LDP in MPLS: Label Distribution Protocol Simply Explained

Hello! This is Lazarus at telecom.Tech,  where telecom and networking technologies...   are simply explained! Today we'll be simply  explaining Label Distribution Protocol for MPLS.  This video series takes a look at MPLS from first  principles. We start from scratch and build up to  

more complex and detailed deployment scenarios,  so whether you're new to MPLS or you just want to   brush up on your understanding of the technology,  this series is helpful for you! You can be   notified of every new video that's published  by subscribing and enabling notifications,  

and if you find this video useful, please do  click that thumbs up button! In previous videos,   we introduced MPLS as a technology that is  used to efficiently direct data between nodes,   and we examined a simplistic traversal of a packet  across an MPLS infrastructure. In this video,  

we'll be taking a more detailed look at how  MPLS routers exchange information about labels,   and how that information is used to populate  various internal constructs in an MPLS router   that make MPLS possible. First of all, when you  have an MPLS core network such as this one here,  

you must have an underlying routing protocol that  allows all of these devices to communicate with   each other. So typically you'll have something  like OSPF or IS-IS running so that these nodes   here are able to find each other and communicate  with each other. Now mind you, the purpose of this  

underlying routing is not to get to the customer  networks! It's only for the purpose of allowing   these routers to find each other and to create  the underlay network that will allow MPLS to   operate. Now this is kind of counterintuitive  because you will typically want to use MPLS  

and label switching instead of routing, so why do  we need routing? Remember MPLS label switching is   responsible for transmitting customer traffic.  This routing underlay here that we're talking   about is here to allow the control plane to  function. Now secondly, once you've established  

that underlay routing, when you have two MPLS  routers and you enable MPLS on these interfaces   here, LDP will automatically attempt to make an  LDP adjacency. MPLS is enabled on a per-interface   basis, and we'll see that in practice when we  configure MPLS routers in a future video. LDP  

adjacencies take place only between directly  connected MPLS routers. Once an adjacency like   this is established, labels begin to be exchanged.  But where are these labels stored and what is   their significance and how are they related to  the IP prefixes of our customers that we want to  

relay over the infrastructure? To answer this,  let's first take a look at the constructs used   by normal IP routing that we all have come to know  and love. When normal routing takes place, we have   various entities within the router that enable it.  We have the routing protocol such as EIGRP, OSPF,  

BGP, or IS-IS for example.These protocols will run  their algorithms and will discover the best paths   within this infrastructure. Those best paths will  be placed within the routing information base or   the RIB. Now this is just another name for the  routing table. Using various rules including  

parameters such as administrative distance and  routing protocol metrics, the RIB is populated.   Now from the routing table, another construct  is populated which is called the forwarding   information base or the FIB. It's essentially  a subset of the router's RIB that is typically  

implemented in hardware for efficient forwarding  and contains only the best routes selected for   packet forwarding. Now when packets arrive at  this router here it's the FIB that's used to   determine the next hop to send the packet along  its way. Now when MPLS is enabled and LDP begins  

to exchange labels, we have some new constructs  that are created within each LDP router. The LDP   protocol is used to locally generate a label  for each prefix that is found within the RIB.   This creates the label information base or the  LIB, within which the prefixes in the RIB are  

assigned labels. The labels that are generated  are only locally significant. This means that   their actual values have meaning only for this  particular router. The information in the list   is then used to create the label forwarding  information base or the LFIB. Now this is the  

counterpart to the FIB for routing. Essentially  when a packet arrives at an MPLS router,   it is the LFIB that is examined to determine  the next hop and send the packet along its way.   Now the LFIB is often employed in hardware that is  specifically designed for enabling ultra-efficient  

label switching. Now since labels are locally  significant, MPLS routers must know what label   is being used for a particular prefix by its  neighbors. That way when it forwards it to the   next router in the path, it will replace its own  locally significant label value with the value  

that the neighboring router says it uses for that  particular prefix. So for a destination network   of 192.168.1.0/24, which exists on this customer  network here, this PE router via LDP assigns a   local label of 27 and this assignment is recorded  in the LFIB. But this P router here has assigned  

a locally significant label of 29 for this same  prefix and through LDP, the P router informs the   PE router of this information. So from the point  of view of this PE router here, the local label is   27 and the remote label of the next hop node is  29. Now looking at the LFIB of this router here  

using this command on a Cisco router, we see the  local label as 27 and the outgoing label as 29.   So when this router sends a packet for the prefix  192.168.1.0/24 to this router it will swap its own   label of 27 with the label of the P router which  is 29 before it sends it out. And in the LFIB you  

can also see that we have the outgoing interface  as well as the IP address of the next hop. So   the label switching process at this router has all  the information it needs to send the packet to the   next node in the path. Now the process continues.  When the packet gets to P1 you can see that the  

local label is 29 and the remote label of the next  hop rer is 31 so it swaps the labels and forwards   the packet with the new label to P2 which is the  next hop router. Now what we see in the LFIB for   P2 is interesting. We see the local label here  is 31 but for the outgoing label column which is  

essentially the advertised label from the network  PE router here, we see "Pop Label." This simply   means that before sending the packet, it will  simply remove the label and send it to the peer   with no label at all! This is called "penultimate  label popping." The word "penultimate" simply  

means "second to last." By definition, the last  P router you encounter in the path is always the   second to last router that's encountered along  that path. So this penultimate router will pop   the label or remove it before sending it to the PE  router which is the ultimate or the final router  

in the path. Now we look at the LFIB of this PE  router here, and we'll see that we don't find the   192.168.1.0/24 network in the LFIB at all! Why?  Because PE routers don't have a label assigned to   the networks of the customers to which they are  directly connected. This is because PE routers  

and CE routers will have a routing protocol  running between them, so routing here takes   place using conventional routing protocols and  it is for this reason that the penultimate label   popping takes place. So once the packet reaches  the PE, it is simple routing that will direct it  

to the CE and to the final destination network.  So hopefully going through this process, you can   see how LDP works, how labels are exchanged, how  the LFIB is populated, and how local and remote   labels are used to allow MPLS to operate. Well  I hope you found this video useful and if you  

have please click that thumbs up button! If you  have any questions or feedback please feel free   to leave a comment and, please subscribe to get  updates of newly published videos! I'm Lazarus at   telecom.Tech, I hope this video has been helpful  for you, and I'd like to thank you for watching!