350-501 SPCOR Cisco CCNP Service Provider – MPLS Label Distribution Protocol part 1

1. MPLS – Introduction

Introduction to MPLS. Now in this video I’m going to get into some of the very basics of MPLS technology. Like first we’ll start with traditional IP routing method. Initially we’ll be discussing on that. And then we’ll get into some of the basic MPLS features. And then we will correlate with Cisco’s posts forwarding because SF must be enabled in order for the MPLS to add the labels. So we’ll see that how exactly it is going to do that process. And then we’ll try to understand some of the MPLS terms, the terminology like Label switch, Routers Edge, LSR or something like that. And then finally I’m going to list out some of the major benefits of the MPLS technology. Let’s get started with traditional IP routing method.

Now this is something very basic, the normal routing process, how exactly it happens. Now basically, let’s take an example. I got three routers, Router One, Router Two and Router Three. Now Router One is having some network called ten dot network. It is going to advertise that information. That’s what you can see the update advertises to Router Two and Router Two advertises to Router Three. Now whenever Router Three want to communicate with the Ten dot network, it is going to check the routing table, it is going to see the destination network ID that is and then it is going to see what is the next top address. Probably the next top address will be Router Two. And then it will see what is the exit interface to reach that next stop.

And it will simply forward the packet to Router Two. And Router Two also will do the same thing. Check the routing table, check the next stop and check the exit interface and then forward back to the Router One. Now, this is something what happens in our traditional IP routing method, which means each and every router is going to do hop by hop routing. Lookup. The routing protocols are used to distribution this layer three routing information, which means the router one endor network will be advertised to router three by using some different routing protocols like rip, EHRP, OSP protocols. And then the forwarding decision is based on the packet header by seeing the local routing table.

And then the routing routing lookup is going to be done independently on each and every hop, each and every doctor. Now how it is going to correlate with our MPLS. Now MPLS again relies on the traditional IP routing method which means the IP routing is something like a pre requirement for MPLS to work. So just we need to have a good understanding on that. That’s just a revision. Now let’s see how MPLS is going to work. Now MPLS, let’s take an example. This is my service portal network, my embellish network. Now there is a customer customer site here. Want to communicate with another customer site here, site One, Site two. Now when you are sending the information before it enters the service for network, it is going to enter as a normal IP packet.

Now, once it enters the service port and our router is going to add the label, that’s what we call as an MP list terminology. We call it as push, push the label nothing but adding the label. Now it’s going to add one label here and then it will send a label and then router B is going to swap the label with some other label. Let’s say the router A uses the label 20 and then it is going to advertise that label information to router B. Router B is going to maybe router B is going to add its own label, maybe label 30, it’s going to swap with some other label and then like that. So now here, once it enters the service for network the labels, then your packet will be label switched instead of a normal IP packet.

Which means each nail router is going to add the label and it’s going to only see the label and then forward the packet. So that’s what impulse forwarding mechanism in which the packets are forwarded based on the labels. So we’ll see much more in detail like this process, how this is done by some protocol called LDP Label Distribution Protocol and will practically verify this. So the basic thing that’s what you can see, MPLS label switching, the multiprotocol label switching, it means that the entire traffic is inside the service for is forwarded based on the labels. So when you say multiprotocol now multi protocol. The reason we say multi protocol because now MPLS packets can run on any of the layer two technologies, which means now this link, whatever the link, this can be your ATM.

Link, or it can be your Frame layer link. It can be your PPP link, or it can be Ethernet link. It can be any type of layer two connection. It’s going to run on any layer two technology. So that’s the reason we call as multiprotocol label switching technology. And then this labeling is totally dependent on IP routing, which means there should be a routing information propagated between all these routers by using any of the routing protocol like OSP of EHRP or any of the protocol. And then there should be Safe enabled on these devices because based on the safe, like based on the Safe, we have something called forwarding information based. It’s going to build something called label forwarding information based.

So I’ll come to that much more in detail in my next slide. So this is something about the basic MPL is what we say. So your traffic or your customer traffic enters a normal IP packet and within the service port network it is label switched and then before it goes outside as a normal IP packet. Now, what is the bigger advantage in this? Now, the major advantage in this is the good thing is it’s a label switch which is going to reduce the overhead on the service protocol. Now, normally, if I talk about normal traditional IP routing, when a packet enters the provider, it has to do the routing look up, it has to see the routing table, check the next hop and check the exit interface. And then for the packet, the same thing happens here as well.

