Before we jump in and configure this whole network our iBGP peers and eBGP peers, we’ve got to learn one more thing and that’s the loopback interface. Loopback interface is a logical interface on your router that can respond on all your interfaces. Let’s say we have a situation where we got our iBGP peer here, we’ve configure it to be neighbors on this IP address which is this interface. Well if this interface goes down then everything is broken so these peers can no longer talk to each other so we’ve got a single point of failure. What we can do is we can use the loopback interface to configure BGP peers. Most of the time were gonna use it to configure iBGP peers but we can also use to it configure eBGP peers if we work with the administrators and the other autonomous system. Let me show you, let’s say we have another link here so we’ve got, we done ways to get from this router to this router and go this way or I can go this way.
Now, if I use the loopback interface, let’s say for this router its 220.127.116.11 and for this router 18.104.22.168 and were actually gonna configure this. If I’m trying to get to 22.214.171.124, my routing protocol that I’m using in here, let’s say EIGRP is gonna tell me the best route to get there. It might say go this way so I gonna go this way and if this route goes down EIGRP is gonna recalculate the route and it’s gonna say ok, this is down, you go this way so now the traffic is gonna automatically be routed this way and since the loopback interface never goes down the TCP session isn’t turn down and you don’t have to rebuild it. So that’s the major _ loopback interface. Let’s go and configure it, here I am on router 1, config T, interface and its loopback and this is where we just give it a number. Now I gonna call loopback 0, hit enter, now we can give it an IP address, let’s give it 126.96.36.199 with a subnet mask of 255.255.255.255, hit enter, exit, exit. Now watch when I ping 188.8.131.52, it works so I’m pinging myself and if you wanna see a Show interfaces you can see exactly what it looks like, its internet 00000 and here’s loopback 0 and its UP and we can see the IP address. Next let’s configure EIGRP because one of the challenges is our routers have to know how to route traffic when were trying to talk to another routers loopback address. I’m gonna turn ON EIGRP, config T, router EIGRP, let’s give an autonomous system of 100 and I’ll type in network 184.108.40.206. Remember, were just telling which interfaces we want to run on and I’m gonna type in No Auto dash (-) summary and I also want to run on network 220.127.116.11. Let’s configure router 2, config T, router EIGRP 100, No Auto dash (-) summary, network 18.104.22.168, were gonna configure the loopback on this also, hit enter, network 22.214.171.124. EIGRP is good to go and as you can see we formed an adjacency there now were gonna go back in and configure a loopback interface. I can just type in loopback interface 0 (LO 0), IP address 126.96.36.199. subnet mask of 255.255.255.255, hit enter.
BGP Synchronization is kind of a tough topic but remember when we talk about a situation where router 3 advertises a route to router 2 then router 2 advertises a route to router 1 because this router is not running BGP so it’s not an IBGP peer and then this router advertises to router 4. Well everything looks good in our Cisco Lab, but then router 4 tries to talk to this router here and the pack goes from peer to here, to here, to here and then it gets drop because this router doesn’t have the route to get to there because it’s not running BGP. Well, route synchronization are BGP synchronization prevents it because what it does, this router will not advertise a route received from an IBGP peer until it has also learn that route from an IGP or Interior Gateway Protocol such as OSPF because usually were gonna have an IGP protocol running let’s say OSPF and what we have to do is we have to redistribute the routes from BGP into OSPF. What that does is it takes that new route, puts it into our OSPF routing table and then it gets propagated out to our other routers that are also running OSPF. This routers gonna be running an IGP but its not running IBGP so its gonna get that route when this router injects it into the OSPF routing table and then propagates into here. When it propagates into here the BGP synchronization is complete because it learns it via BGP from an IBGP peer and it also learn it from OSPF. It says everything is good because now when we advertises to here and then this router tries to talk, it sends the packet here, it sends the packet here. This router now has that route to get to here via OSPF. Routes synchronization or BGP synchronization isn’t always needed and if all your routers in the path or running BGP then your good, your good to go and you can actually turn OFF BGP distribution and we gonna see how to do that. It’s a very simple command, its just no synchronization so that’s one scenario where you don’t need synchronization. The other scenario is if you don’t have a transit autonomous system. The Transit Autonomous System is when one autonomous system gets to another autonomous system through your autonomous system. Basically you have a path through here to get to here. If we only had one exit, if this fun was gone then we wouldn’t have to worry about synchronization because we’ve got one way out and one way in so then we could turn OFF synchronization.
