DR and BDR

OSPF can cause a lot of traffic on our networks, and sometimes that can be problematic. But OSPF has a way to deal with that on multi-access networks: it has a way to elect what's called a designated router and communicate directly to that designated router.

Point-to-Point and Multi-Access Networks

So what is the difference between a point-to-point network and a multi-access network? When we're talking about the difference here, we're really talking about the layer 2 protocol, what is connecting these devices. But just to step back to a little more simplistic view of this, point-to-point protocol is two devices that are talking directly with each other. A multi-access network is going to be a network that has at least the potential of many different devices connecting to the same layer 2 network. So something like Ethernet would be an example of a multi-access network, whereas HDLC and Point-to-Point Protocol are point-to-point type protocols. So that is the difference between point-to-point and multi-access.

The Problem with Multi-Access Networks

Here I have four different networks. We're going to say that these are all multi-access networks; let's say they're all Ethernet. Although this one looks like it's a point-to-point, it is an Ethernet, so this router that's connected to it doesn't know it's connected to one device, two devices, four devices, eight devices. It's not aware of that. So these are all multi-access networks.

So what is the issue here? Let's take a look at how many neighborships need to be created with this. For this first one right here, there is one router connecting to one router; that is just one relationship that's created there. And this network right here, we have one, two, three, so there's three connections. So with three routers we have three connections. With this network right here we have one, two, three, four, five, six, so this network right here has six different neighborships that it creates, six different communications that it needs to establish with it. And then this network right here, we see that there is one, two, three, four, five, six, seven, eight, nine, ten, so ten different connections.

So even though we're only adding one router each time, the number of connections multiplies. It's exponential. There's actually an equation for this: n times n minus one divided by two equals the number of connections we have to have. So in this example right here we have two, so there's two times two minus one, which is one, times two is two, divided by two equals one. Or if we have this 5 right here, it's five and five minus one, which is four, so five times four is twenty, divided by two equals ten. So this is the equation right here that we have.

Now let's get to the actual issue. Every time we create another relationship, that's another device that we have to communicate out to. So if we have to communicate every time there's a change, we have to communicate 10 times, and that can be really problematic. It can really swamp your network, it can really overload your network. A change could happen and it could flood your network with all sorts of communications to combat that change, and no other communication is happening during that time. So that is the issue here.

The solution is that we can elect a single device that's going to be a designated router, and then all of your devices can talk directly to that designated router. Now the number of connections is going to be just how many devices you have minus one. So in this example right here we've got one, two, three, four, so we've just got four connections that need to be maintained, and that makes these bigger networks much more reliable and able to function better.

DR, BDR and DROTHER

So we have the designated router that allows us to establish these communications, but we actually also have a backup designated router. So let's talk about these different designations. A designated router we would just label with a DR here, and this device is the go-to. We also have a backup designated router, so if something were to happen with this designated router we don't have to really do much; we already know who we're going to start communicating to instead. And so that is the backup designated router. We do have one other designator, and that is the DROTHER. The DROTHER is going to be any other device on here, any of the other routers that are going to be communicated on here. So those are the different designators on this network.

How the DR Solves the Problem

So let's see how that actually is going to solve our problem. We have a designated router, a backup designated router and three DROTHERs on this network right here. Let's say there's a change to this DROTHER right here. What it's going to do is send out a communication to the DR. In fact, it's going to be a multicast address, so it's going to be the multicast address of 224.0.0.6, and the DR, the designated router, is listening to that multicast address and is going to get that update. So this is a link state update that it's getting, and it's sent to this DR. Then it's going to update the database that it has, and then it's going to send out a multicast address to 224.0.0.5, and all of the devices are listening for this 224.0.0.5, and so then they make the changes to all their databases. To be specific, it's the link state database that they're changing.

So this is how it eliminates a huge amount of communication. When this change happens on this device right here, it doesn't need to send out a communication to each of the devices. It just sends it to this DR device, it updates its database, and then lets everyone else know of the update. The BDR's job is just to stand by and listen and update its database, and if for some reason the DR becomes unresponsive, then this backup designated router will step in and take over for the process.

The Election Process

How does a DR become a DR, and the BDR become a backup DR? What is the election process? It boils down to three different things. First of all, it's really who gets there first: who has OSPF turned on first, who has the router start up first, whatever is communicating on this network first. Let's say this DR gets started up and goes through this whole process before any of the other devices have OSPF turned on; then it's going to become the designated router for this device, and as soon as these devices turn on and start communicating they'll recognize that. So that is the first selection process.

But let's say, and most of your test questions are going to rely around this, if all these devices are set up the way they are, how is it going to select the DR? The answer to that: first of all it chooses the highest priority interface. We can actually set the priority of each of these interfaces on here to a certain priority, and whoever has the highest priority, that's going to be communicated out with its communication, and whoever has the highest priority interface is going to get designated as the designated router.