The same thing happens here, the same thing happens here. When you talk about traditional IP routing, it’s really going to add some little bit extra overhead for the normal routing lookup. Now this label switching is going to reduce that overhead inside the code and also it is going to minimize the delay. There is one more advantage apart from that. This MPLS is going to provide some lot of applications like MLS, VPNs and then traffic engineering, quality of service, and many more advantages it supports. So not only that, it’s going to support some very good benefits or very good applications, which makes an MLS as a better protocol to run inside the service for network. The next thing, let’s try to correlate how exactly Syscall Express Forwarding is going to help in the label binding information.

Now, if you remember, we have discussed something about Cisco’s Forwarding. Cisco Forwarding is an advanced layer three IP switching technology which is going to optimize the router and it’s going to make the router to forward the packets much faster than usual normal packets. Now, what it is going to do is based on the routing protocol, whatever we are running, let’s say I’m running OSPF, whatever the protocol you’re running, it’s going to build a routing table. That’s what we call as routing information based table. And this routing information based table is generally placed in a control plane. Now this routing information is brought to the data plane that is that is under your, inside your hardware even before a packet errors.

And it’s going to build one thing called Fib Forwarding information based. So that’s what I have in my next slide here. You can see. Now it’s going to use the IP routing information, that is the routing information base which is built by using the different routing protocols. And it’s going to determine, this routing table is going to determine the direction of a packet, of the destination packet and it’s going to decide what is the next stop to forward the packet. Okay, now it is going to bring down the ear that is your FRB table which is prebuilt even before the packet is arriving. So this is something what we have already learned in our basic video in a separate section called Cisco Express Hovering.

Now how it is going to relate that. Now based on your routing table, based on your routing protocol, it’s going to build something called Rib. Now again, we are going to run another protocol called label distribution protocol. Now, what this label distribution protocol is going to do is whatever the routing table is built inside your siff table. That’s what Seth is going to create. One table called Fib, the forwarding Information based table which contains the destination network and the next stop information and exit interface. It’s going to build another table called Label Forwarding Information List Lfiv.

Now what it is going to do is inside the service folder device whichever is running MPLS and both the interfaces are label switch. Both interface are enabled with MPLS. Now any packet enters the router it is not going to do the routing lookup based on the FRB information. It is going to see the label incoming IV packet. Let’s say the incoming packet is if it is a border router. Let’s take an example. If it is a border router which is connecting to if this is the border router of the service for network incoming packet will be normal IP packet. But when it’s code outside it will be going as a label packet. It’s going to see the Label Forwarding Information based table and then forward it to the next order and then the next order again it enters as a label and goes as a label packet.

Again the same thing happens here. Enters as a label packet and goes as a label packet before it reaches the last router of that where the packet enters as a label packet but goes as a normal IP packet. So if it is an edge router then probably there might be a case like if it is a border router your packet may come as a normal IP packet but goes as a label packet. The same thing happens here as well. Your packet as a label packet goes as a normal IP packet because these are the Butter routers. Now here you can see one of the example for your MPLS architecture here. Now you have a control plane. This control plane is responsible for building the routing table based on your routing protocol. I’m using OSP of let’s say.

Now based on this it is going to build something called Fib Forwarding Information based table. And then we are going to run something called LDP on the routers. We are going to enable LDP protocol, label distribution protocol which is responsible for again creating something called Lfid. Now it’s going to prebuild this before even the packet wraps and it’s going to place that in your data plane. Now your packets inside the service folder enters as a label packet. Let’s say in an example with a label of 24 it’s going to swapped with another label called 17 before it goes outside the router. The next thing, let’s try to understand some of the basic terminology here in MPLS. Now in MPLS more typically we call them as Label switch routers.

You can see LSR nothing but label switched routers. Label switch routers are the routers which are going to forward the packets based on the labels and they are going to swap the labels. Now these are all routers we call them as labels which routers because they are going to receive as a label packet and then forwarding as a label label packet. That’s nothing but swapping, swapping the labels. And then the border router, we call it as ingress Edge LSR. Now, this ingress Edge LSR is a router which is going to receive as a normal IP packet and it’s going to forward as a label packet. And then this border router we call them as Edge LSR. Again Edge LSR, but it will be Egress. Egress it is going to receive as a normal label packet and then forward outside as a normal IP packet.