Look I’m gonna throw a new wrinkle in and talk about iBGP. iBGP is really the same pretty much as eBGP except when were dealing with iBGP were dealing with passing these external routes within our own autonomous system. When were talking about eBGP were passing those routes to routers external to our autonomous system.
That’s the main difference, iBGP is not the same as the interior gateway protocol and that its gathering routes for your internal autonomous system, no it’s not doing that. What it is doing is it is passing these external routes.
Let’s do an example, let’s say router 3 here advertises a new route. It’s gonna advertise it to router 2, now without iBGP, router 1 wouldn’t get that new route but since we have an iBGP connection here it gonna pass it over to router 1 and then router 1 gonna pass it over to router 4 so we’ve got this whole connection here. So far, its pretty simple, now I gonna complicate it a bit, I’ll show you the true difference between iBGP and eBGP. I’ve added a new router here, one of the main purposes of iBGP is to prevent routing loops within your autonomous system.
Let’s go over a scenario here, where router 3 advertises a new route. It advertises it to router 2 then router 2 is gonna advertise it to this new router and to router 1. As long as they’re off BGP, iBGP peers and then router 1 is gonna advertise it over but if iBGP wasn’t configure correctly or it didn’t work what it suppose to, what would happen when router 2 advertise the route to this new router then this new router would advertise it to router 1, then router 1 would advertise it not only to router 4 but back to router 2, then router 2 will advertise it – see where I’m going – we’ve got a loop here and because iBGP prevents those loops what we have is router 2 is gonna advertise it to all of its iBGP peers so its gonna advertise it to this one and this one, that’s it.
Now router 1 can advertise it to eBGP peer but its not gonna advertise it to another iBGP peer ‘coz then we’ll have that loop situation. That is one of the main functions of iBGP, that’s also why your iBGP peers have to be fully meshed – what that means is they have to have a network connection so they have to get to one another. If they couldn’t then router 2 could not actually advertise that new route to router 1.
We can’t depend on router 2 advertising it to this new router then this new router advertising it to router 1 because that’s not the way it works. Every router that’s connected, let’s say this one is gone here and we just have this path to get from router 2 to router 1. Each router in this path should be running BGP, if its not were gonna have a problem and there’s a way around this were gonna go over later but I wanna show you what the problem is here. Let’s say router 3 advertises a new route this is 188.8.131.52. It advertise it to router 2, because we have network connectivity here it advertises it to router 1 and then router 1 advertises it to router 4, ok so far so good.
Now when router 4 tries to get to that, let’s say it tries to ping this address, it knows to send the packet to router 1 and then router 1 knows to send the packet to this router but this router doesn’t know what to do with it because its not running BGP, it doesn’t have the route to get here, it doesn’t know to send the packet this way. If it had a default gateway configured it doesn’t work but that’s not the proper way to hook it up because this router has two ways out of our autonomous system. It’s kind of complicated that’s why we need a fully meshed iBGP network and all of the routers that are in our path here should be running a BGP process and as we round up we gonna go over that later.
Now we’ve configured BGP, let’s take a look at what we’ve got. I gonna do show IP route, right now we really don’t have anything even though the routers peered up and BGPs its fully functioning it does have many routes to pass back and forth. But if we were hooked up to AT & T router and we were peered up, we would received the entire BGP routing table from that router so we would be fully populated here and we would get all of the routes for the internet pretty much and I believe its about a 110,000 routes right now.
We don’t have that because we are just in our Cisco Practice Lab and we need to telnet what routes to advertise and pass on to each BGP peers. In real life we would do this also but we would only have to do it with our own IP address as our own networks that are within our autonomous system because we would wanna advertise those routes out and say Hey!
