5. IPv6 Support for MPLS
IPV Six over MPLS. Now what if you have a customer who wants to run IPV Six on the customer side and how we exactly provide the IPV Six over MPLS? How exactly it’s going to work? Now the entire concept of IPV Six over MPLS is going to be the similar like what we did for IP version four. The only difference is we don’t don’t really need to upgrade anything inside the service poor network because now we know we already have an impulse running inside the service network and it’s already running IGP. It might be using some OSPA for IP version four. And also we have an LDP running inside the service porder which is going to build Lfid label forwarding information base.
Now we don’t really need to upgrade anything on the service port network, we can still use the same existing LDP and IGP for IP version four and no need to upgrade service part of core for IP version six. The control plane, whatever the control plane we have discussed LDP will be running same and if you have a traffic engineering it will be the same and we are going to run IGP it will be the same. The only thing we need to do is we need to upgrade the P to ceilings. The only thing the customer must be running IPV Six. That is what that’s the reason we actually use MPLS. Now on the P router only the interface facing towards the customer must be running IP version six. So you might be providing IPV Four for the same customer.
So which means we have still MPLS running for IPV Four and the customer is able to communicate with IPV Four to IPV Four end to end. Now we are actually providing IPV Six end to end. Now we just need to run dual stack on this interface. So which means it has to run IPV Four along with IP version six will be running and then P to C routing. We need to do the same thing P to C routing concept and then we have to redistribute the routes into IPV Six VRF. Now VRF concept is going to be the same in IPV Six as well. Actually we are going to use the same route distinguisher values and route target values only. The thing is on this router we need to associate route target values. The Rd value will be the same for both V four and V Six.
We need to associate the route target value with IPV Six. So there is a command called Address family IPV six. We need to associate the route target values so that this routes should exchange on the other end as a VPN V Six routes just like we do a VPN V four exchange of the routes for IPV Four over MPLS. We need to ensure that these route target values must be associated with IPV Six and it will be exchanged with a VPN V Six. Routes and reaches the other end under the VRS. And then again we do redistribution of the routes from IPV Six VRF to IGP, whatever we are using. So the entire concept of MPLS is going to be the same.
The only thing we need to upgrade the PE devices to support IPV Six routing between PE to C and the same thing we do do here and on the P routers, we need to add some VRF configurations under that family IPV Six for IPV Six support. Okay, so the route target values, VRFs Rd values are just appended to the existing IP, appended to the IPV Six form of VPN V Six address. So we can still use the same configurations. It’s going to be the same architectural features like MPLS VPN like we do in IP version four. Only the P routers are going to run dual stack and then the core is going to be the same. No need to upgrade anything inside the core.
So we still use the same existing MPs less core for IP version six support as well. Now the Mpvgp must be distributing the VPN. VPN? RTOs. Families. Like we need to configure VPN V six neighborship from p to p. And that VPN V six neighborship will be established using the same IPV four artisans. And the reason is because we the VPN, it’s going to add one label called VPN Label. And that label will be still IPV Four. So which means the traffic enters the service folder as an IPV Six and it’ll be routed based on the label. And that label will be IPV Four again. So next stop orders will be in this format because it’s going to add the label here and the VRF is going to support your VPN V Four. And VPN v six allows.
6. MPLS traffic Engineering
MPLS traffic Engineering most of the time when you are learning some MPLS concepts, most of the time you’ll see something called traffic engineering inside the Surrey spawn network. You might be listening some words like support for traffic Engineering. In this section I’ll give you some basic overview of traffic engineering. Now, traffic engineering is one of the major concept inside the service pawn network. And the main advantage here is it is going to reduce the overall cost operations by efficiently utilizing the resources. Now, what exactly it will do? It prevents a situation where some parts of the network will be overutilized and whereas some other parts remain underutilized. Let me explain you this point here. Now take an example.