So we call this routers as a LSR routers, label Switch routers and then Edge LSR. In that Edge LSR we have ingress and Egress LSR. And then finally you have something called label switch path. Now, sequence of labels to reach the destination is called Label Switch Path. Now, generally Label switch path represents the path. This path, this complete path we call as Label switch path where the complete forwarding of your packets happens based on the labels, not based on the IP packets it’s going to forward based on the labels. The last thing we’ll see in this section here is the benefits of MLS. Now, MPLS provides a wide range of benefits. It’s going to support a lot of applications like using MPLS network. Through MLS network we can have a communication between two different sites.

We can support something called unicast IP routing, which means we can have a ten dot network communicating over 20 dot network, which can be different networks, different physical geographical locations, more like a lease land connection. It’s going to provide a virtual point to point connection between them. We have discussed that in our previous sections like PTP models, overlay model VPNs, that kind of things. And not only that, it will also allow your multicast traffic to go from one side to another side. You can also have a multicast traffic to allow over the MPs. Now, typically we call them as multicast VPNs. In our MindPlus terminology, it’s a far advanced concept where we are going to integrate our MLS with multicasting.

Now, mPLUS is going to decrease overhead on the routers by forwarding the packets in the form of the labels. So that is one of the advantage we get by using MPLS and it’s going to support something called BGP free core. Now, inside the MLS network we are not running BGP at all here without actually running BGP. We can still allow the packets, the customer traffic to grow from one side to another side without running BGP. So it’s a Bgpfree core and whatever the P routers, the provider routers, they don’t really maintain any of the customer routes. The customer routes will not be maintained except the provider is router. So we’ll see this much more in detail when we get into the implementations.

At the same time, it supports something called it also supports non IP protocols. Now if you want, you can use non IP protocols. It doesn’t make difference because even this MPLS also support for IPV Six traffic. Also if you have an IPCs customer you want, you want to have an IPV Six communication over it. You don’t really need to upgrade, upgrade much inside the service four network. The existing MPLS network also supports IPV Six customers or any other non IP protocols. Because it’s going to forward based on the labels. It doesn’t make difference what is inside the header. It’s not going to see that. It’s going to see only the label information.

Which means now the actual payload, the actual payload, it can be any kind of traffic. Apart from that, it supports something called VPN concept which is most widely used. We have some L two VPNs and L three VPNs. We’ll see that much more in detail later on. Traffic Engineering now there’s something called traffic Engineering. It supports and also it supports quality of service. And then it can transport any transport over MPLS. The connectivity between the P to C. It can be any transport which you can run this technology on PPP or HDLC encapsulation. It works on Ethernet and it comes on frame delay. It can be any transport over MLS. You can provide connectivity for any layer to connections over MPLS.

2. MPLS Labels – MPLS Stacks

This video is going to explain you two things. We start with MLS labels and MPLS stacks. Now, if you want to have a very good understanding on MPLS, you need to have a good idea on the format of MPs label. That is something that we will be focusing in this section. Apart from that, we also see about what are MPLS tags. So let’s get started with MPLS labels. Now, MPs label is a four byte. Now, a four byte means it’s a 32 bit information. You can see this is an impulse label which will be a four byte which is added by the core devices. And this four byte label is going to decide the forwarding traffic. Now, all the forwarding traffic forwarding decisions is is done based on this label inside the service for core networks.

Now, these labels are locally significant on the router. So any log, any label assigned by the router one let’s say router 20 has nothing to do with the label assigned by the router 220. So they are just locally significant labels. Now, in this section we will try to understand the format of this 32 bit or a four byte label. So in this four byte label, the first 20 bits is the label itself. Now this is now this 1st 20 bits is the actual label which is assigned for each packet. Now in this in this 20 bit label, now the first 1516 labels, zero to 15 or we call them as reserve labels. And each and every packet will be assigned a label starting from 16 and it’s going to be the very big range so rooted power of 32.