Here’s how you get to my autonomous system and all of my networks. So let’s do that, let’s say AS 100 is our personal autonomous system and AS 200 is our main ISP and AS 300 is sort of a secondary ISP. What we need to do, let’s configure our ISP first, let’s configure router 3 and let’s say router 3 actually owns this subnet and this subnet here its 155.1.1 and 135.1.1. I’m gonna type in config T, router BGP and our autonomous system number, hit enter and now we gonna telnet what routes to advertise or what networks to advertise and this command gonna look familiar but its gonna have a little bit different meaning. I’m gonna type in network 184.108.40.206 then mask 255.255.255.0.
Let’s take a look at our diagram, what were doing here is we’re advertising this network and say Hey! inorder to get through this network you can go through me. This probably looks familiar to RIP or a couple other interior gateway protocols but with them when we type this in we’re actually specifying what interfaces should run in interior gateway protocol. This one is saying which networks were actually gonna advertise, hit enter. Let’s do the other one, type in network 220.127.116.11 mask 255.255.255.0, hit enter, that’s pretty much it. Now we’ve configured our “main ISP routers” so that it’s advertising its networks. Now we’ve run router 4 and let’s do a Show IP route.
Now we have something here, we got a B and B stands for BGP and we got a new route 18.104.22.168 and we see our subnet here. It’s got an administrative distance of 20 and to get to it you go through 22.214.171.124. Now we started share routes between our BGP peers. Let’s do one more example of this, I’m gonna configure router 4 to advertise this network 126.96.36.199 and once quickly configure router 2 to advertise this network because we want this network available on the public internet so if someone out here to get to it, it would be in our BGP peer routing tables.
I’m gonna type in config T, router BGP 300, hit enter, I’ll type in network, tell which network to advertise 188.8.131.52 with a subnet mask of 255.255.255.0, hit enter. Let’s go to router 2 and configure that quick, config T, router BGP 100, network 184.108.40.206 mask 255.255.255.0, now were good so we’ve got networks to advertise and our BGP peers are all setup so that’s how quick and simple overview on how EBGP is gonna work.
Now in this Cisco Lab we are going to configure router 3 and 4 pretty much the same way we configure router 2. I type in config T, hit enter, type in router, BGP and the autonomous system number that router 3 is in. In which you can see it’s in autonomous system 200, type in 200 and hit enter. So it turns ON BGP, now we need to go and configure the neighbors. Our first neighbor is 220.127.116.11, type in remote AS and we’ll take a look at the autonomous system that router 2 is in 100 so I’ll type in 100, hit enter. Now we need to configure our other neighbor, other neighbors are 18.104.22.168, the remote autonomous system of 300 and you can see we pretty got a message here say we got an adjacency change, the neighbor 22.214.171.124 is UP so we hardly form an adjacency , hit enter. Adjacency is our peer relationship so we successfully established a peer relationship and now were able to share our routing using BGP. Now, let’s go ahead and configure router 4, we gonna do the same thing, router BGP. The autonomous system number is 300 for router 4 – 126.96.36.199 with the remote autonomous system of 100 this is our router 2, they were become neighbors with and form a peer relationship. Next one is 188.8.131.52 with the remote autonomous system number of 200, as you can see we got a couple of message here letting us know our neighbors are UP. That’s pretty much it for configuring basic, basic, basic BGP, in real life its much more complex you gonna see that. This is the basics of EBGP and in a second here we gonna throw in IBGP.
