router ID, which is the highest loopback address on that router. The same loopback interface is usually used for multiple tasks. For example, if you need to send traffic to the router for management and for BGP, we’ll usually just have one loopback interface and use the same IP address for everything. Multiple loopbacks can be configured, but this is uncommon. We don’t normally do that. It’s usually only done for special use cases where an additional loopback is required. So not typical. Okay, so here you’ll see an example of why we’re using a loopback.
We’ve got a PC, let’s say, that is behind router RF4, and it’s got an IP address of 10:12:10, and we want to connect to the R1 router to manage it. Well, we’ve got two paths that we can take to get to R One via R Four. From here, we can either go along the top path or we can go along the bottom path. But if the top path goes down, we can’t connect to the 100-one IP address, and ours is down. If the bottom path collapses, we will be unable to connect to 100,31. So we’re going to use a loopback, so that way we get a single IP address that we can use to connect to Route One, even if one of those paths goes down. So what we do is add interface and loopback zero, and we give the IP address in our example, 192-168-1132. You can use any IP address you want for the loopback, and we then advertise that in the routing protocol.
R Four will then learn the two paths that it can use to get to 1 9 2 1 6 8 1, and it will use whichever one has the lowest cost or both if they’re equal costs.And our four computers can still connect to each other even if either path goes down. So this is useful for management, and it’s really critical for other things like BGP and IP to Ethanol. Assume we’re sending IP to eaphone traffic from R 4 to R 1. We want to make sure that it’s always going to get there, even if one of the paths goes down. So we don’t direct it at a physical address in R1, which can go down, we direct it at the logical address, and that way, even if one path goes down, the traffic is still going to get there across the other path. So that’s why we use loopbacks. So let’s configure this in the lab. It’s going to be a quick and easy lab;
I will do it right now. So I’m going to assign R1 the loopback interface (19211) IP address 32 and see how our four hands got the two paths to work. Okay, so I’m in the lab, I’m on R-4 here, and I’ll just check that I’ve got EIGRP running everywhere. So I’ll run a Show IP route and see that I have an EIGRP route going out fast Ethernet20, which is via R Five, as well as other EIGRP routes going out fast, which is our three. So I want to have two different paths available and one IP address that I can use across those two different paths to get to R 1. So I appear on R One and perform a show. IP.Interface brief. I’ve just got my physical interfaces configured there right now. So I’ll go to Global Configuration and then, to create a loopback interface, the command is just InterfaceLoopback and then the number you want to use that creates the interface as well as taking you to the configuration mode for the interface.
And notice that the interface goes up immediately because it is a loop back. There’s no need to do a no-shutdown. If I do a no shutdown, it won’t do any harm, so that’s okay. Okay, I need to configure the IP address here. So I’m going to give it the IP address 19216eight, dot one, dot one, and I’m going to use the best practise of using a slash 32 subnetmask, so that’s two 5525-525-5255. So that is how my loopback interface is configured. And given an IP address, I also need to make sure that it’s being advertised in my routing protocol. So I’ll do a do show run, and for this section, EIGRP, I can see I’m using EIGRP 100, and it’s just network100 that’s included in there right now. So I need to include my loopback address as well.
So I’ll go to router EIGRP 100 and then network 192.168.1.1, and it’s a wildcard mask that’s the inverse of the subnet mask. So type 0 0 and press Enter. And now if I go back over to R for EIGRP, it converges pretty quickly. So let’s see if the route is there yet. It’s there already. I’ve got a route going to 019-21-6811. There are two paths, but one of them has a better cost. As a result, Fast Etherneteropath has been added to the routing table. Right now I can pay 192, 168, and 1, and that works. And if I trace to one nine two, one six eighteen, I can see that it’s going along the toppath with the next hop of ten one two.
And that was our interface, fast Ethernet 0. So let’s double-check that. I can fail over and still get to the loop back.So I’ll go to configuration and then interface fast 0. and I’m going to shut down that interface. So that first path is not going to be available anymore. I see my EIGRP adjacency going down. And now if I do a “Show IP route,” so before the route to the loopback was going via Fast Efrontzerozero, if I do a “Show IP route,” I can see that it’s in the routing table from EIGRP again, and now it’s failed over to the other path. It’s going to use Fast Ethernet 20, so I can still ping 192-1681 one.And if I trace the route to it, I’m going to see it going down the bottom path via R five. Okay, so that’s why we want to use loopbacks, so that we can get to our routers no matter what path we’ve got available. still using that same IP address. Okay, that was it, box. See you in the next class.
