BGP Load Balancing Single Homed- Understanding Per Packet and Per Destination LB

hello everyone my name is Sandeep welcome back to another session of BGP wherein we are going to discuss BGP load balancing with loopback IP address and we will understand what is load balancing okay so today we have got this topology wherein we are going to discuss how we are going to load balance the traffic from autonomous system 100 to autonomous system 200 and we are going to configure that using look back I Patricia so in our topology we will do a BGP peering from our one loop back I play this 1 1 1 1 2 are 2 loop back IP address and we are going to advertise some routes from router 2 to router 1 and we will see how the route looks into the BGP table and when we do a ping or trace route so we should see that there the one packet is going or the first link and the another packet is going over the second link ok so let's see we have got these interfaces Ethernet 0 / 0 & 0 / 1 here ok so before we start with this let's understand how many types of load balancing we have and what is what is the meaning of load balancing okay so let's say that we have these two routers let me take a screenshot of this and I'll try to explain there what is load balancing or load sharing okay so here we are going to discuss let's say I have a route here with 200 dot 0 dot 0 dot 0 and I learned this route on the router 1 okay so I have this route on to the router 1 now let's say I have I have some computers in my land who wants to access this 200 0 0 1 network or 200 0 0 0 network which is into the service provider and let's say I have a server here 200 0 0 1 so when I want to access this 200 0 0 1 server from this client machine so the traffic goes here on to the router 1 and router 1 has two parts to the 200 0 0 1 1 by Ethernet 0 slash 0 and 1 y 800 0 / 1 okay so when you configure the load balancing in your routing table you you will see the to exit paths for this destination so one packet will go where at this interface the other packet will go at this interface ok so this is called the load balancing which means that the traffic is distributed between the two links ok so basically there are two types of load balancing one is called as per destination load balancing and the another is called as per packet load balancing okay so we will discuss about the per destination load balancing first now the same topology let's say I have one more network here with 200 and or 130 dot 0 dot 0 dot 0 okay so this network is learnt on router 1 as well and 200 0 0 0 network is also learnt on router 1 so when this computer wants to access 130 Network it will take the first path Ethernet 0 / 1 when he's accessing 200 0 0 0 Network it will take the second path this can link this is called poor destination load-balancing which means means that the router one has destination 130 as well as 200 and it has two links so whenever any users trying to access the first destination it will use the first link and when the users from our land side trying to access the second destination it will use the second link this is called as per destination load-balancing the second type of load balancing we have got is per packet load balancing okay so per packet load balancing is that let's say you have only a single Network let's say we don't have one 30 Network we have only 200 zero zero zero Network and you are accessing this server from this client let's say my destination server is 200 zero zero one so whenever I'm accessing this 200 zero zero one from my client side one packet will be sent on Ethernet 0/1 one packet will be sent on Ethernet 0/0 this is called as per packet load balancing so when when you configure the per packet load balancing it is going to take more processed processor hits and the CPU and RAM will be utilized more on the routers so most of the routers by default are configured as per destination load balancing okay please understand that by default all the routers comes with a per load the load balancing scheme which is per destination load balancing which is written here you can disable the per you know you can disable this per destination load balancing and enable per packet load balancing using a single command on the it on the Ethernet interface that is no IP route cache so if you want to enable the per packet load balancing on this router on this interface you are going to configure this command no IP route cache also on the second interface you can configure this command so the per packet load-balancing will be enabled here all the new routers are coming with a new switching scheme that is called as safe or cisco express forwarding and mostly the new Ras after Levin dot X are using safe by default so safe has the capability to to do both of this per packet and per destination load balancing so let's read some notes from the Cisco EPSA directly so here you see that per destination load balancing means the router distributes the packet based on the destination address we have discussed this if you have two to two paths to the same network all packaged all package to destination one will go or one path and all to the destination two will go another path that's what we discussed here perfect load balancing also we have discussed one packet will be sent on first link and second packet will be sent on second link so per destination or per packet load balancing depends on the type of switching scheme used in used for IP packets so routers are performing our routing so IP packets are observed at the router by default on most cisco routers fast switching is enabled on under the interface fast switching is nothing but your per packet load balancing if you want to disable the fast switching and enable process switching then you can use this command no IP route cache okay I hope that overview was clear to you so let's go ahead and do our configuration in the related to the BGP and we will enable the load balancing in a single home BGP environment single home means that you have a single router from your customer edge to the customer edge to the provider edge if you have multiple router let's say I have one more router then it is called as dual home in this case if I have one more router here and I have one link going to this router then that is called as dual home okay so in today's topology we are going to use a simple topology and we'll discuss this so let's jump onto the router one so let me start working on the router one let me execute a command show IP interface brief quickly to see if we have any IP addresses so okay so I have these two interfaces configured interface loopback one idle configure and I will configure IP address one dot one dot one dot one okay so what I have here is I have configured these interfaces okay so 10-1 12.1 23.1 and 1 1 1 this is loopback IP address and these two are the Ethernet