This is my service for core network and I got a customer maybe customer A one, a two and account a three. Now, there is a communication happening between these three different sites of the same customer, customer A. Now customer, let’s take an example. Customer A is trying to send a traffic at 20 MPs from a one to a three. And the customer A two is also trying to send some amount of traffic between a two to a three. Might be 40 mbps. Now what happens? Now the traffic from the customer side from the Ce enters the service core network. Now, the job of the service poor network is to make sure that it reaches the other end of the provider edge router. That is through label switch path. That is what MPLS is going to do.
Now, once the traffic enters the service for network, it has multiple paths to reach the destination, the other edge of the P router. Now, what it is going to do so in this scenario again, I got two paths. One is let’s say this is path one, wire router three. I got one path wire router three. And maybe I got one path wire router six. Not of these two paths that router two is going to use the best route as per the IGP running inside the service for network, I got two paths. It’s going to select any one path as the best route. And that is totally depending on the IGP. Because based on the IGP, it is going to build your forwarding information based table. And based on that IGP, again, it is going to use that as a label forwarding information base and then forward the packet.
Now, maybe according to IGP, this is the best route, this is the best route. And any traffic coming from router two, going to router six, it will always use the path one, which means this particular path is overutilized and the second path will not be used at all. Okay? And the second path will be only used maybe for another P to maybe this route to router maybe or it’s not at all utilized. So in this scenario, I got two paths, whereas one path and it’s going to forward all the traffic via path one. And the second path will be only utilized if the first path goes on. If there is some failure on the path one, then second path is used. Now in this scenario, this link is over utilized.
Now what happens in our scenario here? Now let us come back again to our scenario where a one is actually sending some 20 Mbps of traffic and a two is also sending some 40 Mbps of traffic. It enters the provider’s router. Now, as per our calculation, we are assuming that this is the best path it is going to send via this path. Now, this link is 155 Mbps link. So probably 20 plus 40 is going to be 60 Mbps. So this link says, okay, I have enough bandwidth. I’m going to send it to router three. Because of IGP, it is the best route. And the router says, okay, I have to send it via this router. This is the best path. According to my routing table, it will start sending this path. But if you see this link, this link is just a 34 Mbps link.
But actually we want to send 60 Mbps of traffic. We have to combine both the different sides. Which means only it can send at a time of 34 Mbps of traffic. And it’s going to automatically drop excess of the traffic. Unless if you have some quality of service implemented, it’s going to automatically drop the remaining amount of traffic. Or maybe it is queued. Which means this link is actually overutilized. Whereas we also have an alternate route to reach router four. That is this route. But that route is not at all utilized. If you see the route, it is one gig links but it is not at all utilized. Now this is the problem. In general. We have inside the service core networks where some links might be overutilized which can lead to the drop of the traffic.
Whereas the alternate links are not at all used because they are not the best routes. Now, I want to utilize both the links. I want to utilize both the links. In this case, we can also do some policy based routing. Those stuff what we have learned in the basic things, but that is not scalable solutions. Now what traffic engineering can do in this kind of scenarios now with traffic engineering, what we can do is we can actually configure some tunnel like configurations from B to B. Now we can actually tell the routers, the service order routers that I need 20 Mbps of bandwidth. So we can actually tell the service portal routers that I need at 20 Mbps of bandwidth to go from this P to another P.
Now what service routers are going to do is using this traffic engineering, there is a protocol called RSVP. We’ll be getting into that more in detail when we get into some more in depth of traffic engineering concepts. So based on that protocol. It is going to first calculate or send a request saying that do you have 20 MPs of bandwidth? Now, if the response is yes, if we get a positive response and then it will pass on this message to next starter do we have 20 MB of bandwidth? Because I want to send 20 minutes of bandwidth, it says yes automatically it is going to use this route. Now let’s take an example. I got one more. Customer also want to send 40 megabits of bandwidth.