13. Adjacencies and Passive Interfaces
In this lecture, you’ll learn about adjacencies and passive interfaces, our interior gateway routing protocols. So Rep, EIGRP, and OSPF are configured under global configuration, and then they’re either enabled or not on the routers. Individual interfaces. When the routing protocol is enabled on the interface, the router will look for other devices on that directly connected link that are also running the routing protocol in order to peer with them. The router does this by sending out and listening for hello packets for that particular routing protocol. And when a matching peer is found, the routers form an adjacency with each other and exchange routing updates with each other. Modern routing protocols use multicast for the hello packets.
This is more efficient than broadcast, which was used by earlier protocols like Rip version one with multicast, and it’s specific to the particular routing protocol. So a device is only going to process that packet if it’s interested in forming adjacency with that routing protocol. And like broadcast traffic, that has to be processed by all hosts, so it’s more efficient. Okay, an adjacency example. So here we’ve got Router R1 in the middle and RA, RB, and RC. And on R1, we’ve got a loopback configuration there with an IP address of 192-168-1132. The IP subnets configured on the routing protocol-enabled interfaces will, of course, be included in its routing protocol updates. For example, our one here has a routing protocol enabled on the loopback zero interface and interfaces fastfinite zero and 10, but it’s not enabled on fast 20.
The reason we’ve done that is that RC belongs to a partner organization, and we need connectivity with them but don’t want to be sending internal network information to them. That would be a security issue. R One will send out and listen for hello packets on the loopback zero interface and Fastifferent zero and 10 because those are the interfaces that we enabled the routing protocol on, and it will form adjacencies with any routers that are running that same routing protocol that it finds on those links. So in this case, we’ve also enabled the routing protocol on RA and on RB on the interfaces that are facing R One.
So the routers will discover each other through hello packets, and they will then establish adjacency and share routing updates. But R1 will not send out or listen for hello packets on Fast 20 because we didn’t enable the routing protocol on that interface. So it will not form an adjacency with RC. It’s not going to be giving out any network information to RC. So in the example here where RC is a partner, we need to have connectivity to them so that we’re not going to be giving them internal information. In that case, we could use static routes between us and them just to give very limited connectivity.
As an example, because it formed adjacencies with RA and RB, R1 will advertise its IP subnets to them. So it will advertise with ten 00:24 subnets, 100:10, and 192-168:1132 as its loopback, but it will not advertise with 100:200:24 because that interface was not included in the routing protocol. When you enable a routing protocol globally and then enable it on an interface, the router will try to form an adjacent network to you on that interface by sending out hello packets, and it will also advertise the subnet that is on that interface as well. But if an interface is not included in the routing protocol, then the router won’t send hello packets there and also won’t advertise the subnet configured on that link to other routers either. So, in this scenario, we will not send information to RC, but RA and RB will not learn routes to 100 and 200:24 because we did not include them in the routing protocol.
So what if we do actually need RA and RB to learn a route to get to 100 and 2? That’s where passive interfaces come in. Passive interfaces allow you to include an IP subnet in the routing protocol without sending updates out of the interface. So if fast two slashes zero is configured as a passive interface, RB will learn routes to tend to, but internal network information will not be sent out to RC. So that’s what we wanted to do. In this situation, it’s best practise to always configure your loopback interfaces as passive interfaces always.So this has nothing to do with giving out network information about them. This is because it’s impossible to maintain adjacency on a loopback interface. It’s impossible for another router to be directly connected to the loopback interface because it’s not a physical interface. It’s just logical, so there’s no way we’re going to ever form an adjacency on a loopback.