interfaces so let's do the similar configuration on router 2 interface loopback to IP address quickly verify our interfaces okay so we have these interfaces configured and we have this up so let's start doing the BGP configurations with this on the router one route of BGP let's say 100 and I am forming a neighbor ship okay so when I'm having the loopback interfaces we are going to configure the BGP neighbor ship with this loopback loopback IP addresses and not with any of the IP addresses here because if this link goes down let's say I have configured busy preparing with 10.1.1.1 goes down then you know you don't have the connectivity with BGP or the BGP will go down so the best option for us to configure the loopback IP addresses on the router and we form a BGP neighbor ship with the loopback IP addresses this is the reason we have configured loopback IP addresses so that we can have we can form a BGP neighbor ship there okay so just one more thing if you notice that before we start doing a BGP configuration we should have a rich ability to 1.1.1 donate to door-to-door to dot network from ping I should have a rich ability but it will I will not be having any rich ability because we don't have this route if you do show IP route you don't see this route here because we are not running any routing protocol so to get this route advertised into the router one we configure any of the IGP we can configure either the static route or eid are pure OSPF or any routing protocol in most of the scenarios it will just go ahead and configure the network and now we should see that I have the two door-to-door to network here in my topology so I have the reach ability from my source IP address of one dot one dot one dot one okay so now I have the reach ability to the two door-to-door to dot network I'm good to configure the BGP now Rooter BGP hundred me per one dot one dot one dot d / 2.2 dot 2.2 and remote autonomous system is what let me do a similar configuration here of T Rooter BGP 200 neighbor one dot one dot one dot one remote autonomous system is hundred so IP with the P summary okay so you will still not be able to form a BJP leadership here it will be cheap in a worship the reason being that you still have many configurations to configure in the BGP there are many configurations are missing here into the BGP so the first configuration is that you need to configure the update source with its own loopback IP address because what happens is that when you send a packet from this router here okay so I'm going to send the packet from this interface so my source the BGP sauce packet will be from 2.2 dot 2.2 similarly I will be sourcing my BGP updates from 1.1.11 so that the three-way handshake of TCP will happen this what would happen is that if I don't configure this loop back to to door to door to door to as my update source it will be source from one of the interface let's say I am sending a packet source packet with 10.1 or 12.2 the same packet and then when I'm getting a back cynic packet I will be getting it from some of the interfaces I'm gonna get the update from 10.1 or 23.2 so out disappear out of TCP out of packet things happens here and the BGP will not be formed there's one one more piece in the BGP which is missing see that it is still it is still idle state one more piece of it is that we need to configure ebgp multi how because this network here let's say 2.2 2.2 network is two hops away from one dot one dot one dot network so the first hop is this one and then it has to go here go here so it is 2 hops away from the to door to door to network because my first loopback interface is here let's say my first hop is here so I should configure ebgp multi-hop here you could say that it has not because we are not you could see that I have received this doubt from BGP hundred you could see that this route is received here 0-200 zero zero zero okay so to raise this about 200 Network I need you know it is showing that the routing is learn from to door to door to door to so to door to door to door to is nothing but the loopback IP address of look like IP address of router to could see that this to door to door to network is learned from two interfaces one is Ethernet 0 / 1 + 1 is saying that it is learned from eternal 0/0 okay you could also see that the traffic share count is 1 here and traffic share count is 1 here which means that you know the traffic is equally distributed between the two links 800 0 / 0 + 0 / 1 okay you could see that the load balancing is happening here the first packet is going where 23.2 okay so you could see that the first packet is going Maya okay 23.2 23.2 is nothing but this interface Ethernet 0 / 0 and then it is moving without went 12.2 so the so the very next packet is going back to 0 / 1 ok so again the next packet is going like this so it is doing a load balancing across the two links basically we are using the BGP here however it is doing the load balancing using the ICP IGP we have used his AI j RP routing protocol and it is doing the load balancing using a IJ t so if you do a show IP route however 0 you could see that if you look at the routing table you are seeing that this this to raise this 200 network I need to send a packet to to door to door to door to or the 200 network is learn from to door to door to door - so this router will do one more lookup routing lookup and then it will define how to route the traffic so if you look at this you would think that the traffic is not getting shared ok so basically the load balancing is not happening using the bgp routing protocol it is happening using the ARP routing protocol if you see that traffic share count is 1 here okay however when you do a trace route it is using so how it would be done is you see so I put out to the to door-to-door to network you could see that it has it has two putts routing entry no no two parts one is eternal zero slash zero and one is why it ended 0 slash point let's do an fix and also we will do one more command use a command called maximum path if you want to if you really want to do a load balancing using any other routing protocol for example if you want to do it by BGP then you need to configure it it allows multiple paths okay however it's showing that it has only a single path to reach for at 2.2 or 2.2 and you know internally it is doing the load balancing using ICP that is ERP we are using so in my next video I'm going to discuss the BGP multihoming wherein I will show how we can use this maximum path command in BGP to load balance the traffic and we will discuss that in more details if you liked this video please hit the like button and subscribe to my channel to get more videos thanks a lot for watching and have a good day

Channel: Sandeep S

Views: 1,086

Keywords:

Id: 9TCoIFIn3sE

Length: 25min 30sec (1530 seconds)

Published: Fri Oct 26 2018