Basic MPLS Configuration on Cisco IOS: Step-by-Step Guide

Hello! This is Lazarus at telecomTech, where  telecom and networking technologies... are   simply explained! Today we'll be simply  explaining a basic configuration of MPLS on   Cisco IOS routers. Now don't forget to subscribe  to enable notifications and to like this video if  

you like this video. And if you have questions or  feedback please write them in the comments below,   I'm always happy to read them and to respond!  As you will see very clearly in this video MPLS   is rarely configured alone. It is almost always  deployed in combination with a whole series of  

other network protocols and features including  BGP as well as the use of one or more IGPs such   as OSPF or IS-IS. Now the topology that we'll be  working with today is this one here. Now let me   explain a few things about what we're going to  do. We have a topology that consists of three  

BGP autonomous systems. We have two customer  sites where the CE1 router at the first site   is found within BGP AS1 and the CE2 router is at  the second site and it's found within BGP AS2. The   ISP network consists of a single BGP AS with an  AS number of 100. Now all routers in the MPLS ISP  

network will have a loopback interface and an IP  address configured on those interfaces. For now,   I've only labeled the PE and the P routers that we  will be dealing with along the expected MPLS path.   But this is done for all PE and P routers in the  ISP MPLS network. We'll see how those loopbacks  

are going to be used in our configuration. Then  we'll configure an eBGP peering between CE1 and   PE1 and another eBGP peering between CE2 and  PE2. This is done so that customer networks   such as those ones here will be advertised  to the PE router. Then we'll configure the  

OSPF underlay on all of the P and PE routers. We  will then configure a full mesh of iBGP peerings   between the PE routers. And finally, once all of  that is done, then we'll enable MPLS. Well, let's   get configuring! Now the routing protocol between  the CE and PE routers can actually be anything,  

but for the purposes of this video, I've chosen  to use BGP. These will be eBGP peerings because   the routers are in different ASes. We have the  10.10.10.0/24 network between these two routers   and the IP addresses of each interface are the  .1 and .2 addresses. Now because I've covered BGP  

configurations extensively in my BGP Video series,  I will just show you the configurations briefly   here. You can check out my BGP videos in the links  in the description below for more details. This is   the configuration that essentially establishes  the eBGP peerings between CE1 and PE1. And we  

have a similar configuration at the CE2 location  between CE2 and PE2 and using the appropriate   AS numbers and IP addresses of eBGP neighbors.  Now notice here that it is the customer network   on the CE2 router that is being advertised. So  these eBGP peerings will cause the CE routers to  

advertise their networks to the PE routers. Next  we configure the P and PE routers within the ISP   infrastructure with OSPF such that all routers  within the topology can find each other and ping   each other. In this configuration, we ensure that  the loopbacks of all devices are also reachable by  

advertising them via OSPF. Now this is vital  because ultimately what we want is for LDP   adjacencies to form between MPLS enabled routers  and LDP adjacencies will always form between the   loopback addresses by default, so they should be  advertised by the IGP and reachable to each other.  

Now here's a sample of the OSPF configuration  that we would find within our PE1 router. We're   advertising the loopback address as well as the  network serving this link here. But notice that we   are not advertising the network of the interface  that connects to the CE1 router. This is where the  

OSPF domain ends. It doesn't extend into the CE  at all! Ideally we should configure that interface   as a passive interface so that no OSPF messages  like hello messages are sent to the CE router at   all. And here we see a sample OSPF configuration  found within a P router. Here all networks on all  

interfaces are advertised as well as the loopback  address. Now ultimately all IP addresses of all P   and PE routers on physical interfaces and loopback  interfaces, except those facing the customer edge   devices, should be reachable from each other. So  that's the IGP underlay configured using OSPF. Now  

that all of the P and PE routers are reachable to  each other, we can configure iBGP peerings between   all PEs. Now remember BGP here is used to exchange  customer networks only. For our topology here,   we'll be looking at the iBGP peering that will  establish between PE1 and PE2. Now to configure  

these peerings, we issue the following commands.  Within each PE, we create the peerings between the   loopbacks. That's one more reason why we created  those loopback interfaces, and why we want them   advertised by the routing protocol. Since we're  using the loopbacks, we must also use the `next  

hop self` command and the `update source` command.  If you're not sure what those are, take a look at   the related BGP videos linked in the description  below. Now the P routers remain unchanged. They   don't need to know about BGP because they won't be  participating in it. They don't need to know the  

customer networks but their OSPF configurations  allow them to convey iBGP peering messages between   the PE routers, allowing them to become iBGP  peers and exchange routes. Now I'm not sure if   you realized that after all of that configuration  we haven't even started configuring MPLS! But not  

to worry, because enabling MPLS is actually  among the simplest configuration steps in this   process. Now MPLS on Cisco devices is enabled on  a per-interface basis and this is done by issuing   the `mpls ip` command on all interfaces within the  MPLS core network. In other words, all interfaces  

except those facing customers. This essentially  enables the LDP protocol and allows these routers   to create LDP adjacencies and exchange labels.  And after a few seconds, the topology should be   converged! So here's the configuration on P2 as  well as on PE2. So now all of these interfaces  

have MPLS enabled and LDP adjacencies are forming  between the loopback interfaces. Now if we ping   from CE1's loopback address to CE2's loopback  address, we will get a successful ping. But let's   see what's actually happening under the hood.  First, let's take a look at one of our P routers.  

Here in the IP table, we don't see the source or  destination address of that ping, and yet our ping   was successful! That means that we're no longer  using IP routing at this point in the topology   but we're using label switching which is what we  want. Secondly let's take a look at our routing  

table for our destination within PE1. We can  see here that in order to get to our destination   from PE1 we need to get to the PE2 iBGP neighbor  which is indicated here by the 4.4.4.4 loopback IP   address. PE1 learned about this customer route  from its iBGP peering with PE2 and we can see  

that the next hop, in order to get there, is PE2  itself. Let's see what the LFIB tells us about how   to get to that destination. Here we can see that  the next hop IP of PE2 corresponds to label 17,   with an exit interface of GigabitEthernet0/2,  and a next hop IP of 10.10.20.2 which is that  

of this P router here. So this label is  pushed and the packet is sent. Notice   that the actual customer destination network  doesn't appear anywhere in the LFIB. It's only   the IP address of the destination PE that is seen  and the labels are used to get to that PE device.  

Now I won't continue the verification steps on the  rest of the P and PE routers in detail, because   we've seen those in previous videos, but I will  just say that label switching continues to occur   until you get to the last P router. At this point,  penultimate hop popping takes place, or PHP, where  

the label is popped or removed. Once that's done  it's sent to the PE2 router which has a route to   the destination and it is sent to the destination  using simple IP routing. So that, in a nutshell,   is how you configure and verify MPLS. Now it's  amazing how the MPLS configuration is just a very  

small part of the whole configuration procedure  because of the fact that you need routing to be   configured between CEs and PEs and an IGP to be  configured on the ISP infrastructure, and BGP to   be configured between all PE routers, and only  then are you able to enable the functionality  

of MPLS to get the results you want! Anyway, I  hope you found this video useful and if you have   please click that thumbs up button. If you have  any questions, feel free to leave a comment below   and please subscribe to get updates to newly  published videos. I'm Lazarus at telecomTech,  

I hope this has been helpful for you,  and I'd like to thank you for watching!