Making the loopback passive means that it will be advertised through the routing protocol. We want to do that. We want other routers to learn how to join the loop, but we don’t want to waste time sending out and listening for hello packets when we know there will never be another router connected to that link. So this makes things more efficient. Always make your loop return to your passive interface. So the use cases for passive interfaces are, to summarize, that we use them on our loopback interfaces and also on physical interfaces where the device on the other side belongs to another organization, or maybe it’s not another organization but the device on the other side. We don’t want to send routing information out to it, but we do want our other internal devices to know about that link. Okay, so that was the theory for our passive interfaces. Let’s actually configure them in the lab. We’ll do that in the next lecture.
14. Adjacencies and Passive Interfaces Lab Demo
You’ll learn how to configure passive interfaces through a lab demo. And in our example topology here, we’ve got routers A or B, R Three, R Four, and R Five, and they’re all internal to our organization. And we’ve got R6, which is a router, which is at a partner. So we want to have connectivity with that partner company, but we don’t want to send them information about our internal networking. So we want to make sure that we don’t start peering, forming an adjacency with Six, and sending them that information. We are also going to have a loopback interface on R1, and it’s best practice to configure your loopback interfaces as passive interfaces.
You’re never going to have another router on the other side of that link you can’t have because it’s not a physical link. So we’ll configure that as a passive interface as well. I’ve already configured routing on all of my other routers within the company. So R2, R3, R4, and R-V-I haven’t done anything yet on our one. So if I jump on R-4, I should see all the internal routes but no route to get to the 100-1 network behind R One. So let’s just verify that. So I’m going to R-4 and I’ll do the Show IP protocols, and I’m running rip here for the lab demo and a Show IP route. And I can see that I’ve got rip routes going everywhere, but I don’t have a route to that network that was behind R One. And the reason for this is that when I go on one and do a Show IP Protocols on here, I don’t see any routing protocols configured.
Yeah, okay, first thing, let’s configure the loop back. So, 19216 eight, dot one, dot one. So if I do a show IP interface brief, I can see the loopback isn’t there yet. So I’ll go to configuration and then interface loopback, 192.168.1. And it’s a loopback, so I’ll give it a hash of 32, which is 255-25-5255, dot two, five five. The next thing I want to do is get this router to start sharing routes and learning routes from my other internal routers. So I will go to Global Confidential, then I’ll say “Router rip.” This will be version two, and the auto summary will go over the rip configuration in greater detail later. And then I want to say passive interface loopback zero, and the other passive interface is going to be the interface facing the partner company. Let’s check which one was so fast: 20. I don’t want to form a connection with R Six and provide them with routing information. So let’s get back to the router. It was also a passive interface with a fast 20.
Now, right now, I haven’t put in my network statements, so I haven’t enabled rip on the interfaces yet. I’ve enabled it globally, but nothing’s going to happen until I specify the interfaces that I want this to be enabled on. So let’s have a look at the network topology again. And all my internal interfaces are in the 10x network, so I’ll add that. I also need to add 192.168.1 for my loopback. So let’s do that. Network 192-1681 doesn’t really do anything, and neither does network 100. And that should be my configuration now. So I might just need to give Rip a second to do its thing. While we’re waiting for Rip to converge, let’s check what that subnet was again. So ten dot o dot one is behind our one and also tends to o dot to so ten dot o dot two. Let’s see if R Four picked up on those networks. So I’ll go into our four and run a Show Pirouette now, and the results were 100-1 and 100-2. And there you go.
It has obviously learned those networks. So you can see why I had to make fast-2 slash zero a passive interface. I don’t want to make a normal interface. I don’t want to send information to the partner company, but if I just didn’t include the interface in my configuration, then my internal routers wouldn’t learn about the route to get there. So that’s why I created it as a passive interface. I can also see I’ve got the route there to get to my loop back behind R1 as well. So that’s great. All of my other internal routers, including the loop back in the passive interface, are learning all of the routes everywhere. And let’s have a look at the partner router, which was R Six. So if I go on there, you’ll see that I’ve actually configured Rip on this router already and specified network 10, so it’s going to try to peer with our one. But if I do a Show IP route, you’ll see that it hasn’t learned any of the routes to the internal networks because I marked it as a passive interface on R One. So R1 is not sending it any information. Okay, that was it. That was our passive interfaces—why we have them and how we can get to them. See you in the next class.