But let's say we don't set the priority on that. Well, the default priority is one, and if all of these are set to the default priority of one, then that's not going to be a good selection process. Then it goes to the highest router ID. There's quite a bit to the router ID; it actually can be several different things in there. But just realize that each one of these devices has a router ID, and it carries that router ID, which actually carries out a few functions, so that is going to be an important part of this whole process.

Now, a couple of things to realize about this: once this is established, it's not going to actively change. So let's say that this device right here comes online, and all the rest of these have the default priority interface of one, so all of these have one. This comes online and it's at a hundred, and so it gets on there. Well, it should be the designated router, right? Well, no. It's not going to actively change who is the DR and who isn't, who's the backup designated router. It's not going to actively change all of that. So when this device comes online, it's going to become a DROTHER and is not going to try to fight for the DR.

So when does this become a designated router, if it is set to 100 here and all of these other interfaces are set to one by default? If this is the designated router right here and this is the backup designated router, let's say that the designated router goes down, so this designated router goes offline. It could be the whole hardware gets shut down, or something goes wrong on the interface and the interface shuts down, or OSPF just turns off on this device altogether. There could be an interruption with OSPF, the interface, or the router itself, so it could be a few different disruptions here. So in this scenario right here, what's going to happen is this backup designated router will now become the designated router, and so now this one has the master database that it's keeping track of and is going to do all of the communication. But we do have the role of a backup designated router now open, and someone needs to take that spot, and so that's where it goes through an election, and then this becomes the backup designated router. So that's where that priority is going to come into play, when there is a re-election process that happens.

Priority

So let's take a closer look at that priority. We said that we set that priority on the interface, and we've already said the default priority for that is one. So the default is one. We can set it to be something different, like we could set it to be zero. If we set it to be zero, that means that this device will not become the designated router; it doesn't have the opportunity to become the designated router or the backup designated router. The number that we can set is between 0 and 255, and the higher it is, the better chance, the more priority it has over the other devices if the devices are equal. So let's say we set this to 100 and this to 100, and a new election process takes place. Then it goes on to our next selection process of, okay, what is the highest router ID?

The Router ID

Now let's talk about the router ID. The router ID is used for a couple of purposes. One we've already talked about: if a DR needs to be selected, it will look at the highest priority interface, who has the highest priority on that interface, and if that's not set or if they're similar, then it will go to who has the highest router ID. So a router has to have a router ID. It's not just used for DR election; it's also used to verify who people are as neighbors, and so communication happens back and forth using this router ID.

So how does a router ID get set? Well, first of all we can statically set it on it, and that is to go into the OSPF process and define what the router ID is going to be. Or it has to create it, and what it would do is take a look first of all to see if you have any loopback interfaces. If you do, it would choose the lowest numbered loopback interface, so if you have loopback zero that would be the lowest, or loopback one would be lower than loopback two. So it takes a look at that lowest loopback interface and then it takes on that IP address. Which is interesting: notice that I say the IP address, because the router ID actually looks like an IP address. It's the same format, it's 32 bits long, and when we set it statically we set it as if it looked like an IP address. So it takes the IP address of the lowest numbered loopback interface. If that doesn't exist, if there's no loopback interfaces, then it takes a look at the lowest interface number and takes on that IP address. So that's how it comes up with this.

Point-to-Point Connections

So this issue that we're talking about, which is solved by a designated router and backup designated router, really becomes more prevalent, really becomes more crucial, the bigger the network is. That's why your point-to-point networks don't really need to elect a designated router; they don't need to elect a backup designated router, so they don't go through this whole process. So this process is really for the multi-access networks.

But there are times when our multi-access networks are just two devices, and two devices are talking back and forth. So for those we could actually get onto the interface and we can designate this as being a point-to-point. So even though it's technically not a point-to-point protocol or something that's directly from one router to another, we could go and designate it so it doesn't have to go through all of this electing a designated router and all of this process for OSPF to get up and running. A good example of that is our network has HDLC and point-to-point networks here and here, so this process would not take place there. It does take place, however, between this switch and this router, because that is an Ethernet connection; it's a multi-access network right there. So what we could do is get onto this network right here, these interfaces, and change that to a point-to-point network right there, and then we would eliminate that whole process there.

Recap

That is some of the issues with a multi-access network when it comes to OSPF. As the more connections there are and more neighbors that are established, that network can become congested with so much traffic. So to alleviate the traffic we elect a backup designated router and a designated router and send traffic to that designated router, and then that designated router disperses the information out.

So how does a designated router get elected? Number one, just who is first on the link. But after that, if there are any re-elections that happen, then what it will do is first of all choose the one that has the interface with the highest priority level, and then if that doesn't exist or if it's the same across the board, if they're all set to default or they happen to be the same, then it will look at the router ID. Now the router ID is set based off of a few different things. It's whatever is the highest router ID, and it's based off of, first of all, if you've statically set it, that will be the router ID. If it's not statically set, then it will look at the lowest loopback number and see the IP address. And then if that's not set, if there is no loopback address, then it'll look at the lowest interface and use the IP address of the lowest interface. So that's how it comes up with the router ID.