When we start off configuring router 2, remember this is an entire autonomous system. All that shown here is our router that’s on the edge of our network or the border and it’s talking to other autonomous systems. Realistically we would have a lot more routers in this autonomous system that would be coming out this way, we’d have our whole network inside of here that would probably use in OSPF and some of the route between but then once it hits this router, when it needs to leave our autonomous system that’s what we gonna deal with. Just keep in mind that this is the only router in here normally but this is the one were focused on right now. Let’s configure router 2, the first thing we’ll gonna do is Enable BGP on our router, its pretty easy and it looks familiar so I gonna type in config T, and I’m gonna type in router BGP and then were gonna type in the autonomous system number. If we go back and look in our chart, we can see that router 2 is in autonomous system 100 and in real life this is gonna be handed out to by the American Registry for Internet Numbers (ARIN) and I just made that now because I’m not actually registered with ARIN but inorder for this to work because this is the routing protocol of the internet basically you have to get this autonomous system number from ARIN. That’s where the lab is ok, I’m gonna type in 100 for the autonomous system number, hit enter. The next thing were gonna do is configure the neighbors and BGP has what’s called peers. Let’s go back to our diagram, a PEER is who you’re going to share your routing information with and we’ve seen this before with BGP it’s the same but were gonna configure a little bit differently because its kind of a two way trust relationship with BGP. I need to say this guy over here is my peer and this guy needs to say this guy over here is my peer inorder for a relationship to be established. I’ve got two peers I’m working with, I’m working with router 3 here and the interface is 184.108.40.206 and also router 4 here with the interface of 220.127.116.11. To set this up, I’m gonna type neighbor and then the IP address of my neighbor 18.104.22.168, then I gonna type in remote, dash (-) AS which stands for remote autonomous system. This autonomous system number was 200, I’ll hit enter. Now, we configure our other neighbor at exactly the same way, hit enter, ok, now we’ve got our neighbor setup and remember it’s a two way trust here. I just can’t say these guys are my neighbors and everything will work, those guys have to say I’m their neighbor as well. So now were gonna configure the other routers in the next movie.
Up to this point we’ve been working with Interior Gateway Protocols such as EIGRP, OSPF or we probably seen RIP and also SS. All of those deal with Intra Autonomous System routing, routing within our autonomous system. Now were gonna focus on an Exterior Gateway Protocol called BGP or Border Gateway Protocol – it focuses pretty much exclusively on routing between different autonomous systems. These are areas – areas were different – areas were within our autonomous system. We’re going from my network over here to AT & T network over here and maybe some other ISP network over here. We don’t actually have control of the other autonomous systems but we need to communicate between them and route traffic between them and we do that with BGP. In any routing device that’s running a BGP routing process is known as the BGP speaker so this router that’s running BGP is gonna be BGP speaker and same with this router. This is a very simple design that we gonna start with and it’s gonna get more complicated. But there are two concepts were gonna have to work with and there pretty tough get initially. It’s EBGP which stands for External Border Gateway Protocol and IBGP which stands for Internal Border Gateway Protocol. They can be super confusing because when we hear IBGP and when we see it, we might start to think, ok this has to do with internal intra autonomous system routing. But it really doesn’t, it focus on exterior routing but the BGP process that goes on within your autonomous system because a lot of times were actually gonna need to pass this external routes through autonomous system and maybe out another side that goes to another autonomous system. We’ve got two different things here, we’ve got IGP – Interior Gateway Protocol, EGP – Exterior Gateway Protocol which is BGP and Exterior Gateway Protocol also includes IBGP which is Internal Border Gateway Protocol. It’s pretty confusing but as we go through it, you’ll begin to get it and get it more and then all of the sudden it will click. Initially when we start off working exclusively with EBGP – External Border Gateway Protocol because that’s probably the easiest to get. Then when we throw IBGP and let’s gonna grade a line a little bit but all of a sudden will get it so let’s go ahead and get started.
In our network we have packets a day that flying around everywhere. These packets can actually collide and when that happens, the packet needs to be resend and a collision occurs. The collision domain separates traffic into little segments on our network so that the collision domain is smaller and therefore it reduces the chance that will have a collision and have to resend the packet. What divide our collision domain is layer 2 devices. Layer 2 devices are switches and also you wanna know for the tests are bridges.
Bridges and switches are similar, bridge normally only has two ports on it and therefore pretty much absolutely now but in the real world we gonna work with switches normally. What we’ve got is a switch that remembers the MAC address of my router and any other network interface. With this computer sense a packet to my router, the traffic is routed directly from this computer to the router and a collision domain is created. We’ve got a collision domain right here from this computer to the switch and then from the switch to the router. Does the hub do the same thing?