If you just calculate that is the maximum number which MPLS generally uses to assign a label for each IP packet in the next three bits here you can see the next three bits. We call them as experimental bits. Now, this field is typically used to define the class of service or IP presence values. Now, generally these three bits are generally used if you are implementing quality of service. It is used in providing quality of service over MPLS networks. So it is majorly used in MPLS QS implementations for marking and giving some priority to specific set of traffic or doing some policing and restricting the bandwidth and something like that.

So to provide quality of service over MPLS networks, these three bits are used which is going to carry your marking information done by the quality of service. The next one bit, you can see this one bit we call it as bottom of stack bit. The next one bit we call as bottom of stack bit indicator. It is going to determine if this label, if this label is the last label or not. Now, in general, when we are talking about any MPLS implementations, we’ll discuss in our next section where there is a possibility that normal a single packet can have more than one label. Now, you can have more than one label assigned or it can be two labels or it can be three labels depending upon the different MPs implementations.

Like if I’m using MPLS VPN’s concept now you will have two labels assigned to each packet. If I’m using MLS VPN along with Traffic Engineering, you will have one more auto, one more label which will be called as Traffic engineering label generally. Now this bottom up tag with is going to tell if this label is the last label or not. Now, whenever this field is indicated with one, whenever this bottom of tag bit is set to one, it indicates that this label, whatever the label, this label is the last label of the packet. If that label is not last label, then typically it will be zero value. Now, based on this bottom of stack bit indicator, the MPLS is going to understand whether it is the last label or not the last label. Now, finally you have something called TTL value.

It will be your eight bits. The last eight bits will be your TTL value. This TTL value, it has the same purpose of the TTL field we have in a normal IP packet. The normal IP packet you have TTL value set to 255 and in every hop it is going to decrease to 254 and when it goes to next hop it is going to decrease to 253. Now, it is actually used for preventing the loops in the networks and it is going to decrease in each and every hop. The next thing we’ll see we’ll talk about MPLS label stack. Now, MPLS label stack is something leading to multiple labels assigned to a single IP packet. Normally every packet will be assigned just one label for each and every packet. However, the MPs is going to allow to have multiple labels in the same label stack.

Now, this is generally possible when you are trying to use some applications like MPLS VPNs, but also you’re using MPLS Traffic Engineering or maybe you’re using both MPLS VPNs as well as traffic Engineering. Now, in this you have a normal payload track payload that is your normal packet is going to add it with some label. If you’re using MPLS VPNs we create generally if you talk about MPLS VPNs here MPLS VPNs. If you remember we discussed in an overview, if this is your service for network, it’s going to create one logical, logical connection or logical MP BGP Peering we call it as VPN V Four peering between PE to PE router. Now, by default your normal packet will have a normal payload. It’s going to add default label.

So normal packet added with some label and this label is generally going to do the forwarding decision which is a normal label. And when you are confirming the VPN V Four is going to add one more label, we call it as VPN label. Now, this VPN label is typically we call a second label which is going to add which is going to identify the VPN, the type of the VPN, whereas the first label which is assigned here, it is going to point to the Egress router, it’s going to identify the exit router, whereas the VPN label is going to identify the specific VPN. Now the same case applies if you are using traffic engineering.

Normally if you’re using traffic engineering, traffic engineering generally uses some other protocol called RSVP instead of using LDP for assigning the label for traffic engineering tunnels. Okay. So now in this case if you’re using traffic nearing it’s going to assign two or more than two labels. Like if you’re using traffic nearing you will have a one normal label. This is your normal payload will be added with one label and the top label is going to point to the destination and then it’s going to add one more label here. Now this label is going to point the end point of the tunnel. So normally just like VPNs we also create some traffic engine tunnels from one side to another side. Now the second label is going, the top level is going to point to the end of the tunnels.

So in simple I can say one label is going to point to the destination and the other label is going to point to the end point of the traffic engineers. Now the combination of these applications or if you’re using any advanced features, it might result to have two or more than two labels which is being used instead of one label. And this kind of labels we call as label stack. We’re there is a possibility of having one label or more than one label. So probably you’ll see these things when we start using this MPLS VPNs concept. There will definitely see two kind of labels aside. One for the VPN label that will be the one VPN label and the other one is going to point the Egress router, the exit router which is going to define the direction of the endpoint.