Now according to this, our calculation, this is the best route. It’s going to this path. It says do you have 40 Mbps of bandwidth? It says yes. Again it will send the request here. The request goes here do you have 40 Mbps of bandwidth? It says no because this link is just 34 Mbps of link and out of that 34 Mbps, 20 Mbps is already utilized by customer a one side and it is just having only 14 Mbps of available bandwidth. Now, if there is a negative response on this side, if the respondent book realizes that on the other side we don’t have enough bandwidth, it will automatically start redirecting from alternate route. It will again check this root is having 40 megs of bandwidth. It says yes, this route is having 40 megs of bandwidth. Yes.
40 meters of bandwidth. Yes. In that case, it will start using the alternate link without actually it has nothing to do with OSPF. So according to OSPF calculation, this might be the best route. But MPL Traffic engineering is not going to rely on the OSPF to decide what route it has to take. It will only work just based on the requirement. Based on the requirement, if my requirement is 40 MPs, it will see the first route. If that first route is not having available bandwidth in a bandwidth, it will start verifying the second route. If that second route is also not having a required bandwidth, if we have a third route, it will start using the third route. Now, in this way, we are actually utilizing all the possible links more efficiently than before.
Because in the previous case, without having any exiting, it will be utilizing only one link and totally based on the OSP of best route. But here we are not doing that. We are not actually relying on complete OSPF. In fact, still you need to rely on OSPF to know each and every possible route. But we are not not using the best route given by the OSPO. We’ll see actually the forwarding of the traffic is done based on the requirement of the available bandwidth. So that is something Traffic engineering is going to do. So it’s going to be the very important concept. If you’re working inside the service for network and if you’re preparing for CCI service exams, you will be definitely tested on questions on MPLS Traffic Engineering.
7. INter AS MPLS L3 VPN
Now in this section, I’ll give you some basic overview on inter as options. Now, intersection is one of the major core topic inside the service for network. If you are working in the service for a core, and also if you are preparing for CCI service, for example, we got multiple options to provide interas MPLS VPNs. We call them as when exactly this intersection is more applicable. Now, let’s take an example. I got customer A, the customer side A one want to exchange their routes with the customer side A two. And to provide the connectivity, I’m going to take the MPLS as my service portal back one. Okay, this is my MPLS service portal backbone, which is providing the reachability end to end using the L three VPNs.
But now you can come across a scenario where you have a customer side in India probably, and maybe in Dubai you got another customer side. You don’t have it. Maybe you’re not exactly using the same service folder because on both the sides you may have two different service Fodder associated. Now, what if you want to provide a connectivity between your two sites which are associated with two different service portals? So that’s where your enter is. MPLS VPN comes into picture. Now, what exactly we do here is we configure PTC routing between these two routers here, and you’re advertising the routes, and the router four is going to place those routes in a VR routing table. And then we’ll configure VPN V four pairing between our P router and to the border router here. The same thing happens here also.
Now we got a customer router exchange P to C routing. It goes to the P router and the peer router is going to have a VPN V four pairing between these two routers. And then just like we do EBGP pairing, we are going to have an EBGP pairing between these two routers. EBGP VPN vivo pairing. So using this EBGP VPN vivo pairing, we are actually having end to end VPN. So either there are multiple options in this, we got an option A, option B, option C. In each and every option we got a different set of configurations we actually use. But the end result is in case if you have a customer site, customer peer router, the service part of peer router and the other end of the peer router belongs to two different customers.
How we can provide end to end labels, which path, how we can. So both the service portal has to come on a common interrace option to use. And based on that, we can still have a communication between the two different customer sites, even though they are connecting between two different responses. So we call this as interrays MPLS L three VPNs. So this is one of the major concept. If you’re preparing for CCI service. For example, you will definitely be tested on some intersections. Maybe depending upon the scenarios, you may be tested. Option A. Option B. Option C. The basic difference between these options are actually the way you configure the things, the way we exchange the routes between the two. Or more different service for us.