A hub actually does not create collision domains so with this side of the network all these connections are actually one collision domain. So this side of the network we’ve got several collision domains, we’ve got one here, we’ve got one going to this computer to the router, one going from this computer to this computer so we have several collision domains, over here we just have one because a hub repeats the traffic and works on layer 1, remember that. Collision domains are separated by layer 2 devices which is switches.
Broadcast domain is a packet that sent to the broadcast address. We have to figure out the broadcast address but this 192.168.6.255 is the broadcast address. When that occurs the packet is set to every network device that is part of this subnetwork.
Broadcast it out, hits the switch and gets set to every port on the switch. A switch does not separate broadcast
domain, a router or a layer 3 device separates broadcast domain. So when this computer broadcast the messages, it hits the switch, it hits this server and its router but the router stops it right there. It doesn’t broadcast on to every other computer on the other subnetworks. We have one broadcast domain over here and another broadcast domain over here.
Things to remember, layer 2 devices stops collision domain, layer 3 devices such as routers separates broadcast domains.
Let’s talk about CSMA/CD and it stands for Career Sense Multiple Access with Collision Detection and this is how data is actually put on a wire in the internet network. What CSMA does is when a collision occurs CSMA resends that piece of data so that it actually gets through where it tries to get through. The CD part or Collision Detection is an improvement on CSMA because normally if you had a collision the two computers that were involve in the collision would resend the packet at exactly the same time and you would have another collision. You resend it again and so on and so forth so what the collision detection does is it terminates the transmission as soon as the collision is detected and waits for a run of period of time before resends that piece of data so that reduces the probability of getting another collision.
Now let’s take a look at how this actually occurs. Let’s say this computer is trying to send a packet to this router interface and this router interface is trying to send a packet to this computer at exactly the same time. What would happen is the two packets would meet in the middle and would have a collision.
Then will see a CMA the computer and the router which resends the piece of data again and you would get another collision because they would probably resends it at exactly the same time. What the collision detection does is these two devices send a packet to each other and collision occurs. This computer will wait for random period of time and this random period of time is split seconds because that’s how fast data is transferred and this device will wait a run period of time to resend the packet and therefore the packet will make it all the way through without another collision. That’s CSMA/CD.
Go for the test.
Let’s talk about hubs, switches and routers. What we have here in our diagram is we have one subnetwork and it’s the 192.168.6.0/24 subnet and will get in with what all this numbers mean later on but what we have is one subnet here and computers in a subnet connected to a switch and then the switch is connected to the router. On the other side of the router we have another subnet our 192.168.1.0/24 subnet. We have some computers on a subnet connected to a hub then the hub is connected to a router.
Now let’s talk about the function of a hub, switch or router. What a hub does is it works on layer 1 and it’s simply repeats the signal. That’s all it does.
It’s in their to get signals from the computers connected to it and it broadcast that signals out in every port it has. So if a packet is coming from 192.168.1.30 it gonna send the packet out all its port so it gonna send to this computer, this computer and back to the router. That’s all a hub does and a hub works on layer 1, remember that so just repeats 1 and 0,that’
s all it does.
A switch is more than a hub and it’s the next step. It works on layer 2. What a switch does is it remembers the MAC address of each network interface on the servers and on the router and we’ll talk about MAC address a little bit later but basically a MAC address is a hexa decimal number that is furled into every network adaptor. So it’s individual and its unique for every network adaptor in the world. So you won’t have two network adaptors on your network with the same MAC address.
So I remember just computers MAC address is so and so, so long as we fix the decimal number, you remember what this MAC addresses and this MAC address. When we route traffic on layer 2, let’s say this computer sends a frame to the switch and the switch knows which computer is suppose to go to by the MAC address and it sends it directly there. That’s what a switch
does. A router passes traffic or routes traffic in between subnets so if this computer is trying to get to an IP address that’s on different subnets, let’s say trying to talk to 192.168.1.45, this computer sends the packet to this router and then this router knows Hey! the 192.168.178 is on my other interface so its gonna route that packet over to the hub then the hub gonna broadcast it everywhere and its gonna get to 192.168.1.45. The principal role here that we wanna know is collision domains and broadcast domains.
We gonna talk about that in the next session.